1 systemd System and Service Manager
5 * The service manager gained basic support for cgroup v2 freezer. Units
6 can now be suspended or resumed either using new systemctl verbs,
7 freeze and thaw respectively, or via D-Bus.
9 * PID 1 may now automatically load pre-compiled AppArmor policies from
10 /etc/apparmor/earlypolicy during early boot.
12 * The CPUAffinity= setting in service unit files now supports a new
13 special value "numa" that causes the CPU affinity masked to be set
14 based on the NUMA mask.
16 * systemd will now log about all left-over processes remaining in a
17 unit when the unit is stopped. It will now warn about services using
18 KillMode=none, as this is generally an unsafe thing to make use of.
20 * Two new unit file settings
21 ConditionPathIsEncrypted=/AssertPathIsEncrypted= have been
22 added. They may be used to check whether a specific file system path
23 resides on a block device that is encrypted on the block level
24 (i.e. using dm-crypt/LUKS).
26 * Another pair of new settings ConditionEnvironment=/AssertEnvironment=
27 has been added that may be used for simple environment checks. This
28 is particularly useful when passing in environment variables from a
29 container manager (or from PAM in case of the systemd --user
32 * .service unit files now accept a new setting CoredumpFilter= which
33 allows configuration of the memory sections coredumps of the
34 service's processes shall include.
36 * .mount units gained a new ReadWriteOnly= boolean option. If set
37 it will not be attempted to mount a file system read-only if mounting
38 in read-write mode doesn't succeed. An option x-systemd.rw-only is
39 available in /etc/fstab to control the same.
41 * .socket units gained a new boolean setting PassPacketInfo=. If
42 enabled, the kernel will attach additional per-packet metadata to all
43 packets read from the socket, as ancillary message. This controls the
44 IP_PKTINFO, IPV6_RECVPKTINFO, NETLINK_PKTINFO socket options,
45 depending on socket type.
47 * .service units gained a new setting RootHash= which may be used to
48 specify the root hash for verity enabled disk images which are
49 specified in RootImage=. RootVerity= may be used to specify a path to
50 the Verity data matching a RootImage= file system. (The latter is
51 only useful for images that do not contain the Verity data embedded
52 into the same image that carries a GPT partition table following the
53 Discoverable Partition Specification). Similarly, systemd-nspawn
54 gained a new switch --verity-data= that takes a path to a file with
55 the verity data of the disk image supplied in --image=, if the image
56 doesn't contain the verity data itself.
58 * .service units gained a new setting RootHashSignature= which takes
59 either a base64 encoded PKCS#7 signature of the root hash specified
60 with RootHash=, or a path to a file to read the signature from. This
61 allows validation of the root hash against public keys available in
62 the kernel keyring, and is only supported on recent kernels
63 (>= 5.4)/libcryptsetup (>= 2.30). A similar switch has been added to
64 systemd-nspawn and systemd-dissect (--root-hash-sig=). Support for
65 this mechanism has also been added to systemd-veritysetup.
67 * .service unit files gained two new options
68 TimeoutStartFailureMode=/TimeoutStopFailureMode= that may be used to
69 tune behaviour if a start or stop timeout is hit, i.e. whether to
70 terminate the service with SIGTERM, SIGABRT or SIGKILL.
72 * Most options in systemd that accept hexadecimal values prefixed with
73 0x in additional to the usual decimal notation now also support octal
74 notation when the 0o prefix is used and binary notation if the 0b
77 * Unit files, tmpfiles.d/ snippets, sysusers.d/ snippets and other
78 configuration files that support specifier expansion learnt six new
79 specifiers: %a resolves to the current architecture, %o/%w/%B/%W
80 resolve to the various ID fields from /etc/os-release, %l resolves to
81 the "short" hostname of the system, i.e. the hostname configured in
82 the kernel truncated at the first dot.
84 * Support for the .include syntax in unit files has been removed. The
85 concept has been obsolete for 6 years and we started warning about
86 its pending removal 2 years ago (also see NEWS file below). It's
89 * StandardError= and StandardOutput= in unit files no longer support
90 the "syslog" and "syslog-console" switches. They were long removed
91 from the documentation, but will now result in warnings when used,
92 and be converted to "journal" and "journal+console" automatically.
94 * If the service setting User= is set to the "nobody" user, a warning
95 message is now written to the logs (but the value is nonetheless
96 accepted). Setting User=nobody is unsafe, since the primary purpose
97 of the "nobody" user is to own all files whose owner cannot be mapped
98 locally. It's in particular used by the NFS subsystem and in user
99 namespacing. By running a service under this user's UID it might get
100 read and even write access to all these otherwise unmappable files,
101 which is quite likely a major security problem.
103 * A new kernel command line option systemd.hostname= has been added
104 that allows controlling the hostname that is initialized early during
107 * A kernel command line option "udev.blockdev_read_only" has been
108 added. If specified all hardware block devices that show up are
109 immediately marked as read-only by udev. This option is useful for
110 making sure that a specific boot under no circumstances modifies data
111 on disk. Use "blockdev --setrw" to undo the effect of this, per
114 * A new boolean kernel command line option systemd.swap= has been
115 added, which may be used to turn off automatic activation of swap
116 devices listed in /etc/fstab.
118 * New kernel command line options systemd.condition-needs-update= and
119 systemd.condition-first-boot= have been added, which override the
120 result of the ConditionNeedsUpdate= and ConditionFirstBoot=
123 * A new kernel command line option systemd.clock-usec= has been added
124 that allows setting the system clock to the specified time in µs
125 since Jan 1st, 1970 early during boot. This is in particular useful
126 in order to make test cases more reliable.
128 * The fs.suid_dumpable sysctl is set to 2 / "suidsafe". This allows
129 systemd-coredump to save core files for suid processes. When saving
130 the core file, systemd-coredump will use the effective uid and gid of
131 the process that faulted.
133 * The /sys/module/kernel/parameters/crash_kexec_post_notifiers file is
134 now automatically set to "Y" at boot, in order to enable pstore
135 generation for collection with systemd-pstore.
137 * A new 'hwdb' file has been added that collects information about PCI
138 and USB devices that correctly support auto-suspend, on top of the
139 databases for this we import from the ChromiumOS project. If you have
140 a device that supports auto-suspend correctly and where it should be
141 enabled by default, please submit a patch that adds it to the
142 database (see /usr/lib/udev/hwdb.d/60-autosuspend.hwdb).
144 * systemd-udevd gained new configuration option timeout_signal= as well
145 as corresponding kernel command line option udev.timeout_signal=.
146 The option can be used to configure the UNIX signal that the main
147 daemon sends to the worker processes on timeout. Setting the signal
148 to SIGABRT is useful for debugging.
150 * .link files managed by systemd-udevd gained options RxFlowControl=,
151 TxFlowControl=, AutoNegotiationFlowControl= in the [Link] section, in
152 order to configure various flow control parameters. They also gained
153 RxMiniBufferSize= and RxJumboBufferSize= in order to configure jumbo
154 frame ring buffer sizes.
156 * networkd.conf gained a new boolean setting ManageForeignRoutes=. If
157 enabled systemd-networkd manages all routes configured by other tools.
159 * .network files managed by systemd-networkd gained a new section
160 [SR-IOV], in order to configure SR-IOV capable network devices.
162 * systemd-networkd's [IPv6Prefix] section in .network files gained a
163 new boolean setting Assign=. If enabled an address from the prefix is
164 automatically assigned to the interface.
166 * systemd-networkd's [Network] section gained a new setting
167 IPv6PDSubnetId= that allows explicit configuration of the preferred
168 subnet that networkd's Prefix Delegation logic assigns to interfaces.
170 * systemd-networkd's [Network] section gained a new setting
171 IPv4AcceptLocal=. If enabled the interface accepts packets with local
174 * systemd-networkd gained support for configuring the HTB queuing
175 discipline in the [HierarchyTokenBucket] and
176 [HierarchyTokenBucketClass] sections. Similar the "pfifo" qdisc may
177 be configured in the [PFIFO] section, "GRED" in
178 [GenericRandomEarlyDetection], "SFB" in [StochasticFairBlue], "cake"
179 in [CAKE], "PIE" in [PIE], "DRR" in [DeficitRoundRobinScheduler] and
180 [DeficitRoundRobinSchedulerClass], "BFIFO" in [BFIFO],
181 "PFIFOHeadDrop" in [PFIFOHeadDrop], "PFIFOFast" in [PFIFOFast], "HHF"
182 in [HeavyHitterFilter], "ETS" in [EnhancedTransmissionSelection] and
183 "QFQ" in [QuickFairQueueing] and [QuickFairQueueingClass].
185 * systemd-networkd gained support for a new Termination= setting in the
186 [CAN] section for configuring the termination resistor. It also
187 gained a new ListenOnly= setting for controlling whether to only
188 listen on CAN interfaces, without interfering with traffic otherwise
189 (which is useful for debugging/monitoring CAN network
190 traffic). DataBitRate=, DataSamplePoint=, FDMode=, FDNonISO= have
191 been added to configure various CAN-FD aspects.
193 * systemd-networkd's [DHCPv6] section gained a new option WithoutRA=.
194 When enabled, DHCPv6 will be attempted right-away without requiring an
195 Router Advertisement packet suggesting it first (i.e. without the 'M'
196 or 'O' flags set). The [IPv6AcceptRA] section gained a boolean option
197 DHCPv6Client= that may be used to turn off the DHCPv6 client even if
198 the RA packets suggest it.
200 * systemd-networkd's [DHCPv4] section gained a new setting UseGateway=
201 which may be used to turn off use of the gateway information provided
202 by the DHCP lease. A new FallbackLeaseLifetimeSec= setting may be
203 used to configure how to process leases that lack a lifetime option.
205 * systemd-networkd's [DHCPv4] and [DHCPServer] sections gained a new
206 setting SendVendorOption= allowing configuration of additional vendor
207 options to send in the DHCP requests/responses. The [DHCPv6] section
208 gained a new SendOption= setting for sending arbitrary DHCP
209 options. RequestOptions= has been added to request arbitrary options
210 from the server. UserClass= has been added to set the DHCP user class
213 * systemd-networkd's [DHCPServer] section gained a new set of options
214 EmitPOP3=/POP3=, EmitSMTP=/SMTP=, EmitLPR=/LPR= for including server
215 information about these three protocols in the DHCP lease. It also
216 gained support for including "MUD" URLs ("Manufacturer Usage
217 Description"). Support for "MUD" URLs was also added to the LLDP
218 stack, configurable in the [LLDP] section in .network files.
220 * systemd-networkd's .netdev files now support a new setting
221 VLANProtocol= in the [Bridge] section that allows configuration of
222 the VLAN protocol to use.
224 * systemd-networkd supports a new Group= setting in the [Link] section
225 of the .network files, to control the link group.
227 * systemd-networkd's [Network] section gained a new
228 IPv6LinkLocalAddressGenerationMode= setting, which specifies how IPv6
229 link local address is generated.
231 * A new default .network file is now shipped that matches TUN/TAP
232 devices that begin with "vt-" in their name. Such interfaces will
233 have IP routing onto the host links set up automatically. This is
234 supposed to be used by VM managers to trivially acquire a network
235 interface which is fully set up for host communication, simply by
236 carefully picking an interface name to use.
238 * A new boolean option AssignAcquiredDelegatedPrefixAddress= has been
239 added to the [DHCPv6] section of .network files. If enabled (which is
240 the default) an address from any acquired delegated prefix is
241 automatically chosen and assigned to the interface.
243 * systemd-networkd's [DHCPv6] section gained a new setting RouteMetric=
244 which sets the route priority for routes specified by the DHCP server.
246 * systemd-networkd's [DHCPv6] section gained a new setting VendorClass=
247 which configures the vendor class information sent to DHCP server.
249 * The BlackList= settings in .network files' [DHCPv4] and
250 [IPv6AcceptRA] sections have been renamed DenyList=. The old names
251 are still understood to provide compatibility.
253 * networkctl gained the new "forcerenew" command for forcing all DHCP
254 server clients to renew their lease. The interface "status" output
255 will now show numerous additional fields of information about an
256 interface. There are new "up" and "down" commands to bring specific
257 interfaces up or down.
259 * systemd-resolved's DNS= configuration option now optionally accepts
260 DNS server addresses suffixed by "#" followed by a host name. If
261 used, the DNS-over-TLS certificate is validated to match the
264 * systemd-resolved may be configured to forward single-label DNS names.
265 This is not standard-conformant, but may make sense in setups where
266 public DNS servers are not used.
268 * systemd-resolved's DNS-over-TLS support gained SNI validation.
270 * systemd-nspawn's --resolv-conf= switch gained a number of new
271 supported values. Specifically, options starting with "replace-" are
272 like those prefixed "copy-" but replace any existing resolv.conf
273 file. And options ending in "-uplink" and "-stub" can now be used to
274 propagate other flavours of resolv.conf into the container (as
275 defined by systemd-resolved).
277 * The various programs included in systemd can now optionally output
278 their log messages on stderr prefixed with a timestamp, controlled by
279 the $SYSTEMD_LOG_TIME environment variable.
281 * systemctl gained a new "-P" switch that is a shortcut for "--value
284 * "systemctl list-units" and "systemctl list-machines" no longer hide
285 their first output column with --no-legend. To hide the first column,
288 * "systemctl reboot" takes the option "--reboot-argument=".
289 The optional positional argument to "systemctl reboot" is now
290 being deprecated in favor of this option.
292 * systemd-run gained a new switch --slice-inherit. If specified the
293 unit it generates is placed in the same slice as the systemd-run
296 * systemd-journald gained support for zstd compression of large fields
297 in journal files. The hash tables in journal files have been hardened
298 against hash collisions. This is an incompatible change and means
299 that journal files created with new systemd versions are not readable
300 with old versions. If the $SYSTEMD_JOURNAL_KEYED_HASH boolean
301 environment variable for systemd-journald.service is set to 0 this
302 new hardening functionality may be turned off, so that generated
303 journal files remain compatible with older journalctl
306 * journalctl will now include a clickable link in the default output for
307 each log message for which an URL with further documentation is
308 known. This is only supported on terminal emulators that support
309 clickable hyperlinks, and is turned off if a pager is used (since
310 "less" still doesn't support hyperlinks,
311 unfortunately). Documentation URLs may be included in log messages
312 either by including a DOCUMENTATION= journal field in it, or by
313 associating a journal message catalog entry with the log message's
314 MESSAGE_ID, which then carries a "Documentation:" tag.
316 * journald.conf gained a new boolean setting Audit= that may be used to
317 control whether systemd-journald will enable audit during
320 * when systemd-journald's log stream is broken up into multiple lines
321 because the PID of the sender changed this is indicated in the
322 generated log records via the _LINE_BREAK=pid-change field.
324 * journalctl's "-o cat" output mode will now show one or more journal
325 fields specified with --output-fields= instead of unconditionally
326 MESSAGE=. This is useful to retrieve a very specific set of fields
327 without any decoration.
329 * coredumpctl gained a new --file= switch, matching the same one in
330 journalctl: a specific journal file may be specified to read the
333 * coredumps collected by systemd-coredump may now be compressed using
336 * systemd-binfmt gained a new switch --unregister for unregistering all
337 registered entries at once. This is now invoked automatically at
338 shutdown, so that binary formats registered with the "F" flag will
339 not block clean file system unmounting.
341 * systemd-notify's --pid= switch gained new values: "parent", "self",
342 "auto" for controlling which PID to send to the service manager: the
343 systemd-notify process' PID, or the one of the process invoking it.
345 * systemd-logind's Session bus object learnt a new method call
346 SetType() for temporarily updating the session type of an already
347 allocated session. This is useful for upgrading tty sessions to
348 graphical ones once a compositor is invoked.
350 * systemd-socket-proxy gained a new switch --exit-idle-time= for
351 configuring an exit-on-idle time.
353 * systemd-repart's --empty= setting gained a new value "create". If
354 specified a new empty regular disk image file is created under the
355 specified name. Its size may be specified with the new --size=
356 option. The latter is also supported without the "create" mode, in
357 order to grow existing disk image files to the specified size. These
358 two new options are useful when creating or manipulating disk images
359 instead of operating on actual block devices.
361 * systemd-repart drop-ins now support a new UUID= setting to control
362 the UUID to assign to a newly created partition.
364 * systemd-repart's SizeMin= per-partition parameter now defaults to 10M
367 * systemd-repart's Label= setting now support the usual, simple
370 * systemd-homed's LUKS backend gained the ability to discard empty file
371 system blocks automatically when the user logs out. This is enabled
372 by default to ensure that home directories take minimal space when
373 logged out but get full size guarantees when logged in. This may be
374 controlled with the new --luks-offline-discard= switch to homectl.
376 * If systemd-homed detects that /home/ is encrypted as a whole it will
377 now default to the directory or subvolume backends instead of the
378 LUKS backend, in order to avoid double encryption. The default
379 storage and file system may now be configured explicitly, too, via
380 the new /etc/systemd/homed.conf configuration file.
382 * systemd-homed now supports unlocking home directories with FIDO2
383 security tokens that support the 'hmac-secret' extension, in addition
384 to the existing support for PKCS#11 security token unlocking
385 support. Note that many recent hardware security tokens support both
386 interfaces. The FIDO2 support is accessible via homectl's
387 --fido2-device= option.
389 * homectl's --pkcs11-uri= setting now accepts two special parameters:
390 if "auto" is specified and only one suitable PKCS#11 security token
391 is plugged in, its URL is automatically determined and enrolled for
392 unlocking the home directory. If "list" is specified a brief table of
393 suitable PKCS#11 security tokens is shown. Similar, the new
394 --fido2-device= option also supports these two special values, for
395 automatically selecting and listing suitable FIDO2 devices.
397 * The /etc/crypttab tmp option now optionally takes an argument
398 selecting the file system to use. Moreover, the default is now
399 changed from ext2 to ext4.
401 * There's a new /etc/crypttab option "keyfile-erase". If specified the
402 key file listed in the same line is removed after use, regardless if
403 volume activation was successful or not. This is useful if the key
404 file is only acquired transiently at runtime and shall be erased
405 before the system continues to boot.
407 * There's also a new /etc/crypttab option "try-empty-password". If
408 specified, before asking the user for a password it is attempted to
409 unlock the volume with an empty password. This is useful for
410 installing encrypted images whose password shall be set on first boot
411 instead of at installation time.
413 * systemd-cryptsetup will now attempt to load the keys to unlock
414 volumes with automatically from files in
415 /etc/cryptsetup-keys.d/<volume>.key and
416 /run/cryptsetup-keys.d/<volume>.key, if any of these files exist.
418 * systemd-cryptsetup may now activate Microsoft BitLocker volumes via
419 /etc/crypttab, during boot.
421 * logind.conf gained a new RuntimeDirectoryInodesMax= setting to
422 control the inode limit for the per-user $XDG_RUNTIME_DIR tmpfs
425 * A new generator systemd-xdg-autostart-generator has been added. It
426 generates systemd unit files from XDG autostart .desktop files, and
427 may be used to let the systemd user instance manage services that are
428 started automatically as part of the desktop session.
430 * "bootctl" gained a new verb "reboot-to-firmware" that may be used
431 to query and change the firmware's 'reboot into firmware' setup flag.
433 * systemd-firstboot gained a new switch --kernel-command-line= that may
434 be used to initialize the /etc/kernel/cmdline file of the image. It
435 also gained a new switch --root-password-hashed= which is like
436 --root-password= but accepts a pre-hashed UNIX password as
437 argument. The new option --delete-root-password may be used to unset
438 any password for the root user (dangerous!). A new --force option may
439 be used to override any already set settings with the parameters
440 specified on the command line (by default, the tool will not override
441 what has already been set before, i.e. is purely incremental).
443 * systemd-firstboot gained support for a new --image= switch, which is
444 similar to --root= but accepts the path to a disk image file, on
445 which it then operates.
447 * A new sd-path.h API has been added to libsystemd. It provides a
448 simple API for retrieving various search paths and primary
449 directories for various resources.
451 * A new call sd_notify_barrier() has been added to the sd-daemon.h
452 API. The call will block until all previously sent sd_notify()
453 messages have been processed by the service manager. This is useful
454 to remove races caused by a process already having disappeared at the
455 time a notification message is processed by the service manager,
456 making correct attribution impossible. The systemd-notify tool will
457 now make use of this call implicitly, but this can be turned off again
458 via the new --no-block switch.
460 * When sending a file descriptor (fd) to the service manager to keep
461 track of, using the sd_notify() mechanism, a new parameter FDPOLL=0
462 may be specified. If passed the service manager will refrain from
463 poll()ing on the file descriptor. Traditionally (and when the
464 parameter is not specified), the service manager will poll it for
465 POLLHUP or POLLERR events, and immediately close the fds in that
468 * The service manager (PID1) gained a new D-Bus method call
469 SetShowStatus() which may be used to control whether it shall show
470 boot-time status output on the console. This method has a similar
471 effect to sending SIGRTMIN+20/SIGRTMIN+21 to PID 1.
473 * The sd-bus API gained a number of convenience functions that take
474 va_list arguments rather than "...". For example, there's now
475 sd_bus_call_methodv() to match sd_bus_call_method(). Those calls make
476 it easier to build wrappers that accept variadic arguments and want
477 to pass a ready va_list structure to sd-bus.
479 * sd-bus vtable entries can have a new SD_BUS_VTABLE_ABSOLUTE_OFFSET
480 flag which alters how the userdata pointer to pass to the callbacks
481 is determined. When the flag is set, the offset field is converted
482 as-is into a pointer, without adding it to the object pointer the
483 vtable is associated with.
485 * sd-bus now exposes four new functions:
486 sd_bus_interface_name_is_valid() + sd_bus_service_name_is_valid() +
487 sd_bus_member_name_is_valid() + sd_bus_object_path_is_valid() will
488 validate strings to check if they qualify as various D-Bus concepts.
490 * The sd-bus API gained the SD_BUS_METHOD_WITH_ARGS(),
491 SD_BUS_METHOD_WITH_ARGS_OFFSET() and SD_BUS_SIGNAL_WITH_ARGS() macros
492 that simplify adding argument names to D-Bus methods and signals.
494 * The man pages for the sd-bus and sd-hwdb APIs have been completed.
496 * Various D-Bus APIs of systemd daemons now have man pages that
497 document the methods, signals and properties.
499 * The expectations on user/group name syntax are now documented in
500 detail; documentation how classic home directories may be converted
501 into home directories managed by homed has been added; documentation
502 regarding integration of homed/userdb functionality in desktops has
505 https://systemd.io/USER_NAMES
506 https://systemd.io/CONVERTING_TO_HOMED
507 https://systemd.io/USERDB_AND_DESKTOPS
509 * Documentation for the on-disk Journal file format has been updated
510 and has now moved to:
512 https://systemd.io/JOURNAL_FILE_FORMAT
514 * The interface for containers (https://systemd.io/CONTAINER_INTERFACE)
515 has been extended by a set of environment variables that expose
516 select fields from the host's os-release file to the container
517 payload. Similarly, host's os-release files can be mounted into the
518 container underneath /run/hosts. Together, those mechanisms provide a
519 standardized way to expose information about the host to the
520 container payload. Both interfaces are implemented in systemd-nspawn.
522 * All D-Bus services shipped in systemd now implement the generic
523 LogControl1 D-Bus API which allows clients to change log level +
524 target of the service during runtime.
526 Contributions from: 24bisquitz, Adam Nielsen, Alan Perry, Alexander
527 Malafeev, Alin Popa, Amos Bird, Andreas Rammhold, AndreRH, Andrew
528 Doran, Anita Zhang, Ankit Jain, antznin, Arnaud Ferraris, Arthur Moraes
529 do Lago, Arusekk, Balaji Punnuru, Balint Reczey, Bastien Nocera,
530 bemarek, Benjamin Berg, Benjamin Dahlhoff, Benjamin Robin, Chris Down,
531 Chris Kerr, Christian Göttsche, Christian Hesse, Christian Oder,
532 Ciprian Hacman, codicodi, Corey Hinshaw, Daan De Meyer, Dana Olson, Dan
533 Callaghan, Daniel Fullmer, Daniel Rusek, Dan Streetman, Dave Reisner,
534 David Edmundson, David Wood, Denis Pronin, Diego Escalante Urrelo,
535 Dimitri John Ledkov, dolphrundgren, duguxy, Einsler Lee, Elisei Roca,
536 Emmanuel Garette, Eric Anderson, Eric DeVolder, Evgeny Vereshchagin,
537 ExtinctFire, fangxiuning, Ferran Pallarès Roca, Filipe Brandenburger,
538 Finn, Florian Klink, Franck Bui, Frantisek Sumsal, Gaoyi, gaurav, Georg
539 Müller, Gergely Polonkai, Giedrius Statkevičius, Gigadoc2, gogogogi,
540 gzjsgdsb, Hans de Goede, Haochen Tong, ianhi, ignapk, Jakov Smolic,
541 James T. Lee, Jan Janssen, Jan Klötzke, Jan Palus, Jay Burger, Jeremy
542 Cline, Jérémy Rosen, Jian-Hong Pan, Jiri Slaby, Joel Shapiro, Joerg
543 Behrmann, Jörg Thalheim, Jouke Witteveen, Kai-Heng Feng, Kenny
544 Levinsen, Kevin Kuehler, Kumar Kartikeya Dwivedi, layderv, laydervus,
545 Lénaïc Huard, Lennart Poettering, Lidong Zhong, Luca Boccassi, Luca
546 BRUNO, Lucas Werkmeister, Lukas Klingsbo, Lukáš Nykrýn, Łukasz
547 Stelmach, Maciej S. Szmigiero, MadMcCrow, Marc-André Lureau, Marcel
548 Holtmann, Marc Kleine-Budde, Martin Hundebøll, Matthew Leeds, Matt
549 Ranostay, Maxim Fomin, MaxVerevkin, Michael Biebl, Michael Chapman,
550 Michael Gubbels, Michael Marley, Michał Bartoszkiewicz, Michal Koutný,
551 Michal Sekletar, Michal Sekletár, Mike Gilbert, Mike Kazantsev, ml,
552 Motiejus Jakštys, nabijaczleweli, nerdopolis, Niccolò Maggioni, Niklas
553 Hambüchen, Norbert Lange, Paul Cercueil, pelzvieh, Peter Hutterer,
554 Piero La Terza, Pieter Lexis, Piotr Drąg, Rafael Fontenelle, Richard
555 Petri, Ronan Pigott, Ross Lagerwall, Rubens Figueiredo, satmandu,
556 Sean-StarLabs, Sebastian Jennen, sterlinghughes, Susant Sahani, Thomas
557 Haller, Tobias Hunger, Tom, Tomáš Pospíšek, Tomer Shechner, Tom Hughes,
558 Topi Miettinen, Tudor Roman, Uwe Kleine-König, Valery0xff, Vito Caputo,
559 Vladimir Panteleev, Vladyslav Tronko, Wen Yang, Yegor Vialov, Yigal
560 Korman, YmrDtnJu, Yuri Chornoivan, Yu Watanabe, Zbigniew
561 Jędrzejewski-Szmek, Zhu Li, Дамјан Георгиевски, наб
567 * A new tool "systemd-repart" has been added, that operates as an
568 idempotent declarative repartitioner for GPT partition tables.
569 Specifically, a set of partitions that must or may exist can be
570 configured via drop-in files, and during every boot the partition
571 table on disk is compared with these files, creating missing
572 partitions or growing existing ones based on configurable relative
573 and absolute size constraints. The tool is strictly incremental,
574 i.e. does not delete, shrink or move partitions, but only adds and
575 grows them. The primary use-case is OS images that ship in minimized
576 form, that on first boot are grown to the size of the underlying
577 block device or augmented with additional partitions. For example,
578 the root partition could be extended to cover the whole disk, or a
579 swap or /home partitions could be added on first boot. It can also be
580 used for systems that use an A/B update scheme but ship images with
581 just the A partition, with B added on first boot. The tool is
582 primarily intended to be run in the initrd, shortly before
583 transitioning into the host OS, but can also be run after the
584 transition took place. It automatically discovers the disk backing
585 the root file system, and should hence not require any additional
586 configuration besides the partition definition drop-ins. If no
587 configuration drop-ins are present, no action is taken.
589 * A new component "userdb" has been added, along with a small daemon
590 "systemd-userdb.service" and a client tool "userdbctl". The framework
591 allows defining rich user and group records in a JSON format,
592 extending on the classic "struct passwd" and "struct group"
593 structures. Various components in systemd have been updated to
594 process records in this format, including systemd-logind and
595 pam-systemd. The user records are intended to be extensible, and
596 allow setting various resource management, security and runtime
597 parameters that shall be applied to processes and sessions of the
598 user as they log in. This facility is intended to allow associating
599 such metadata directly with user/group records so that they can be
600 produced, extended and consumed in unified form. We hope that
601 eventually frameworks such as sssd will generate records this way, so
602 that for the first time resource management and various other
603 per-user settings can be configured in LDAP directories and then
604 provided to systemd (specifically to systemd-logind and pam-system)
605 to apply on login. For further details see:
607 https://systemd.io/USER_RECORD
608 https://systemd.io/GROUP_RECORD
609 https://systemd.io/USER_GROUP_API
611 * A small new service systemd-homed.service has been added, that may be
612 used to securely manage home directories with built-in encryption.
613 The complete user record data is unified with the home directory,
614 thus making home directories naturally migratable. Its primary
615 back-end is based on LUKS volumes, but fscrypt, plain directories,
616 and other storage schemes are also supported. This solves a couple of
617 problems we saw with traditional ways to manage home directories, in
618 particular when it comes to encryption. For further discussion of
619 this, see the video of Lennart's talk at AllSystemsGo! 2019:
621 https://media.ccc.de/v/ASG2019-164-reinventing-home-directories
623 For further details about the format and expectations on home
624 directories this new daemon makes, see:
626 https://systemd.io/HOME_DIRECTORY
628 * systemd-journald is now multi-instantiable. In addition to the main
629 instance systemd-journald.service there's now a template unit
630 systemd-journald@.service, with each instance defining a new named
631 log 'namespace' (whose name is specified via the instance part of the
632 unit name). A new unit file setting LogNamespace= has been added,
633 taking such a namespace name, that assigns services to the specified
634 log namespaces. As each log namespace is serviced by its own
635 independent journal daemon, this functionality may be used to improve
636 performance and increase isolation of applications, at the price of
637 losing global message ordering. Each instance of journald has a
638 separate set of configuration files, with possibly different disk
639 usage limitations and other settings.
641 journalctl now takes a new option --namespace= to show logs from a
642 specific log namespace. The sd-journal.h API gained
643 sd_journal_open_namespace() for opening the log stream of a specific
644 log namespace. systemd-journald also gained the ability to exit on
645 idle, which is useful in the context of log namespaces, as this means
646 log daemons for log namespaces can be activated automatically on
647 demand and will stop automatically when no longer used, minimizing
650 * When systemd-tmpfiles copies a file tree using the 'C' line type it
651 will now label every copied file according to the SELinux database.
653 * When systemd/PID 1 detects it is used in the initrd it will now boot
654 into initrd.target rather than default.target by default. This should
655 make it simpler to build initrds with systemd as for many cases the
656 only difference between a host OS image and an initrd image now is
657 the presence of the /etc/initrd-release file.
659 * A new kernel command line option systemd.cpu_affinity= is now
660 understood. It's equivalent to the CPUAffinity= option in
661 /etc/systemd/system.conf and allows setting the CPU mask for PID 1
662 itself and the default for all other processes.
664 * When systemd/PID 1 is reloaded (with systemctl daemon-reload or
665 equivalent), the SELinux database is now reloaded, ensuring that
666 sockets and other file system objects are generated taking the new
667 database into account.
669 * systemd/PID 1 accepts a new "systemd.show-status=error" setting, and
670 "quiet" has been changed to imply that instead of
671 "systemd.show-status=auto". In this mode, only messages about errors
672 and significant delays in boot are shown on the console.
674 * The sd-event.h API gained native support for the new Linux "pidfd"
675 concept. This permits watching processes using file descriptors
676 instead of PID numbers, which fixes a number of races and makes
677 process supervision more robust and efficient. All of systemd's
678 components will now use pidfds if the kernel supports it for process
679 watching, with the exception of PID 1 itself, unfortunately. We hope
680 to move PID 1 to exclusively using pidfds too eventually, but this
681 requires some more kernel work first. (Background: PID 1 watches
682 processes using waitid() with the P_ALL flag, and that does not play
683 together nicely with pidfds yet.)
685 * Closely related to this, the sd-event.h API gained two new calls
686 sd_event_source_send_child_signal() (for sending a signal to a
687 watched process) and sd_event_source_get_child_process_own() (for
688 marking a process so that it is killed automatically whenever the
689 event source watching it is freed).
691 * systemd-networkd gained support for configuring Token Bucket Filter
692 (TBF) parameters in its qdisc configuration support. Similarly,
693 support for Stochastic Fairness Queuing (SFQ), Controlled-Delay
694 Active Queue Management (CoDel), and Fair Queue (FQ) has been added.
696 * systemd-networkd gained support for Intermediate Functional Block
697 (IFB) network devices.
699 * systemd-networkd gained support for configuring multi-path IP routes,
700 using the new MultiPathRoute= setting in the [Route] section.
702 * systemd-networkd's DHCPv4 client has been updated to support a new
703 SendDecline= option. If enabled, duplicate address detection is done
704 after a DHCP offer is received from the server. If a conflict is
705 detected, the address is declined. The DHCPv4 client also gained
706 support for a new RouteMTUBytes= setting that allows to configure the
707 MTU size to be used for routes generated from DHCPv4 leases.
709 * The PrefixRoute= setting in systemd-networkd's [Address] section of
710 .network files has been deprecated, and replaced by AddPrefixRoute=,
711 with its sense inverted.
713 * The Gateway= setting of [Route] sections of .network files gained
714 support for a special new value "_dhcp". If set, the configured
715 static route uses the gateway host configured via DHCP.
717 * New User= and SuppressPrefixLength= settings have been implemented
718 for the [RoutingPolicyRule] section of .network files to configure
719 source routing based on UID ranges and prefix length, respectively.
721 * sd-bus gained a new API call sd_bus_message_sensitive() that marks a
722 D-Bus message object as "sensitive". Those objects are erased from
723 memory when they are freed. This concept is intended to be used for
724 messages that contain security sensitive data. A new flag
725 SD_BUS_VTABLE_SENSITIVE has been introduced as well to mark methods
726 in sd-bus vtables, causing any incoming and outgoing messages of
727 those methods to be implicitly marked as "sensitive".
729 * sd-bus gained a new API call sd_bus_message_dump() for dumping the
730 contents of a message (or parts thereof) to standard output for
733 * systemd-sysusers gained support for creating users with the primary
734 group named differently than the user.
736 * systemd-growfs (i.e. the x-systemd.growfs mount option in /etc/fstab)
737 gained support for growing XFS partitions. Previously it supported
738 only ext4 and btrfs partitions.
740 * The support for /etc/crypttab gained a new x-initrd.attach option. If
741 set, the specified encrypted volume is unlocked already in the
742 initrd. This concept corresponds to the x-initrd.mount option in
745 * systemd-cryptsetup gained native support for unlocking encrypted
746 volumes utilizing PKCS#11 smartcards, i.e. for example to bind
747 encryption of volumes to YubiKeys. This is exposed in the new
748 pkcs11-uri= option in /etc/crypttab.
750 * The /etc/fstab support in systemd now supports two new mount options
751 x-systemd.{required,wanted}-by=, for explicitly configuring the units
752 that the specified mount shall be pulled in by, in place of
753 the usual local-fs.target/remote-fs.target.
755 * The https://systemd.io/ web site has been relaunched, directly
756 populated with most of the documentation included in the systemd
757 repository. systemd also acquired a new logo, thanks to Tobias
760 * systemd-udevd gained support for managing "alternative" network
761 interface names, as supported by new Linux kernels. For the first
762 time this permits assigning multiple (and longer!) names to a network
763 interface. systemd-udevd will now by default assign the names
764 generated via all supported naming schemes to each interface. This
765 may be further tweaked with .link files and the AlternativeName= and
766 AlternativeNamesPolicy= settings. Other components of systemd have
767 been updated to support the new alternative names wherever
768 appropriate. For example, systemd-nspawn will now generate
769 alternative interface names for the host-facing side of container
770 veth links based on the full container name without truncation.
772 * systemd-nspawn interface naming logic has been updated in another way
773 too: if the main interface name (i.e. as opposed to new-style
774 "alternative" names) based on the container name is truncated, a
775 simple hashing scheme is used to give different interface names to
776 multiple containers whose names all begin with the same prefix. Since
777 this changes the primary interface names pointing to containers if
778 truncation happens, the old scheme may still be requested by
779 selecting an older naming scheme, via the net.naming-scheme= kernel
782 * PrivateUsers= in service files now works in services run by the
783 systemd --user per-user instance of the service manager.
785 * A new per-service sandboxing option ProtectClock= has been added that
786 locks down write access to the system clock. It takes away device
787 node access to /dev/rtc as well as the system calls that set the
788 system clock and the CAP_SYS_TIME and CAP_WAKE_ALARM capabilities.
789 Note that this option does not affect access to auxiliary services
790 that allow changing the clock, for example access to
793 * The systemd-id128 tool gained a new "show" verb for listing or
794 resolving a number of well-known UUIDs/128bit IDs, currently mostly
795 GPT partition table types.
797 * The Discoverable Partitions Specification has been updated to support
798 /var and /var/tmp partition discovery. Support for this has been
799 added to systemd-gpt-auto-generator. For details see:
801 https://systemd.io/DISCOVERABLE_PARTITIONS
803 * "systemctl list-unit-files" has been updated to show a new column
804 with the suggested enablement state based on the vendor preset files
805 for the respective units.
807 * "systemctl" gained a new option "--with-dependencies". If specified
808 commands such as "systemctl status" or "systemctl cat" will now show
809 all specified units along with all units they depend on.
811 * networkctl gained support for showing per-interface logs in its
814 * systemd-networkd-wait-online gained support for specifying the maximum
815 operational state to wait for, and to wait for interfaces to
818 * The [Match] section of .link and .network files now supports a new
819 option PermanentMACAddress= which may be used to check against the
820 permanent MAC address of a network device even if a randomized MAC
823 * The [TrafficControlQueueingDiscipline] section in .network files has
824 been renamed to [NetworkEmulator] with the "NetworkEmulator" prefix
825 dropped from the individual setting names.
827 * Any .link and .network files that have an empty [Match] section (this
828 also includes empty and commented-out files) will now be
829 rejected. systemd-udev and systemd-networkd started warning about
830 such files in version 243.
832 * systemd-logind will now validate access to the operation of changing
833 the virtual terminal via a polkit action. By default, only users
834 with at least one session on a local VT are granted permission.
836 * When systemd sets up PAM sessions that invoked service processes
837 shall run in, the pam_setcred() API is now invoked, thus permitting
838 PAM modules to set additional credentials for the processes.
840 * portablectl attach/detach verbs now accept --now and --enable options
841 to combine attachment with enablement and invocation, or detachment
842 with stopping and disablement.
844 * UPGRADE ISSUE: a bug where some jobs were trimmed as redundant was
845 fixed, which in turn exposed bugs in unit configuration of services
846 which have Type=oneshot and should only run once, but do not have
847 RemainAfterExit=yes set. Without RemainAfterExit=yes, a one-shot
848 service may be started again after exiting successfully, for example
849 as a dependency in another transaction. Affected services included
850 some internal systemd services (most notably
851 systemd-vconsole-setup.service, which was updated to have
852 RemainAfterExit=yes), and plymouth-start.service. Please ensure that
853 plymouth has been suitably updated or patched before upgrading to
854 this systemd release. See
855 https://bugzilla.redhat.com/show_bug.cgi?id=1807771 for some
856 additional discussion.
858 Contributions from: AJ Bagwell, Alin Popa, Andreas Rammhold, Anita
859 Zhang, Ansgar Burchardt, Antonio Russo, Arian van Putten, Ashley Davis,
860 Balint Reczey, Bart Willems, Bastien Nocera, Benjamin Dahlhoff, Charles
861 (Chas) Williams, cheese1, Chris Down, Chris Murphy, Christian Ehrhardt,
862 Christian Göttsche, cvoinf, Daan De Meyer, Daniele Medri, Daniel Rusek,
863 Daniel Shahaf, Dann Frazier, Dan Streetman, Dariusz Gadomski, David
864 Michael, Dimitri John Ledkov, Emmanuel Bourg, Evgeny Vereshchagin,
865 ezst036, Felipe Sateler, Filipe Brandenburger, Florian Klink, Franck
866 Bui, Fran Dieguez, Frantisek Sumsal, Greg "GothAck" Miell, Guilhem
867 Lettron, Guillaume Douézan-Grard, Hans de Goede, HATAYAMA Daisuke, Iain
868 Lane, James Buren, Jan Alexander Steffens (heftig), Jérémy Rosen, Jin
869 Park, Jun'ichi Nomura, Kai Krakow, Kevin Kuehler, Kevin P. Fleming,
870 Lennart Poettering, Leonid Bloch, Leonid Evdokimov, lothrond, Luca
871 Boccassi, Lukas K, Lynn Kirby, Mario Limonciello, Mark Deneen, Matthew
872 Leeds, Michael Biebl, Michal Koutný, Michal Sekletár, Mike Auty, Mike
873 Gilbert, mtron, nabijaczleweli, Naïm Favier, Nate Jones, Norbert Lange,
874 Oliver Giles, Paul Davey, Paul Menzel, Peter Hutterer, Piotr Drąg, Rafa
875 Couto, Raphael, rhn, Robert Scheck, Rocka, Romain Naour, Ryan Attard,
876 Sascha Dewald, Shengjing Zhu, Slava Kardakov, Spencer Michaels, Sylvain
877 Plantefeve, Stanislav Angelovič, Susant Sahani, Thomas Haller, Thomas
878 Schmitt, Timo Schlüßler, Timo Wilken, Tobias Bernard, Tobias Klauser,
879 Tobias Stoeckmann, Topi Miettinen, tsia, WataruMatsuoka, Wieland
880 Hoffmann, Wilhelm Schuster, Will Fleming, xduugu, Yong Cong Sin, Yuri
881 Chornoivan, Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zeyu
888 * Support for the cpuset cgroups v2 controller has been added.
889 Processes may be restricted to specific CPUs using the new
890 AllowedCPUs= setting, and to specific memory NUMA nodes using the new
891 AllowedMemoryNodes= setting.
893 * The signal used in restart jobs (as opposed to e.g. stop jobs) may
894 now be configured using a new RestartKillSignal= setting. This
895 allows units which signals to request termination to implement
896 different behaviour when stopping in preparation for a restart.
898 * "systemctl clean" may now be used also for socket, mount, and swap
901 * systemd will also read configuration options from the EFI variable
902 SystemdOptions. This may be used to configure systemd behaviour when
903 modifying the kernel command line is inconvenient, but configuration
904 on disk is read too late, for example for the options related to
905 cgroup hierarchy setup. 'bootctl systemd-efi-options' may be used to
906 set the EFI variable.
908 * systemd will now disable printk ratelimits in early boot. This should
909 allow us to capture more logs from the early boot phase where normal
910 storage is not available and the kernel ring buffer is used for
911 logging. Configuration on the kernel command line has higher priority
912 and overrides the systemd setting.
914 systemd programs which log to /dev/kmsg directly use internal
915 ratelimits to prevent runaway logging. (Normally this is only used
916 during early boot, so in practice this change has very little
919 * Unit files now support top level dropin directories of the form
920 <unit_type>.d/ (e.g. service.d/) that may be used to add configuration
921 that affects all corresponding unit files.
923 * systemctl gained support for 'stop --job-mode=triggering' which will
924 stop the specified unit and any units which could trigger it.
926 * Unit status display now includes units triggering and triggered by
927 the unit being shown.
929 * The RuntimeMaxSec= setting is now supported by scopes, not just
930 .service units. This is particularly useful for PAM sessions which
931 create a scope unit for the user login. systemd.runtime_max_sec=
932 setting may used with the pam_systemd module to limit the duration
933 of the PAM session, for example for time-limited logins.
935 * A new @pkey system call group is now defined to make it easier to
936 allow-list memory protection syscalls for containers and services
937 which need to use them.
939 * systemd-udevd: removed the 30s timeout for killing stale workers on
940 exit. systemd-udevd now waits for workers to finish. The hard-coded
941 exit timeout of 30s was too short for some large installations, where
942 driver initialization could be prematurely interrupted during initrd
943 processing if the root file system had been mounted and init was
944 preparing to switch root. If udevd is run without systemd and workers
945 are hanging while udevd receives an exit signal, udevd will now exit
946 when udev.event_timeout is reached for the last hanging worker. With
947 systemd, the exit timeout can additionally be configured using
948 TimeoutStopSec= in systemd-udevd.service.
950 * udev now provides a program (fido_id) that identifies FIDO CTAP1
951 ("U2F")/CTAP2 security tokens based on the usage declared in their
952 report and descriptor and outputs suitable environment variables.
953 This replaces the externally maintained allow lists of all known
954 security tokens that were used previously.
956 * Automatically generated autosuspend udev rules for allow-listed
957 devices have been imported from the Chromium OS project. This should
958 improve power saving with many more devices.
960 * udev gained a new "CONST{key}=value" setting that allows matching
961 against system-wide constants without forking a helper binary.
962 Currently "arch" and "virt" keys are supported.
964 * udev now opens CDROMs in non-exclusive mode when querying their
965 capabilities. This should fix issues where other programs trying to
966 use the CDROM cannot gain access to it, but carries a risk of
967 interfering with programs writing to the disk, if they did not open
968 the device in exclusive mode as they should.
970 * systemd-networkd does not create a default route for IPv4 link local
971 addressing anymore. The creation of the route was unexpected and was
972 breaking routing in various cases, but people who rely on it being
973 created implicitly will need to adjust. Such a route may be requested
974 with DefaultRouteOnDevice=yes.
976 Similarly, systemd-networkd will not assign a link-local IPv6 address
977 when IPv6 link-local routing is not enabled.
979 * Receive and transmit buffers may now be configured on links with
980 the new RxBufferSize= and TxBufferSize= settings.
982 * systemd-networkd may now advertise additional IPv6 routes. A new
983 [IPv6RoutePrefix] section with Route= and LifetimeSec= options is
986 * systemd-networkd may now configure "next hop" routes using the
987 [NextHop] section and Gateway= and Id= settings.
989 * systemd-networkd will now retain DHCP config on restarts by default
990 (but this may be overridden using the KeepConfiguration= setting).
991 The default for SendRelease= has been changed to true.
993 * The DHCPv4 client now uses the OPTION_INFORMATION_REFRESH_TIME option
994 received from the server.
996 The client will use the received SIP server list if UseSIP=yes is
999 The client may be configured to request specific options from the
1000 server using a new RequestOptions= setting.
1002 The client may be configured to send arbitrary options to the server
1003 using a new SendOption= setting.
1005 A new IPServiceType= setting has been added to configure the "IP
1006 service type" value used by the client.
1008 * The DHCPv6 client learnt a new PrefixDelegationHint= option to
1009 request prefix hints in the DHCPv6 solicitation.
1011 * The DHCPv4 server may be configured to send arbitrary options using
1012 a new SendOption= setting.
1014 * The DHCPv4 server may now be configured to emit SIP server list using
1015 the new EmitSIP= and SIP= settings.
1017 * systemd-networkd and networkctl may now renew DHCP leases on demand.
1018 networkctl has a new 'networkctl renew' verb.
1020 * systemd-networkd may now reconfigure links on demand. networkctl
1021 gained two new verbs: "reload" will reload the configuration, and
1022 "reconfigure DEVICE…" will reconfigure one or more devices.
1024 * .network files may now match on SSID and BSSID of a wireless network,
1025 i.e. the access point name and hardware address using the new SSID=
1026 and BSSID= options. networkctl will display the current SSID and
1027 BSSID for wireless links.
1029 .network files may also match on the wireless network type using the
1030 new WLANInterfaceType= option.
1032 * systemd-networkd now includes default configuration that enables
1033 link-local addressing when connected to an ad-hoc wireless network.
1035 * systemd-networkd may configure the Traffic Control queueing
1036 disciplines in the kernel using the new
1037 [TrafficControlQueueingDiscipline] section and Parent=,
1038 NetworkEmulatorDelaySec=, NetworkEmulatorDelayJitterSec=,
1039 NetworkEmulatorPacketLimit=, NetworkEmulatorLossRate=,
1040 NetworkEmulatorDuplicateRate= settings.
1042 * systemd-tmpfiles gained a new w+ setting to append to files.
1044 * systemd-analyze dump will now report when the memory configuration in
1045 the kernel does not match what systemd has configured (usually,
1046 because some external program has modified the kernel configuration
1049 * systemd-analyze gained a new --base-time= switch instructs the
1050 'calendar' verb to resolve times relative to that timestamp instead
1051 of the present time.
1053 * journalctl --update-catalog now produces deterministic output (making
1054 reproducible image builds easier).
1056 * A new devicetree-overlay setting is now documented in the Boot Loader
1059 * The default value of the WatchdogSec= setting used in systemd
1060 services (the ones bundled with the project itself) may be set at
1061 configuration time using the -Dservice-watchdog= setting. If set to
1062 empty, the watchdogs will be disabled.
1064 * systemd-resolved validates IP addresses in certificates now when GnuTLS
1067 * libcryptsetup >= 2.0.1 is now required.
1069 * A configuration option -Duser-path= may be used to override the $PATH
1070 used by the user service manager. The default is again to use the same
1071 path as the system manager.
1073 * The systemd-id128 tool gained a new switch "-u" (or "--uuid") for
1074 outputting the 128bit IDs in UUID format (i.e. in the "canonical
1077 * Service units gained a new sandboxing option ProtectKernelLogs= which
1078 makes sure the program cannot get direct access to the kernel log
1079 buffer anymore, i.e. the syslog() system call (not to be confused
1080 with the API of the same name in libc, which is not affected), the
1081 /proc/kmsg and /dev/kmsg nodes and the CAP_SYSLOG capability are made
1082 inaccessible to the service. It's recommended to enable this setting
1083 for all services that should not be able to read from or write to the
1084 kernel log buffer, which are probably almost all.
1086 Contributions from: Aaron Plattner, Alcaro, Anita Zhang, Balint Reczey,
1087 Bastien Nocera, Baybal Ni, Benjamin Bouvier, Benjamin Gilbert, Carlo
1088 Teubner, cbzxt, Chen Qi, Chris Down, Christian Rebischke, Claudio
1089 Zumbo, ClydeByrdIII, crashfistfight, Cyprien Laplace, Daniel Edgecumbe,
1090 Daniel Gorbea, Daniel Rusek, Daniel Stuart, Dan Streetman, David
1091 Pedersen, David Tardon, Dimitri John Ledkov, Dominique Martinet, Donald
1092 A. Cupp Jr, Evgeny Vereshchagin, Fabian Henneke, Filipe Brandenburger,
1093 Franck Bui, Frantisek Sumsal, Georg Müller, Hans de Goede, Haochen
1094 Tong, HATAYAMA Daisuke, Iwan Timmer, Jan Janssen, Jan Kundrát, Jan
1095 Synacek, Jan Tojnar, Jay Strict, Jérémy Rosen, Jóhann B. Guðmundsson,
1096 Jonas Jelten, Jonas Thelemann, Justin Trudell, J. Xing, Kai-Heng Feng,
1097 Kenneth D'souza, Kevin Becker, Kevin Kuehler, Lennart Poettering,
1098 Léonard Gérard, Lorenz Bauer, Luca Boccassi, Maciej Stanczew, Mario
1099 Limonciello, Marko Myllynen, Mark Stosberg, Martin Wilck, matthiasroos,
1100 Michael Biebl, Michael Olbrich, Michael Tretter, Michal Sekletar,
1101 Michal Sekletár, Michal Suchanek, Mike Gilbert, Mike Kazantsev, Nicolas
1102 Douma, nikolas, Norbert Lange, pan93412, Pascal de Bruijn, Paul Menzel,
1103 Pavel Hrdina, Peter Wu, Philip Withnall, Piotr Drąg, Rafael Fontenelle,
1104 Renaud Métrich, Riccardo Schirone, RoadrunnerWMC, Ronan Pigott, Ryan
1105 Attard, Sebastian Wick, Serge, Siddharth Chandrasekara, Steve Ramage,
1106 Steve Traylen, Susant Sahani, Thibault Nélis, Tim Teichmann, Tom
1107 Fitzhenry, Tommy J, Torsten Hilbrich, Vito Caputo, ypf791, Yu Watanabe,
1108 Zach Smith, Zbigniew Jędrzejewski-Szmek
1110 – Warsaw, 2019-11-29
1114 * This release enables unprivileged programs (i.e. requiring neither
1115 setuid nor file capabilities) to send ICMP Echo (i.e. ping) requests
1116 by turning on the "net.ipv4.ping_group_range" sysctl of the Linux
1117 kernel for the whole UNIX group range, i.e. all processes. This
1118 change should be reasonably safe, as the kernel support for it was
1119 specifically implemented to allow safe access to ICMP Echo for
1120 processes lacking any privileges. If this is not desirable, it can be
1121 disabled again by setting the parameter to "1 0".
1123 * Previously, filters defined with SystemCallFilter= would have the
1124 effect that any calling of an offending system call would terminate
1125 the calling thread. This behaviour never made much sense, since
1126 killing individual threads of unsuspecting processes is likely to
1127 create more problems than it solves. With this release the default
1128 action changed from killing the thread to killing the whole
1129 process. For this to work correctly both a kernel version (>= 4.14)
1130 and a libseccomp version (>= 2.4.0) supporting this new seccomp
1131 action is required. If an older kernel or libseccomp is used the old
1132 behaviour continues to be used. This change does not affect any
1133 services that have no system call filters defined, or that use
1134 SystemCallErrorNumber= (and thus see EPERM or another error instead
1135 of being killed when calling an offending system call). Note that
1136 systemd documentation always claimed that the whole process is
1137 killed. With this change behaviour is thus adjusted to match the
1140 * On 64 bit systems, the "kernel.pid_max" sysctl is now bumped to
1141 4194304 by default, i.e. the full 22bit range the kernel allows, up
1142 from the old 16bit range. This should improve security and
1143 robustness, as PID collisions are made less likely (though certainly
1144 still possible). There are rumours this might create compatibility
1145 problems, though at this moment no practical ones are known to
1146 us. Downstream distributions are hence advised to undo this change in
1147 their builds if they are concerned about maximum compatibility, but
1148 for everybody else we recommend leaving the value bumped. Besides
1149 improving security and robustness this should also simplify things as
1150 the maximum number of allowed concurrent tasks was previously bounded
1151 by both "kernel.pid_max" and "kernel.threads-max" and now effectively
1152 only a single knob is left ("kernel.threads-max"). There have been
1153 concerns that usability is affected by this change because larger PID
1154 numbers are harder to type, but we believe the change from 5 digits
1155 to 7 digits doesn't hamper usability.
1157 * MemoryLow= and MemoryMin= gained hierarchy-aware counterparts,
1158 DefaultMemoryLow= and DefaultMemoryMin=, which can be used to
1159 hierarchically set default memory protection values for a particular
1160 subtree of the unit hierarchy.
1162 * Memory protection directives can now take a value of zero, allowing
1163 explicit opting out of a default value propagated by an ancestor.
1165 * systemd now defaults to the "unified" cgroup hierarchy setup during
1166 build-time, i.e. -Ddefault-hierarchy=unified is now the build-time
1167 default. Previously, -Ddefault-hierarchy=hybrid was the default. This
1168 change reflects the fact that cgroupsv2 support has matured
1169 substantially in both systemd and in the kernel, and is clearly the
1170 way forward. Downstream production distributions might want to
1171 continue to use -Ddefault-hierarchy=hybrid (or even =legacy) for
1172 their builds as unfortunately the popular container managers have not
1173 caught up with the kernel API changes.
1175 * Man pages are not built by default anymore (html pages were already
1176 disabled by default), to make development builds quicker. When
1177 building systemd for a full installation with documentation, meson
1178 should be called with -Dman=true and/or -Dhtml=true as appropriate.
1179 The default was changed based on the assumption that quick one-off or
1180 repeated development builds are much more common than full optimized
1181 builds for installation, and people need to pass various other
1182 options to when doing "proper" builds anyway, so the gain from making
1183 development builds quicker is bigger than the one time disruption for
1186 Two scripts are created in the *build* directory to generate and
1187 preview man and html pages on demand, e.g.:
1189 build/man/man systemctl
1190 build/man/html systemd.index
1192 * libidn2 is used by default if both libidn2 and libidn are installed.
1193 Please use -Dlibidn=true if libidn is preferred.
1195 * The D-Bus "wire format" of the CPUAffinity= attribute is changed on
1196 big-endian machines. Before, bytes were written and read in native
1197 machine order as exposed by the native libc __cpu_mask interface.
1198 Now, little-endian order is always used (CPUs 0–7 are described by
1199 bits 0–7 in byte 0, CPUs 8–15 are described by byte 1, and so on).
1200 This change fixes D-Bus calls that cross endianness boundary.
1202 The presentation format used for CPUAffinity= by "systemctl show" and
1203 "systemd-analyze dump" is changed to present CPU indices instead of
1204 the raw __cpu_mask bitmask. For example, CPUAffinity=0-1 would be
1205 shown as CPUAffinity=03000000000000000000000000000… (on
1206 little-endian) or CPUAffinity=00000000000000300000000000000… (on
1207 64-bit big-endian), and is now shown as CPUAffinity=0-1, matching the
1208 input format. The maximum integer that will be printed in the new
1209 format is 8191 (four digits), while the old format always used a very
1210 long number (with the length varying by architecture), so they can be
1211 unambiguously distinguished.
1213 * /usr/sbin/halt.local is no longer supported. Implementation in
1214 distributions was inconsistent and it seems this functionality was
1217 To replace this functionality, users should:
1218 - either define a new unit and make it a dependency of final.target
1219 (systemctl add-wants final.target my-halt-local.service)
1220 - or move the shutdown script to /usr/lib/systemd/system-shutdown/
1221 and ensure that it accepts "halt", "poweroff", "reboot", and
1222 "kexec" as an argument, see the description in systemd-shutdown(8).
1224 * When a [Match] section in .link or .network file is empty (contains
1225 no match patterns), a warning will be emitted. Please add any "match
1226 all" pattern instead, e.g. OriginalName=* or Name=* in case all
1227 interfaces should really be matched.
1229 * A new setting NUMAPolicy= may be used to set process memory
1230 allocation policy. This setting can be specified in
1231 /etc/systemd/system.conf and hence will set the default policy for
1232 PID1. The default policy can be overridden on a per-service
1233 basis. The related setting NUMAMask= is used to specify NUMA node
1234 mask that should be associated with the selected policy.
1236 * PID 1 will now listen to Out-Of-Memory (OOM) events the kernel
1237 generates when processes it manages are reaching their memory limits,
1238 and will place their units in a special state, and optionally kill or
1239 stop the whole unit.
1241 * The service manager will now expose bus properties for the IO
1242 resources used by units. This information is also shown in "systemctl
1243 status" now (for services that have IOAccounting=yes set). Moreover,
1244 the IO accounting data is included in the resource log message
1245 generated whenever a unit stops.
1247 * Units may now configure an explicit timeout to wait for when killed
1248 with SIGABRT, for example when a service watchdog is hit. Previously,
1249 the regular TimeoutStopSec= timeout was applied in this case too —
1250 now a separate timeout may be set using TimeoutAbortSec=.
1252 * Services may now send a special WATCHDOG=trigger message with
1253 sd_notify() to trigger an immediate "watchdog missed" event, and thus
1254 trigger service termination. This is useful both for testing watchdog
1255 handling, but also for defining error paths in services, that shall
1256 be handled the same way as watchdog events.
1258 * There are two new per-unit settings IPIngressFilterPath= and
1259 IPEgressFilterPath= which allow configuration of a BPF program
1260 (usually by specifying a path to a program uploaded to /sys/fs/bpf/)
1261 to apply to the IP packet ingress/egress path of all processes of a
1262 unit. This is useful to allow running systemd services with BPF
1263 programs set up externally.
1265 * systemctl gained a new "clean" verb for removing the state, cache,
1266 runtime or logs directories of a service while it is terminated. The
1267 new verb may also be used to remove the state maintained on disk for
1268 timer units that have Persistent= configured.
1270 * During the last phase of shutdown systemd will now automatically
1271 increase the log level configured in the "kernel.printk" sysctl so
1272 that any relevant loggable events happening during late shutdown are
1273 made visible. Previously, loggable events happening so late during
1274 shutdown were generally lost if the "kernel.printk" sysctl was set to
1275 high thresholds, as regular logging daemons are terminated at that
1276 time and thus nothing is written to disk.
1278 * If processes terminated during the last phase of shutdown do not exit
1279 quickly systemd will now show their names after a short time, to make
1280 debugging easier. After a longer timeout they are forcibly killed,
1283 * journalctl (and the other tools that display logs) will now highlight
1284 warnings in yellow (previously, both LOG_NOTICE and LOG_WARNING where
1285 shown in bright bold, now only LOG_NOTICE is). Moreover, audit logs
1286 are now shown in blue color, to separate them visually from regular
1287 logs. References to configuration files are now turned into clickable
1288 links on terminals that support that.
1290 * systemd-journald will now stop logging to /var/log/journal during
1291 shutdown when /var/ is on a separate mount, so that it can be
1292 unmounted safely during shutdown.
1294 * systemd-resolved gained support for a new 'strict' DNS-over-TLS mode.
1296 * systemd-resolved "Cache=" configuration option in resolved.conf has
1297 been extended to also accept the 'no-negative' value. Previously,
1298 only a boolean option was allowed (yes/no), having yes as the
1299 default. If this option is set to 'no-negative', negative answers are
1300 not cached while the old cache heuristics are used positive answers.
1301 The default remains unchanged.
1303 * The predictable naming scheme for network devices now supports
1304 generating predictable names for "netdevsim" devices.
1306 Moreover, the "en" prefix was dropped from the ID_NET_NAME_ONBOARD
1309 Those two changes form a new net.naming-policy-scheme= entry.
1310 Distributions which want to preserve naming stability may want to set
1311 the -Ddefault-net-naming-scheme= configuration option.
1313 * systemd-networkd now supports MACsec, nlmon, IPVTAP and Xfrm
1314 interfaces natively.
1316 * systemd-networkd's bridge FDB support now allows configuration of a
1317 destination address for each entry (Destination=), as well as the
1318 VXLAN VNI (VNI=), as well as an option to declare what an entry is
1319 associated with (AssociatedWith=).
1321 * systemd-networkd's DHCPv4 support now understands a new MaxAttempts=
1322 option for configuring the maximum number of DHCP lease requests. It
1323 also learnt a new BlackList= option for deny-listing DHCP servers (a
1324 similar setting has also been added to the IPv6 RA client), as well
1325 as a SendRelease= option for configuring whether to send a DHCP
1326 RELEASE message when terminating.
1328 * systemd-networkd's DHCPv4 and DHCPv6 stacks can now be configured
1329 separately in the [DHCPv4] and [DHCPv6] sections.
1331 * systemd-networkd's DHCP support will now optionally create an
1332 implicit host route to the DNS server specified in the DHCP lease, in
1333 addition to the routes listed explicitly in the lease. This should
1334 ensure that in multi-homed systems DNS traffic leaves the systems on
1335 the interface that acquired the DNS server information even if other
1336 routes such as default routes exist. This behaviour may be turned on
1337 with the new RoutesToDNS= option.
1339 * systemd-networkd's VXLAN support gained a new option
1340 GenericProtocolExtension= for enabling VXLAN Generic Protocol
1341 Extension support, as well as IPDoNotFragment= for setting the IP
1342 "Don't fragment" bit on outgoing packets. A similar option has been
1343 added to the GENEVE support.
1345 * In systemd-networkd's [Route] section you may now configure
1346 FastOpenNoCookie= for configuring per-route TCP fast-open support, as
1347 well as TTLPropagate= for configuring Label Switched Path (LSP) TTL
1348 propagation. The Type= setting now supports local, broadcast,
1349 anycast, multicast, any, xresolve routes, too.
1351 * systemd-networkd's [Network] section learnt a new option
1352 DefaultRouteOnDevice= for automatically configuring a default route
1353 onto the network device.
1355 * systemd-networkd's bridging support gained two new options ProxyARP=
1356 and ProxyARPWifi= for configuring proxy ARP behaviour as well as
1357 MulticastRouter= for configuring multicast routing behaviour. A new
1358 option MulticastIGMPVersion= may be used to change bridge's multicast
1359 Internet Group Management Protocol (IGMP) version.
1361 * systemd-networkd's FooOverUDP support gained the ability to configure
1362 local and peer IP addresses via Local= and Peer=. A new option
1363 PeerPort= may be used to configure the peer's IP port.
1365 * systemd-networkd's TUN support gained a new setting VnetHeader= for
1366 tweaking Generic Segment Offload support.
1368 * The address family for policy rules may be specified using the new
1369 Family= option in the [RoutingPolicyRule] section.
1371 * networkctl gained a new "delete" command for removing virtual network
1372 devices, as well as a new "--stats" switch for showing device
1375 * networkd.conf gained a new setting SpeedMeter= and
1376 SpeedMeterIntervalSec=, to measure bitrate of network interfaces. The
1377 measured speed may be shown by 'networkctl status'.
1379 * "networkctl status" now displays MTU and queue lengths, and more
1380 detailed information about VXLAN and bridge devices.
1382 * systemd-networkd's .network and .link files gained a new Property=
1383 setting in the [Match] section, to match against devices with
1384 specific udev properties.
1386 * systemd-networkd's tunnel support gained a new option
1387 AssignToLoopback= for selecting whether to use the loopback device
1388 "lo" as underlying device.
1390 * systemd-networkd's MACAddress= setting in the [Neighbor] section has
1391 been renamed to LinkLayerAddress=, and it now allows configuration of
1394 * systemd-networkd's handling of the kernel's disable_ipv6 sysctl is
1395 simplified: systemd-networkd will disable the sysctl (enable IPv6) if
1396 IPv6 configuration (static or DHCPv6) was found for a given
1397 interface. It will not touch the sysctl otherwise.
1399 * The order of entries is $PATH used by the user manager instance was
1400 changed to put bin/ entries before the corresponding sbin/ entries.
1401 It is recommended to not rely on this order, and only ever have one
1402 binary with a given name in the system paths under /usr.
1404 * A new tool systemd-network-generator has been added that may generate
1405 .network, .netdev and .link files from IP configuration specified on
1406 the kernel command line in the format used by Dracut.
1408 * The CriticalConnection= setting in .network files is now deprecated,
1409 and replaced by a new KeepConfiguration= setting which allows more
1410 detailed configuration of the IP configuration to keep in place.
1412 * systemd-analyze gained a few new verbs:
1414 - "systemd-analyze timestamp" parses and converts timestamps. This is
1415 similar to the existing "systemd-analyze calendar" command which
1416 does the same for recurring calendar events.
1418 - "systemd-analyze timespan" parses and converts timespans (i.e.
1419 durations as opposed to points in time).
1421 - "systemd-analyze condition" will parse and test ConditionXYZ=
1424 - "systemd-analyze exit-status" will parse and convert exit status
1425 codes to their names and back.
1427 - "systemd-analyze unit-files" will print a list of all unit
1428 file paths and unit aliases.
1430 * SuccessExitStatus=, RestartPreventExitStatus=, and
1431 RestartForceExitStatus= now accept exit status names (e.g. "DATAERR"
1432 is equivalent to "65"). Those exit status name mappings may be
1433 displayed with the systemd-analyze exit-status verb describe above.
1435 * systemd-logind now exposes a per-session SetBrightness() bus call,
1436 which may be used to securely change the brightness of a kernel
1437 brightness device, if it belongs to the session's seat. By using this
1438 call unprivileged clients can make changes to "backlight" and "leds"
1439 devices securely with strict requirements on session membership.
1440 Desktop environments may use this to generically make brightness
1441 changes to such devices without shipping private SUID binaries or
1442 udev rules for that purpose.
1444 * "udevadm info" gained a --wait-for-initialization switch to wait for
1445 a device to be initialized.
1447 * systemd-hibernate-resume-generator will now look for resumeflags= on
1448 the kernel command line, which is similar to rootflags= and may be
1449 used to configure device timeout for the hibernation device.
1451 * sd-event learnt a new API call sd_event_source_disable_unref() for
1452 disabling and unref'ing an event source in a single function. A
1453 related call sd_event_source_disable_unrefp() has been added for use
1454 with gcc's cleanup extension.
1456 * The sd-id128.h public API gained a new definition
1457 SD_ID128_UUID_FORMAT_STR for formatting a 128bit ID in UUID format
1460 * "busctl introspect" gained a new switch --xml-interface for dumping
1461 XML introspection data unmodified.
1463 * PID 1 may now show the unit name instead of the unit description
1464 string in its status output during boot. This may be configured in
1465 the StatusUnitFormat= setting in /etc/systemd/system.conf or the
1466 kernel command line option systemd.status_unit_format=.
1468 * PID 1 now understands a new option KExecWatchdogSec= in
1469 /etc/systemd/system.conf to set a watchdog timeout for kexec reboots.
1470 Previously watchdog functionality was only available for regular
1471 reboots. The new setting defaults to off, because we don't know in
1472 the general case if the watchdog will be reset after kexec (some
1473 drivers do reset it, but not all), and the new userspace might not be
1474 configured to handle the watchdog.
1476 Moreover, the old ShutdownWatchdogSec= setting has been renamed to
1477 RebootWatchdogSec= to more clearly communicate what it is about. The
1478 old name is still accepted for compatibility.
1480 * The systemd.debug_shell kernel command line option now optionally
1481 takes a tty name to spawn the debug shell on, which allows a
1482 different tty to be selected than the built-in default.
1484 * Service units gained a new ExecCondition= setting which will run
1485 before ExecStartPre= and either continue execution of the unit (for
1486 clean exit codes), stop execution without marking the unit failed
1487 (for exit codes 1 through 254), or stop execution and fail the unit
1488 (for exit code 255 or abnormal termination).
1490 * A new service systemd-pstore.service has been added that pulls data
1491 from /sys/fs/pstore/ and saves it to /var/lib/pstore for later
1494 * timedatectl gained new verbs for configuring per-interface NTP
1495 service configuration for systemd-timesyncd.
1497 * "localectl list-locales" won't list non-UTF-8 locales anymore. It's
1498 2019. (You can set non-UTF-8 locales though, if you know their name.)
1500 * If variable assignments in sysctl.d/ files are prefixed with "-" any
1501 failures to apply them are now ignored.
1503 * systemd-random-seed.service now optionally credits entropy when
1504 applying the seed to the system. Set $SYSTEMD_RANDOM_SEED_CREDIT to
1505 true for the service to enable this behaviour, but please consult the
1506 documentation first, since this comes with a couple of caveats.
1508 * systemd-random-seed.service is now a synchronization point for full
1509 initialization of the kernel's entropy pool. Services that require
1510 /dev/urandom to be correctly initialized should be ordered after this
1513 * The systemd-boot boot loader has been updated to optionally maintain
1514 a random seed file in the EFI System Partition (ESP). During the boot
1515 phase, this random seed is read and updated with a new seed
1516 cryptographically derived from it. Another derived seed is passed to
1517 the OS. The latter seed is then credited to the kernel's entropy pool
1518 very early during userspace initialization (from PID 1). This allows
1519 systems to boot up with a fully initialized kernel entropy pool from
1520 earliest boot on, and thus entirely removes all entropy pool
1521 initialization delays from systems using systemd-boot. Special care
1522 is taken to ensure different seeds are derived on system images
1523 replicated to multiple systems. "bootctl status" will show whether
1524 a seed was received from the boot loader.
1526 * bootctl gained two new verbs:
1528 - "bootctl random-seed" will generate the file in ESP and an EFI
1529 variable to allow a random seed to be passed to the OS as described
1532 - "bootctl is-installed" checks whether systemd-boot is currently
1535 * bootctl will warn if it detects that boot entries are misconfigured
1536 (for example if the kernel image was removed without purging the
1539 * A new document has been added describing systemd's use and support
1540 for the kernel's entropy pool subsystem:
1542 https://systemd.io/RANDOM_SEEDS
1544 * When the system is hibernated the swap device to write the
1545 hibernation image to is now automatically picked from all available
1546 swap devices, preferring the swap device with the highest configured
1547 priority over all others, and picking the device with the most free
1548 space if there are multiple devices with the highest priority.
1550 * /etc/crypttab support has learnt a new keyfile-timeout= per-device
1551 option that permits selecting the timeout how long to wait for a
1552 device with an encryption key before asking for the password.
1554 * IOWeight= has learnt to properly set the IO weight when using the
1555 BFQ scheduler officially found in kernels 5.0+.
1557 * A new mailing list has been created for reporting of security issues:
1558 systemd-security@redhat.com. For mode details, see
1559 https://systemd.io/CONTRIBUTING#security-vulnerability-reports.
1561 Contributions from: Aaron Barany, Adrian Bunk, Alan Jenkins, Albrecht
1562 Lohofener, Andrej Valek, Anita Zhang, Arian van Putten, Balint Reczey,
1563 Bastien Nocera, Ben Boeckel, Benjamin Robin, camoz, Chen Qi, Chris
1564 Chiu, Chris Down, Christian Göttsche, Christian Kellner, Clinton Roy,
1565 Connor Reeder, Daniel Black, Daniel Lublin, Daniele Medri, Dan
1566 Streetman, Dave Reisner, Dave Ross, David Art, David Tardon, Debarshi
1567 Ray, Dimitri John Ledkov, Dominick Grift, Donald Buczek, Douglas
1568 Christman, Eric DeVolder, EtherGraf, Evgeny Vereshchagin, Feldwor,
1569 Felix Riemann, Florian Dollinger, Francesco Pennica, Franck Bui,
1570 Frantisek Sumsal, Franz Pletz, frederik, Hans de Goede, Iago López
1571 Galeiras, Insun Pyo, Ivan Shapovalov, Iwan Timmer, Jack, Jakob
1572 Unterwurzacher, Jan Chren, Jan Klötzke, Jan Losinski, Jan Pokorný, Jan
1573 Synacek, Jan-Michael Brummer, Jeka Pats, Jeremy Soller, Jérémy Rosen,
1574 Jiri Pirko, Joe Lin, Joerg Behrmann, Joe Richey, Jóhann B. Guðmundsson,
1575 Johannes Christ, Johannes Schmitz, Jonathan Rouleau, Jorge Niedbalski,
1576 Jörg Thalheim, Kai Krakow, Kai Lüke, Karel Zak, Kashyap Chamarthy,
1577 Krayushkin Konstantin, Lennart Poettering, Lubomir Rintel, Luca
1578 Boccassi, Luís Ferreira, Marc-André Lureau, Markus Felten, Martin Pitt,
1579 Matthew Leeds, Mattias Jernberg, Michael Biebl, Michael Olbrich,
1580 Michael Prokop, Michael Stapelberg, Michael Zhivich, Michal Koutný,
1581 Michal Sekletar, Mike Gilbert, Milan Broz, Miroslav Lichvar, mpe85,
1582 Mr-Foo, Network Silence, Oliver Harley, pan93412, Paul Menzel, pEJipE,
1583 Peter A. Bigot, Philip Withnall, Piotr Drąg, Rafael Fontenelle, Robert
1584 Scheck, Roberto Santalla, Ronan Pigott, root, RussianNeuroMancer,
1585 Sebastian Jennen, shinygold, Shreyas Behera, Simon Schricker, Susant
1586 Sahani, Thadeu Lima de Souza Cascardo, Theo Ouzhinski, Thiebaud
1587 Weksteen, Thomas Haller, Thomas Weißschuh, Tomas Mraz, Tommi Rantala,
1588 Topi Miettinen, VD-Lycos, ven, Vladimir Yerilov, Wieland Hoffmann,
1589 William A. Kennington III, William Wold, Xi Ruoyao, Yuri Chornoivan,
1590 Yu Watanabe, Zach Smith, Zbigniew Jędrzejewski-Szmek, Zhang Xianwei
1592 – Camerino, 2019-09-03
1596 * In .link files, MACAddressPolicy=persistent (the default) is changed
1597 to cover more devices. For devices like bridges, tun, tap, bond, and
1598 similar interfaces that do not have other identifying information,
1599 the interface name is used as the basis for persistent seed for MAC
1600 and IPv4LL addresses. The way that devices that were handled
1601 previously is not changed, and this change is about covering more
1602 devices then previously by the "persistent" policy.
1604 MACAddressPolicy=random may be used to force randomized MACs and
1605 IPv4LL addresses for a device if desired.
1607 Hint: the log output from udev (at debug level) was enhanced to
1608 clarify what policy is followed and which attributes are used.
1609 `SYSTEMD_LOG_LEVEL=debug udevadm test-builtin net_setup_link /sys/class/net/<name>`
1610 may be used to view this.
1612 Hint: if a bridge interface is created without any slaves, and gains
1613 a slave later, then now the bridge does not inherit slave's MAC.
1614 To inherit slave's MAC, for example, create the following file:
1616 # /etc/systemd/network/98-bridge-inherit-mac.link
1621 MACAddressPolicy=none
1624 * The .device units generated by systemd-fstab-generator and other
1625 generators do not automatically pull in the corresponding .mount unit
1626 as a Wants= dependency. This means that simply plugging in the device
1627 will not cause the mount unit to be started automatically. But please
1628 note that the mount unit may be started for other reasons, in
1629 particular if it is part of local-fs.target, and any unit which
1630 (transitively) depends on local-fs.target is started.
1632 * networkctl list/status/lldp now accept globbing wildcards for network
1633 interface names to match against all existing interfaces.
1635 * The $PIDFILE environment variable is set to point the absolute path
1636 configured with PIDFile= for processes of that service.
1638 * The fallback DNS server list was augmented with Cloudflare public DNS
1639 servers. Use `-Ddns-servers=` to set a different fallback.
1641 * A new special target usb-gadget.target will be started automatically
1642 when a USB Device Controller is detected (which means that the system
1643 is a USB peripheral).
1645 * A new unit setting CPUQuotaPeriodSec= assigns the time period
1646 relatively to which the CPU time quota specified by CPUQuota= is
1649 * A new unit setting ProtectHostname= may be used to prevent services
1650 from modifying hostname information (even if they otherwise would
1651 have privileges to do so).
1653 * A new unit setting NetworkNamespacePath= may be used to specify a
1654 namespace for service or socket units through a path referring to a
1655 Linux network namespace pseudo-file.
1657 * The PrivateNetwork= setting and JoinsNamespaceOf= dependencies now
1658 have an effect on .socket units: when used the listening socket is
1659 created within the configured network namespace instead of the host
1662 * ExecStart= command lines in unit files may now be prefixed with ':'
1663 in which case environment variable substitution is
1664 disabled. (Supported for the other ExecXYZ= settings, too.)
1666 * .timer units gained two new boolean settings OnClockChange= and
1667 OnTimezoneChange= which may be used to also trigger a unit when the
1668 system clock is changed or the local timezone is
1669 modified. systemd-run has been updated to make these options easily
1670 accessible from the command line for transient timers.
1672 * Two new conditions for units have been added: ConditionMemory= may be
1673 used to conditionalize a unit based on installed system
1674 RAM. ConditionCPUs= may be used to conditionalize a unit based on
1675 installed CPU cores.
1677 * The @default system call filter group understood by SystemCallFilter=
1678 has been updated to include the new rseq() system call introduced in
1681 * A new time-set.target has been added that indicates that the system
1682 time has been set from a local source (possibly imprecise). The
1683 existing time-sync.target is stronger and indicates that the time has
1684 been synchronized with a precise external source. Services where
1685 approximate time is sufficient should use the new target.
1687 * "systemctl start" (and related commands) learnt a new
1688 --show-transaction option. If specified brief information about all
1689 jobs queued because of the requested operation is shown.
1691 * systemd-networkd recognizes a new operation state 'enslaved', used
1692 (instead of 'degraded' or 'carrier') for interfaces which form a
1693 bridge, bond, or similar, and an new 'degraded-carrier' operational
1694 state used for the bond or bridge master interface when one of the
1695 enslaved devices is not operational.
1697 * .network files learnt the new IgnoreCarrierLoss= option for leaving
1698 networks configured even if the carrier is lost.
1700 * The RequiredForOnline= setting in .network files may now specify a
1701 minimum operational state required for the interface to be considered
1702 "online" by systemd-networkd-wait-online. Related to this
1703 systemd-networkd-wait-online gained a new option --operational-state=
1704 to configure the same, and its --interface= option was updated to
1705 optionally also take an operational state specific for an interface.
1707 * systemd-networkd-wait-online gained a new setting --any for waiting
1708 for only one of the requested interfaces instead of all of them.
1710 * systemd-networkd now implements L2TP tunnels.
1712 * Two new .network settings UseAutonomousPrefix= and UseOnLinkPrefix=
1713 may be used to cause autonomous and onlink prefixes received in IPv6
1714 Router Advertisements to be ignored.
1716 * New MulticastFlood=, NeighborSuppression=, and Learning= .network
1717 file settings may be used to tweak bridge behaviour.
1719 * The new TripleSampling= option in .network files may be used to
1720 configure CAN triple sampling.
1722 * A new .netdev settings PrivateKeyFile= and PresharedKeyFile= may be
1723 used to point to private or preshared key for a WireGuard interface.
1725 * /etc/crypttab now supports the same-cpu-crypt and
1726 submit-from-crypt-cpus options to tweak encryption work scheduling
1729 * systemd-tmpfiles will now take a BSD file lock before operating on a
1730 contents of directory. This may be used to temporarily exclude
1731 directories from aging by taking the same lock (useful for example
1732 when extracting a tarball into /tmp or /var/tmp as a privileged user,
1733 which might create files with really old timestamps, which
1734 nevertheless should not be deleted). For further details, see:
1736 https://systemd.io/TEMPORARY_DIRECTORIES
1738 * systemd-tmpfiles' h line type gained support for the
1739 FS_PROJINHERIT_FL ('P') file attribute (introduced in kernel 4.5),
1740 controlling project quota inheritance.
1742 * sd-boot and bootctl now implement support for an Extended Boot Loader
1743 (XBOOTLDR) partition, that is intended to be mounted to /boot, in
1744 addition to the ESP partition mounted to /efi or /boot/efi.
1745 Configuration file fragments, kernels, initrds and other EFI images
1746 to boot will be loaded from both the ESP and XBOOTLDR partitions.
1747 The XBOOTLDR partition was previously described by the Boot Loader
1748 Specification, but implementation was missing in sd-boot. Support for
1749 this concept allows using the sd-boot boot loader in more
1750 conservative scenarios where the boot loader itself is placed in the
1751 ESP but the kernels to boot (and their metadata) in a separate
1754 * A system may now be booted with systemd.volatile=overlay on the
1755 kernel command line, which causes the root file system to be set up
1756 an overlayfs mount combining the root-only root directory with a
1757 writable tmpfs. In this setup, the underlying root device is not
1758 modified, and any changes are lost at reboot.
1760 * Similar, systemd-nspawn can now boot containers with a volatile
1761 overlayfs root with the new --volatile=overlay switch.
1763 * systemd-nspawn can now consume OCI runtime bundles using a new
1764 --oci-bundle= option. This implementation is fully usable, with most
1765 features in the specification implemented, but since this a lot of
1766 new code and functionality, this feature should most likely not
1767 be used in production yet.
1769 * systemd-nspawn now supports various options described by the OCI
1770 runtime specification on the command-line and in .nspawn files:
1771 --inaccessible=/Inaccessible= may be used to mask parts of the file
1772 system tree, --console=/--pipe may be used to configure how standard
1773 input, output, and error are set up.
1775 * busctl learned the `emit` verb to generate D-Bus signals.
1777 * systemd-analyze cat-config may be used to gather and display
1778 configuration spread over multiple files, for example system and user
1779 presets, tmpfiles.d, sysusers.d, udev rules, etc.
1781 * systemd-analyze calendar now takes an optional new parameter
1782 --iterations= which may be used to show a maximum number of iterations
1783 the specified expression will elapse next.
1785 * The sd-bus C API gained support for naming method parameters in the
1788 * systemd-logind gained D-Bus APIs to specify the "reboot parameter"
1789 the reboot() system call expects.
1791 * journalctl learnt a new --cursor-file= option that points to a file
1792 from which a cursor should be loaded in the beginning and to which
1793 the updated cursor should be stored at the end.
1795 * ACRN hypervisor and Windows Subsystem for Linux (WSL) are now
1796 detected by systemd-detect-virt (and may also be used in
1797 ConditionVirtualization=).
1799 * The behaviour of systemd-logind may now be modified with environment
1800 variables $SYSTEMD_REBOOT_TO_FIRMWARE_SETUP,
1801 $SYSTEMD_REBOOT_TO_BOOT_LOADER_MENU, and
1802 $SYSTEMD_REBOOT_TO_BOOT_LOADER_ENTRY. They cause logind to either
1803 skip the relevant operation completely (when set to false), or to
1804 create a flag file in /run/systemd (when set to true), instead of
1805 actually commencing the real operation when requested. The presence
1806 of /run/systemd/reboot-to-firmware-setup,
1807 /run/systemd/reboot-to-boot-loader-menu, and
1808 /run/systemd/reboot-to-boot-loader-entry, may be used by alternative
1809 boot loader implementations to replace some steps logind performs
1810 during reboot with their own operations.
1812 * systemctl can be used to request a reboot into the boot loader menu
1813 or a specific boot loader entry with the new --boot-load-menu= and
1814 --boot-loader-entry= options to a reboot command. (This requires a
1815 boot loader that supports this, for example sd-boot.)
1817 * kernel-install will no longer unconditionally create the output
1818 directory (e.g. /efi/<machine-id>/<kernel-version>) for boot loader
1819 snippets, but will do only if the machine-specific parent directory
1820 (i.e. /efi/<machine-id>/) already exists. bootctl has been modified
1821 to create this parent directory during sd-boot installation.
1823 This makes it easier to use kernel-install with plugins which support
1824 a different layout of the bootloader partitions (for example grub2).
1826 * During package installation (with `ninja install`), we would create
1827 symlinks for getty@tty1.service, systemd-networkd.service,
1828 systemd-networkd.socket, systemd-resolved.service,
1829 remote-cryptsetup.target, remote-fs.target,
1830 systemd-networkd-wait-online.service, and systemd-timesyncd.service
1831 in /etc, as if `systemctl enable` was called for those units, to make
1832 the system usable immediately after installation. Now this is not
1833 done anymore, and instead calling `systemctl preset-all` is
1834 recommended after the first installation of systemd.
1836 * A new boolean sandboxing option RestrictSUIDSGID= has been added that
1837 is built on seccomp. When turned on creation of SUID/SGID files is
1840 * The NoNewPrivileges= and the new RestrictSUIDSGID= options are now
1841 implied if DynamicUser= is turned on for a service. This hardens
1842 these services, so that they neither can benefit from nor create
1843 SUID/SGID executables. This is a minor compatibility breakage, given
1844 that when DynamicUser= was first introduced SUID/SGID behaviour was
1845 unaffected. However, the security benefit of these two options is
1846 substantial, and the setting is still relatively new, hence we opted
1847 to make it mandatory for services with dynamic users.
1849 Contributions from: Adam Jackson, Alexander Tsoy, Andrey Yashkin,
1850 Andrzej Pietrasiewicz, Anita Zhang, Balint Reczey, Beniamino Galvani,
1851 Ben Iofel, Benjamin Berg, Benjamin Dahlhoff, Chris, Chris Morin,
1852 Christopher Wong, Claudius Ellsel, Clemens Gruber, dana, Daniel Black,
1853 Davide Cavalca, David Michael, David Rheinsberg, emersion, Evgeny
1854 Vereshchagin, Filipe Brandenburger, Franck Bui, Frantisek Sumsal,
1855 Giacinto Cifelli, Hans de Goede, Hugo Kindel, Ignat Korchagin, Insun
1856 Pyo, Jan Engelhardt, Jonas Dorel, Jonathan Lebon, Jonathon Kowalski,
1857 Jörg Sommer, Jörg Thalheim, Jussi Pakkanen, Kai-Heng Feng, Lennart
1858 Poettering, Lubomir Rintel, Luís Ferreira, Martin Pitt, Matthias
1859 Klumpp, Michael Biebl, Michael Niewöhner, Michael Olbrich, Michal
1860 Sekletar, Mike Lothian, Paul Menzel, Piotr Drąg, Riccardo Schirone,
1861 Robin Elvedi, Roman Kulikov, Ronald Tschalär, Ross Burton, Ryan
1862 Gonzalez, Sebastian Krzyszkowiak, Stephane Chazelas, StKob, Susant
1863 Sahani, Sylvain Plantefève, Szabolcs Fruhwald, Taro Yamada, Theo
1864 Ouzhinski, Thomas Haller, Tobias Jungel, Tom Yan, Tony Asleson, Topi
1865 Miettinen, unixsysadmin, Van Laser, Vesa Jääskeläinen, Yu, Li-Yu,
1866 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
1868 — Warsaw, 2019-04-11
1872 * The default locale can now be configured at compile time. Otherwise,
1873 a suitable default will be selected automatically (one of C.UTF-8,
1874 en_US.UTF-8, and C).
1876 * The version string shown by systemd and other tools now includes the
1877 git commit hash when built from git. An override may be specified
1878 during compilation, which is intended to be used by distributions to
1879 include the package release information.
1881 * systemd-cat can now filter standard input and standard error streams
1882 for different syslog priorities using the new --stderr-priority=
1885 * systemd-journald and systemd-journal-remote reject entries which
1886 contain too many fields (CVE-2018-16865) and set limits on the
1887 process' command line length (CVE-2018-16864).
1889 * $DBUS_SESSION_BUS_ADDRESS environment variable is set by pam_systemd
1892 * A new network device NamePolicy "keep" is implemented for link files,
1893 and used by default in 99-default.link (the fallback configuration
1894 provided by systemd). With this policy, if the network device name
1895 was already set by userspace, the device will not be renamed again.
1896 This matches the naming scheme that was implemented before
1897 systemd-240. If naming-scheme < 240 is specified, the "keep" policy
1898 is also enabled by default, even if not specified. Effectively, this
1899 means that if naming-scheme >= 240 is specified, network devices will
1900 be renamed according to the configuration, even if they have been
1901 renamed already, if "keep" is not specified as the naming policy in
1902 the .link file. The 99-default.link file provided by systemd includes
1903 "keep" for backwards compatibility, but it is recommended for user
1904 installed .link files to *not* include it.
1906 The "kernel" policy, which keeps kernel names declared to be
1907 "persistent", now works again as documented.
1909 * kernel-install script now optionally takes the paths to one or more
1910 initrd files, and passes them to all plugins.
1912 * The mincore() system call has been dropped from the @system-service
1913 system call filter group, as it is pretty exotic and may potentially
1914 used for side-channel attacks.
1916 * -fPIE is dropped from compiler and linker options. Please specify
1917 -Db_pie=true option to meson to build position-independent
1918 executables. Note that the meson option is supported since meson-0.49.
1920 * The fs.protected_regular and fs.protected_fifos sysctls, which were
1921 added in Linux 4.19 to make some data spoofing attacks harder, are
1922 now enabled by default. While this will hopefully improve the
1923 security of most installations, it is technically a backwards
1924 incompatible change; to disable these sysctls again, place the
1925 following lines in /etc/sysctl.d/60-protected.conf or a similar file:
1927 fs.protected_regular = 0
1928 fs.protected_fifos = 0
1930 Note that the similar hardlink and symlink protection has been
1931 enabled since v199, and may be disabled likewise.
1933 * The files read from the EnvironmentFile= setting in unit files now
1934 parse backslashes inside quotes literally, matching the behaviour of
1937 * udevadm trigger, udevadm control, udevadm settle and udevadm monitor
1938 now automatically become NOPs when run in a chroot() environment.
1940 * The tmpfiles.d/ "C" line type will now copy directory trees not only
1941 when the destination is so far missing, but also if it already exists
1942 as a directory and is empty. This is useful to cater for systems
1943 where directory trees are put together from multiple separate mount
1944 points but otherwise empty.
1946 * A new function sd_bus_close_unref() (and the associated
1947 sd_bus_close_unrefp()) has been added to libsystemd, that combines
1948 sd_bus_close() and sd_bus_unref() in one.
1950 * udevadm control learnt a new option for --ping for testing whether a
1951 systemd-udevd instance is running and reacting.
1953 * udevadm trigger learnt a new option for --wait-daemon for waiting
1954 systemd-udevd daemon to be initialized.
1956 Contributions from: Aaron Plattner, Alberts Muktupāvels, Alex Mayer,
1957 Ayman Bagabas, Beniamino Galvani, Burt P, Chris Down, Chris Lamb, Chris
1958 Morin, Christian Hesse, Claudius Ellsel, dana, Daniel Axtens, Daniele
1959 Medri, Dave Reisner, David Santamaría Rogado, Diego Canuhe, Dimitri
1960 John Ledkov, Evgeny Vereshchagin, Fabrice Fontaine, Filipe
1961 Brandenburger, Franck Bui, Frantisek Sumsal, govwin, Hans de Goede,
1962 James Hilliard, Jan Engelhardt, Jani Uusitalo, Jan Janssen, Jan
1963 Synacek, Jonathan McDowell, Jonathan Roemer, Jonathon Kowalski, Joost
1964 Heitbrink, Jörg Thalheim, Lance, Lennart Poettering, Louis Taylor,
1965 Lucas Werkmeister, Mantas Mikulėnas, Marc-Antoine Perennou,
1966 marvelousblack, Michael Biebl, Michael Sloan, Michal Sekletar, Mike
1967 Auty, Mike Gilbert, Mikhail Kasimov, Neil Brown, Niklas Hambüchen,
1968 Patrick Williams, Paul Seyfert, Peter Hutterer, Philip Withnall, Roger
1969 James, Ronnie P. Thomas, Ryan Gonzalez, Sam Morris, Stephan Edel,
1970 Stephan Gerhold, Susant Sahani, Taro Yamada, Thomas Haller, Topi
1971 Miettinen, YiFei Zhu, YmrDtnJu, YunQiang Su, Yu Watanabe, Zbigniew
1972 Jędrzejewski-Szmek, zsergeant77, Дамјан Георгиевски
1974 — Berlin, 2019-02-14
1978 * NoNewPrivileges=yes has been set for all long-running services
1979 implemented by systemd. Previously, this was problematic due to
1980 SELinux (as this would also prohibit the transition from PID1's label
1981 to the service's label). This restriction has since been lifted, but
1982 an SELinux policy update is required.
1983 (See e.g. https://github.com/fedora-selinux/selinux-policy/pull/234.)
1985 * DynamicUser=yes is dropped from systemd-networkd.service,
1986 systemd-resolved.service and systemd-timesyncd.service, which was
1987 enabled in v239 for systemd-networkd.service and systemd-resolved.service,
1988 and since v236 for systemd-timesyncd.service. The users and groups
1989 systemd-network, systemd-resolve and systemd-timesync are created
1990 by systemd-sysusers again. Distributors or system administrators
1991 may need to create these users and groups if they not exist (or need
1992 to re-enable DynamicUser= for those units) while upgrading systemd.
1993 Also, the clock file for systemd-timesyncd may need to move from
1994 /var/lib/private/systemd/timesync/clock to /var/lib/systemd/timesync/clock.
1996 * When unit files are loaded from disk, previously systemd would
1997 sometimes (depending on the unit loading order) load units from the
1998 target path of symlinks in .wants/ or .requires/ directories of other
1999 units. This meant that unit could be loaded from different paths
2000 depending on whether the unit was requested explicitly or as a
2001 dependency of another unit, not honouring the priority of directories
2002 in search path. It also meant that it was possible to successfully
2003 load and start units which are not found in the unit search path, as
2004 long as they were requested as a dependency and linked to from
2005 .wants/ or .requires/. The target paths of those symlinks are not
2006 used for loading units anymore and the unit file must be found in
2009 * A new service type has been added: Type=exec. It's very similar to
2010 Type=simple but ensures the service manager will wait for both fork()
2011 and execve() of the main service binary to complete before proceeding
2012 with follow-up units. This is primarily useful so that the manager
2013 propagates any errors in the preparation phase of service execution
2014 back to the job that requested the unit to be started. For example,
2015 consider a service that has ExecStart= set to a file system binary
2016 that doesn't exist. With Type=simple starting the unit would be
2017 considered instantly successful, as only fork() has to complete
2018 successfully and the manager does not wait for execve(), and hence
2019 its failure is seen "too late". With the new Type=exec service type
2020 starting the unit will fail, as the manager will wait for the
2021 execve() and notice its failure, which is then propagated back to the
2024 NOTE: with the next release 241 of systemd we intend to change the
2025 systemd-run tool to default to Type=exec for transient services
2026 started by it. This should be mostly safe, but in specific corner
2027 cases might result in problems, as the systemd-run tool will then
2028 block on NSS calls (such as user name look-ups due to User=) done
2029 between the fork() and execve(), which under specific circumstances
2030 might cause problems. It is recommended to specify "-p Type=simple"
2031 explicitly in the few cases where this applies. For regular,
2032 non-transient services (i.e. those defined with unit files on disk)
2033 we will continue to default to Type=simple.
2035 * The Linux kernel's current default RLIMIT_NOFILE resource limit for
2036 userspace processes is set to 1024 (soft) and 4096
2037 (hard). Previously, systemd passed this on unmodified to all
2038 processes it forked off. With this systemd release the hard limit
2039 systemd passes on is increased to 512K, overriding the kernel's
2040 defaults and substantially increasing the number of simultaneous file
2041 descriptors unprivileged userspace processes can allocate. Note that
2042 the soft limit remains at 1024 for compatibility reasons: the
2043 traditional UNIX select() call cannot deal with file descriptors >=
2044 1024 and increasing the soft limit globally might thus result in
2045 programs unexpectedly allocating a high file descriptor and thus
2046 failing abnormally when attempting to use it with select() (of
2047 course, programs shouldn't use select() anymore, and prefer
2048 poll()/epoll, but the call unfortunately remains undeservedly popular
2049 at this time). This change reflects the fact that file descriptor
2050 handling in the Linux kernel has been optimized in more recent
2051 kernels and allocating large numbers of them should be much cheaper
2052 both in memory and in performance than it used to be. Programs that
2053 want to take benefit of the increased limit have to "opt-in" into
2054 high file descriptors explicitly by raising their soft limit. Of
2055 course, when they do that they must acknowledge that they cannot use
2056 select() anymore (and neither can any shared library they use — or
2057 any shared library used by any shared library they use and so on).
2058 Which default hard limit is most appropriate is of course hard to
2059 decide. However, given reports that ~300K file descriptors are used
2060 in real-life applications we believe 512K is sufficiently high as new
2061 default for now. Note that there are also reports that using very
2062 high hard limits (e.g. 1G) is problematic: some software allocates
2063 large arrays with one element for each potential file descriptor
2064 (Java, …) — a high hard limit thus triggers excessively large memory
2065 allocations in these applications. Hopefully, the new default of 512K
2066 is a good middle ground: higher than what real-life applications
2067 currently need, and low enough for avoid triggering excessively large
2068 allocations in problematic software. (And yes, somebody should fix
2071 * The fs.nr_open and fs.file-max sysctls are now automatically bumped
2072 to the highest possible values, as separate accounting of file
2073 descriptors is no longer necessary, as memcg tracks them correctly as
2074 part of the memory accounting anyway. Thus, from the four limits on
2075 file descriptors currently enforced (fs.file-max, fs.nr_open,
2076 RLIMIT_NOFILE hard, RLIMIT_NOFILE soft) we turn off the first two,
2077 and keep only the latter two. A set of build-time options
2078 (-Dbump-proc-sys-fs-file-max=false and -Dbump-proc-sys-fs-nr-open=false)
2079 has been added to revert this change in behaviour, which might be
2080 an option for systems that turn off memcg in the kernel.
2082 * When no /etc/locale.conf file exists (and hence no locale settings
2083 are in place), systemd will now use the "C.UTF-8" locale by default,
2084 and set LANG= to it. This locale is supported by various
2085 distributions including Fedora, with clear indications that upstream
2086 glibc is going to make it available too. This locale enables UTF-8
2087 mode by default, which appears appropriate for 2018.
2089 * The "net.ipv4.conf.all.rp_filter" sysctl will now be set to 2 by
2090 default. This effectively switches the RFC3704 Reverse Path filtering
2091 from Strict mode to Loose mode. This is more appropriate for hosts
2092 that have multiple links with routes to the same networks (e.g.
2093 a client with a Wi-Fi and Ethernet both connected to the internet).
2095 Consult the kernel documentation for details on this sysctl:
2096 https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt
2098 * CPUAccounting=yes no longer enables the CPU controller when using
2099 kernel 4.15+ and the unified cgroup hierarchy, as required accounting
2100 statistics are now provided independently from the CPU controller.
2102 * Support for disabling a particular cgroup controller within a sub-tree
2103 has been added through the DisableControllers= directive.
2105 * cgroup_no_v1=all on the kernel command line now also implies
2106 using the unified cgroup hierarchy, unless one explicitly passes
2107 systemd.unified_cgroup_hierarchy=0 on the kernel command line.
2109 * The new "MemoryMin=" unit file property may now be used to set the
2110 memory usage protection limit of processes invoked by the unit. This
2111 controls the cgroup v2 memory.min attribute. Similarly, the new
2112 "IODeviceLatencyTargetSec=" property has been added, wrapping the new
2113 cgroup v2 io.latency cgroup property for configuring per-service I/O
2116 * systemd now supports the cgroup v2 devices BPF logic, as counterpart
2117 to the cgroup v1 "devices" cgroup controller.
2119 * systemd-escape now is able to combine --unescape with --template. It
2120 also learnt a new option --instance for extracting and unescaping the
2121 instance part of a unit name.
2123 * sd-bus now provides the sd_bus_message_readv() which is similar to
2124 sd_bus_message_read() but takes a va_list object. The pair
2125 sd_bus_set_method_call_timeout() and sd_bus_get_method_call_timeout()
2126 has been added for configuring the default method call timeout to
2127 use. sd_bus_error_move() may be used to efficiently move the contents
2128 from one sd_bus_error structure to another, invalidating the
2129 source. sd_bus_set_close_on_exit() and sd_bus_get_close_on_exit() may
2130 be used to control whether a bus connection object is automatically
2131 flushed when an sd-event loop is exited.
2133 * When processing classic BSD syslog log messages, journald will now
2134 save the original time-stamp string supplied in the new
2135 SYSLOG_TIMESTAMP= journal field. This permits consumers to
2136 reconstruct the original BSD syslog message more correctly.
2138 * StandardOutput=/StandardError= in service files gained support for
2139 new "append:…" parameters, for connecting STDOUT/STDERR of a service
2140 to a file, and appending to it.
2142 * The signal to use as last step of killing of unit processes is now
2143 configurable. Previously it was hard-coded to SIGKILL, which may now
2144 be overridden with the new KillSignal= setting. Note that this is the
2145 signal used when regular termination (i.e. SIGTERM) does not suffice.
2146 Similarly, the signal used when aborting a program in case of a
2147 watchdog timeout may now be configured too (WatchdogSignal=).
2149 * The XDG_SESSION_DESKTOP environment variable may now be configured in
2150 the pam_systemd argument line, using the new desktop= switch. This is
2151 useful to initialize it properly from a display manager without
2152 having to touch C code.
2154 * Most configuration options that previously accepted percentage values
2155 now also accept permille values with the '‰' suffix (instead of '%').
2157 * systemd-resolved may now optionally use OpenSSL instead of GnuTLS for
2160 * systemd-resolved's configuration file resolved.conf gained a new
2161 option ReadEtcHosts= which may be used to turn off processing and
2162 honoring /etc/hosts entries.
2164 * The "--wait" switch may now be passed to "systemctl
2165 is-system-running", in which case the tool will synchronously wait
2166 until the system finished start-up.
2168 * hostnamed gained a new bus call to determine the DMI product UUID.
2170 * On x86-64 systemd will now prefer using the RDRAND processor
2171 instruction over /dev/urandom whenever it requires randomness that
2172 neither has to be crypto-grade nor should be reproducible. This
2173 should substantially reduce the amount of entropy systemd requests
2174 from the kernel during initialization on such systems, though not
2175 reduce it to zero. (Why not zero? systemd still needs to allocate
2176 UUIDs and such uniquely, which require high-quality randomness.)
2178 * networkd gained support for Foo-Over-UDP, ERSPAN and ISATAP
2179 tunnels. It also gained a new option ForceDHCPv6PDOtherInformation=
2180 for forcing the "Other Information" bit in IPv6 RA messages. The
2181 bonding logic gained four new options AdActorSystemPriority=,
2182 AdUserPortKey=, AdActorSystem= for configuring various 802.3ad
2183 aspects, and DynamicTransmitLoadBalancing= for enabling dynamic
2184 shuffling of flows. The tunnel logic gained a new
2185 IPv6RapidDeploymentPrefix= option for configuring IPv6 Rapid
2186 Deployment. The policy rule logic gained four new options IPProtocol=,
2187 SourcePort= and DestinationPort=, InvertRule=. The bridge logic gained
2188 support for the MulticastToUnicast= option. networkd also gained
2189 support for configuring static IPv4 ARP or IPv6 neighbor entries.
2191 * .preset files (as read by 'systemctl preset') may now be used to
2192 instantiate services.
2194 * /etc/crypttab now understands the sector-size= option to configure
2195 the sector size for an encrypted partition.
2197 * Key material for encrypted disks may now be placed on a formatted
2198 medium, and referenced from /etc/crypttab by the UUID of the file
2199 system, followed by "=" suffixed by the path to the key file.
2201 * The "collect" udev component has been removed without replacement, as
2202 it is neither used nor maintained.
2204 * When the RuntimeDirectory=, StateDirectory=, CacheDirectory=,
2205 LogsDirectory=, ConfigurationDirectory= settings are used in a
2206 service the executed processes will now receive a set of environment
2207 variables containing the full paths of these directories.
2208 Specifically, RUNTIME_DIRECTORY=, STATE_DIRECTORY, CACHE_DIRECTORY,
2209 LOGS_DIRECTORY, CONFIGURATION_DIRECTORY are now set if these options
2210 are used. Note that these options may be used multiple times per
2211 service in which case the resulting paths will be concatenated and
2212 separated by colons.
2214 * Predictable interface naming has been extended to cover InfiniBand
2215 NICs. They will be exposed with an "ib" prefix.
2217 * tmpfiles.d/ line types may now be suffixed with a '-' character, in
2218 which case the respective line failing is ignored.
2220 * .link files may now be used to configure the equivalent to the
2221 "ethtool advertise" commands.
2223 * The sd-device.h and sd-hwdb.h APIs are now exported, as an
2224 alternative to libudev.h. Previously, the latter was just an internal
2225 wrapper around the former, but now these two APIs are exposed
2228 * sd-id128.h gained a new function sd_id128_get_boot_app_specific()
2229 which calculates an app-specific boot ID similar to how
2230 sd_id128_get_machine_app_specific() generates an app-specific machine
2233 * A new tool systemd-id128 has been added that can be used to determine
2234 and generate various 128bit IDs.
2236 * /etc/os-release gained two new standardized fields DOCUMENTATION_URL=
2239 * systemd-hibernate-resume-generator will now honor the "noresume"
2240 kernel command line option, in which case it will bypass resuming
2241 from any hibernated image.
2243 * The systemd-sleep.conf configuration file gained new options
2244 AllowSuspend=, AllowHibernation=, AllowSuspendThenHibernate=,
2245 AllowHybridSleep= for prohibiting specific sleep modes even if the
2246 kernel exports them.
2248 * portablectl is now officially supported and has thus moved to
2251 * bootctl learnt the two new commands "set-default" and "set-oneshot"
2252 for setting the default boot loader item to boot to (either
2253 persistently or only for the next boot). This is currently only
2254 compatible with sd-boot, but may be implemented on other boot loaders
2255 too, that follow the boot loader interface. The updated interface is
2256 now documented here:
2258 https://systemd.io/BOOT_LOADER_INTERFACE
2260 * A new kernel command line option systemd.early_core_pattern= is now
2261 understood which may be used to influence the core_pattern PID 1
2262 installs during early boot.
2264 * busctl learnt two new options -j and --json= for outputting method
2265 call replies, properties and monitoring output in JSON.
2267 * journalctl's JSON output now supports simple ANSI coloring as well as
2268 a new "json-seq" mode for generating RFC7464 output.
2270 * Unit files now support the %g/%G specifiers that resolve to the UNIX
2271 group/GID of the service manager runs as, similar to the existing
2272 %u/%U specifiers that resolve to the UNIX user/UID.
2274 * systemd-logind learnt a new global configuration option
2275 UserStopDelaySec= that may be set in logind.conf. It specifies how
2276 long the systemd --user instance shall remain started after a user
2277 logs out. This is useful to speed up repetitive re-connections of the
2278 same user, as it means the user's service manager doesn't have to be
2279 stopped/restarted on each iteration, but can be reused between
2280 subsequent options. This setting defaults to 10s. systemd-logind also
2281 exports two new properties on its Manager D-Bus objects indicating
2282 whether the system's lid is currently closed, and whether the system
2285 * systemd gained support for a generic boot counting logic, which
2286 generically permits automatic reverting to older boot loader entries
2287 if newer updated ones don't work. The boot loader side is implemented
2288 in sd-boot, but is kept open for other boot loaders too. For details
2291 https://systemd.io/AUTOMATIC_BOOT_ASSESSMENT
2293 * The SuccessAction=/FailureAction= unit file settings now learnt two
2294 new parameters: "exit" and "exit-force", which result in immediate
2295 exiting of the service manager, and are only useful in systemd --user
2296 and container environments.
2298 * Unit files gained support for a pair of options
2299 FailureActionExitStatus=/SuccessActionExitStatus= for configuring the
2300 exit status to use as service manager exit status when
2301 SuccessAction=/FailureAction= is set to exit or exit-force.
2303 * A pair of LogRateLimitIntervalSec=/LogRateLimitBurst= per-service
2304 options may now be used to configure the log rate limiting applied by
2305 journald per-service.
2307 * systemd-analyze gained a new verb "timespan" for parsing and
2308 normalizing time span values (i.e. strings like "5min 7s 8us").
2310 * systemd-analyze also gained a new verb "security" for analyzing the
2311 security and sand-boxing settings of services in order to determine an
2312 "exposure level" for them, indicating whether a service would benefit
2313 from more sand-boxing options turned on for them.
2315 * "systemd-analyze syscall-filter" will now also show system calls
2316 supported by the local kernel but not included in any of the defined
2319 * .nspawn files now understand the Ephemeral= setting, matching the
2320 --ephemeral command line switch.
2322 * sd-event gained the new APIs sd_event_source_get_floating() and
2323 sd_event_source_set_floating() for controlling whether a specific
2324 event source is "floating", i.e. destroyed along with the even loop
2327 * Unit objects on D-Bus gained a new "Refs" property that lists all
2328 clients that currently have a reference on the unit (to ensure it is
2331 * The JoinControllers= option in system.conf is no longer supported, as
2332 it didn't work correctly, is hard to support properly, is legacy (as
2333 the concept only exists on cgroup v1) and apparently wasn't used.
2335 * Journal messages that are generated whenever a unit enters the failed
2336 state are now tagged with a unique MESSAGE_ID. Similarly, messages
2337 generated whenever a service process exits are now made recognizable,
2338 too. A tagged message is also emitted whenever a unit enters the
2339 "dead" state on success.
2341 * systemd-run gained a new switch --working-directory= for configuring
2342 the working directory of the service to start. A shortcut -d is
2343 equivalent, setting the working directory of the service to the
2344 current working directory of the invoking program. The new --shell
2345 (or just -S) option has been added for invoking the $SHELL of the
2346 caller as a service, and implies --pty --same-dir --wait --collect
2347 --service-type=exec. Or in other words, "systemd-run -S" is now the
2348 quickest way to quickly get an interactive in a fully clean and
2349 well-defined system service context.
2351 * machinectl gained a new verb "import-fs" for importing an OS tree
2352 from a directory. Moreover, when a directory or tarball is imported
2353 and single top-level directory found with the OS itself below the OS
2354 tree is automatically mangled and moved one level up.
2356 * systemd-importd will no longer set up an implicit btrfs loop-back
2357 file system on /var/lib/machines. If one is already set up, it will
2358 continue to be used.
2360 * A new generator "systemd-run-generator" has been added. It will
2361 synthesize a unit from one or more program command lines included in
2362 the kernel command line. This is very useful in container managers
2365 # systemd-nspawn -i someimage.raw -b systemd.run='"some command line"'
2367 This will run "systemd-nspawn" on an image, invoke the specified
2368 command line and immediately shut down the container again, returning
2369 the command line's exit code.
2371 * The block device locking logic is now documented:
2373 https://systemd.io/BLOCK_DEVICE_LOCKING
2375 * loginctl and machinectl now optionally output the various tables in
2376 JSON using the --output= switch. It is our intention to add similar
2377 support to systemctl and all other commands.
2379 * udevadm's query and trigger verb now optionally take a .device unit
2382 * systemd-udevd's network naming logic now understands a new
2383 net.naming-scheme= kernel command line switch, which may be used to
2384 pick a specific version of the naming scheme. This helps stabilizing
2385 interface names even as systemd/udev are updated and the naming logic
2388 * sd-id128.h learnt two new auxiliary helpers: sd_id128_is_allf() and
2389 SD_ID128_ALLF to test if a 128bit ID is set to all 0xFF bytes, and to
2390 initialize one to all 0xFF.
2392 * After loading the SELinux policy systemd will now recursively relabel
2393 all files and directories listed in
2394 /run/systemd/relabel-extra.d/*.relabel (which should be simple
2395 newline separated lists of paths) in addition to the ones it already
2396 implicitly relabels in /run, /dev and /sys. After the relabelling is
2397 completed the *.relabel files (and /run/systemd/relabel-extra.d/) are
2398 removed. This is useful to permit initrds (i.e. code running before
2399 the SELinux policy is in effect) to generate files in the host
2400 filesystem safely and ensure that the correct label is applied during
2401 the transition to the host OS.
2403 * KERNEL API BREAKAGE: Linux kernel 4.18 changed behaviour regarding
2404 mknod() handling in user namespaces. Previously mknod() would always
2405 fail with EPERM in user namespaces. Since 4.18 mknod() will succeed
2406 but device nodes generated that way cannot be opened, and attempts to
2407 open them result in EPERM. This breaks the "graceful fallback" logic
2408 in systemd's PrivateDevices= sand-boxing option. This option is
2409 implemented defensively, so that when systemd detects it runs in a
2410 restricted environment (such as a user namespace, or an environment
2411 where mknod() is blocked through seccomp or absence of CAP_SYS_MKNOD)
2412 where device nodes cannot be created the effect of PrivateDevices= is
2413 bypassed (following the logic that 2nd-level sand-boxing is not
2414 essential if the system systemd runs in is itself already sand-boxed
2415 as a whole). This logic breaks with 4.18 in container managers where
2416 user namespacing is used: suddenly PrivateDevices= succeeds setting
2417 up a private /dev/ file system containing devices nodes — but when
2418 these are opened they don't work.
2420 At this point it is recommended that container managers utilizing
2421 user namespaces that intend to run systemd in the payload explicitly
2422 block mknod() with seccomp or similar, so that the graceful fallback
2425 We are very sorry for the breakage and the requirement to change
2426 container configurations for newer kernels. It's purely caused by an
2427 incompatible kernel change. The relevant kernel developers have been
2428 notified about this userspace breakage quickly, but they chose to
2431 * PermissionsStartOnly= setting is deprecated (but is still supported
2432 for backwards compatibility). The same functionality is provided by
2433 the more flexible "+", "!", and "!!" prefixes to ExecStart= and other
2436 * $DBUS_SESSION_BUS_ADDRESS environment variable is not set by
2437 pam_systemd anymore.
2439 * The naming scheme for network devices was changed to always rename
2440 devices, even if they were already renamed by userspace. The "kernel"
2441 policy was changed to only apply as a fallback, if no other naming
2444 * The requirements to build systemd is bumped to meson-0.46 and
2447 Contributions from: afg, Alan Jenkins, Aleksei Timofeyev, Alexander
2448 Filippov, Alexander Kurtz, Alexey Bogdanenko, Andreas Henriksson,
2449 Andrew Jorgensen, Anita Zhang, apnix-uk, Arkan49, Arseny Maslennikov,
2450 asavah, Asbjørn Apeland, aszlig, Bastien Nocera, Ben Boeckel, Benedikt
2451 Morbach, Benjamin Berg, Bruce Zhang, Carlo Caione, Cedric Viou, Chen
2452 Qi, Chris Chiu, Chris Down, Chris Morin, Christian Rebischke, Claudius
2453 Ellsel, Colin Guthrie, dana, Daniel, Daniele Medri, Daniel Kahn
2454 Gillmor, Daniel Rusek, Daniel van Vugt, Dariusz Gadomski, Dave Reisner,
2455 David Anderson, Davide Cavalca, David Leeds, David Malcolm, David
2456 Strauss, David Tardon, Dimitri John Ledkov, Dmitry Torokhov, dj-kaktus,
2457 Dongsu Park, Elias Probst, Emil Soleyman, Erik Kooistra, Ervin Peters,
2458 Evgeni Golov, Evgeny Vereshchagin, Fabrice Fontaine, Faheel Ahmad,
2459 Faizal Luthfi, Felix Yan, Filipe Brandenburger, Franck Bui, Frank
2460 Schaefer, Frantisek Sumsal, Gautier Husson, Gianluca Boiano, Giuseppe
2461 Scrivano, glitsj16, Hans de Goede, Harald Hoyer, Harry Mallon, Harshit
2462 Jain, Helmut Grohne, Henry Tung, Hui Yiqun, imayoda, Insun Pyo, Iwan
2463 Timmer, Jan Janssen, Jan Pokorný, Jan Synacek, Jason A. Donenfeld,
2464 javitoom, Jérémy Nouhaud, Jeremy Su, Jiuyang Liu, João Paulo Rechi
2465 Vita, Joe Hershberger, Joe Rayhawk, Joerg Behrmann, Joerg Steffens,
2466 Jonas Dorel, Jon Ringle, Josh Soref, Julian Andres Klode, Jun Bo Bi,
2467 Jürg Billeter, Keith Busch, Khem Raj, Kirill Marinushkin, Larry
2468 Bernstone, Lennart Poettering, Lion Yang, Li Song, Lorenz
2469 Hübschle-Schneider, Lubomir Rintel, Lucas Werkmeister, Ludwin Janvier,
2470 Lukáš Nykrýn, Luke Shumaker, mal, Marc-Antoine Perennou, Marcin
2471 Skarbek, Marco Trevisan (Treviño), Marian Cepok, Mario Hros, Marko
2472 Myllynen, Markus Grimm, Martin Pitt, Martin Sobotka, Martin Wilck,
2473 Mathieu Trudel-Lapierre, Matthew Leeds, Michael Biebl, Michael Olbrich,
2474 Michael 'pbone' Pobega, Michael Scherer, Michal Koutný, Michal
2475 Sekletar, Michal Soltys, Mike Gilbert, Mike Palmer, Muhammet Kara, Neal
2476 Gompa, Neil Brown, Network Silence, Niklas Tibbling, Nikolas Nyby,
2477 Nogisaka Sadata, Oliver Smith, Patrik Flykt, Pavel Hrdina, Paweł
2478 Szewczyk, Peter Hutterer, Piotr Drąg, Ray Strode, Reinhold Mueller,
2479 Renaud Métrich, Roman Gushchin, Ronny Chevalier, Rubén Suárez Alvarez,
2480 Ruixin Bao, RussianNeuroMancer, Ryutaroh Matsumoto, Saleem Rashid, Sam
2481 Morris, Samuel Morris, Sandy Carter, scootergrisen, Sébastien Bacher,
2482 Sergey Ptashnick, Shawn Landden, Shengyao Xue, Shih-Yuan Lee
2483 (FourDollars), Silvio Knizek, Sjoerd Simons, Stasiek Michalski, Stephen
2484 Gallagher, Steven Allen, Steve Ramage, Susant Sahani, Sven Joachim,
2485 Sylvain Plantefève, Tanu Kaskinen, Tejun Heo, Thiago Macieira, Thomas
2486 Blume, Thomas Haller, Thomas H. P. Andersen, Tim Ruffing, TJ, Tobias
2487 Jungel, Todd Walton, Tommi Rantala, Tomsod M, Tony Novak, Tore
2488 Anderson, Trevonn, Victor Laskurain, Victor Tapia, Violet Halo, Vojtech
2489 Trefny, welaq, William A. Kennington III, William Douglas, Wyatt Ward,
2490 Xiang Fan, Xi Ruoyao, Xuanwo, Yann E. Morin, YmrDtnJu, Yu Watanabe,
2491 Zbigniew Jędrzejewski-Szmek, Zhang Xianwei, Zsolt Dollenstein
2493 — Warsaw, 2018-12-21
2497 * NETWORK INTERFACE DEVICE NAMING CHANGES: systemd-udevd's "net_id"
2498 builtin will name network interfaces differently than in previous
2499 versions for virtual network interfaces created with SR-IOV and NPAR
2500 and for devices where the PCI network controller device does not have
2501 a slot number associated.
2503 SR-IOV virtual devices are now named based on the name of the parent
2504 interface, with a suffix of "v<N>", where <N> is the virtual device
2505 number. Previously those virtual devices were named as if completely
2508 The ninth and later NPAR virtual devices will be named following the
2509 scheme used for the first eight NPAR partitions. Previously those
2510 devices were not renamed and the kernel default (eth<n>) was used.
2512 "net_id" will also generate names for PCI devices where the PCI
2513 network controller device does not have an associated slot number
2514 itself, but one of its parents does. Previously those devices were
2515 not renamed and the kernel default (eth<n>) was used.
2517 * AF_INET and AF_INET6 are dropped from RestrictAddressFamilies= in
2518 systemd-logind.service. Since v235, IPAddressDeny=any has been set to
2519 the unit. So, it is expected that the default behavior of
2520 systemd-logind is not changed. However, if distribution packagers or
2521 administrators disabled or modified IPAddressDeny= setting by a
2522 drop-in config file, then it may be necessary to update the file to
2523 re-enable AF_INET and AF_INET6 to support network user name services,
2526 * When the RestrictNamespaces= unit property is specified multiple
2527 times, then the specified types are merged now. Previously, only the
2528 last assignment was used. So, if distribution packagers or
2529 administrators modified the setting by a drop-in config file, then it
2530 may be necessary to update the file.
2532 * When OnFailure= is used in combination with Restart= on a service
2533 unit, then the specified units will no longer be triggered on
2534 failures that result in restarting. Previously, the specified units
2535 would be activated each time the unit failed, even when the unit was
2536 going to be restarted automatically. This behaviour contradicted the
2537 documentation. With this release the code is adjusted to match the
2540 * systemd-tmpfiles will now print a notice whenever it encounters
2541 tmpfiles.d/ lines referencing the /var/run/ directory. It will
2542 recommend reworking them to use the /run/ directory instead (for
2543 which /var/run/ is simply a symlinked compatibility alias). This way
2544 systemd-tmpfiles can properly detect line conflicts and merge lines
2545 referencing the same file by two paths, without having to access
2548 * systemctl disable/unmask/preset/preset-all cannot be used with
2549 --runtime. Previously this was allowed, but resulted in unintuitive
2550 behaviour that wasn't useful. systemctl disable/unmask will now undo
2551 both runtime and persistent enablement/masking, i.e. it will remove
2552 any relevant symlinks both in /run and /etc.
2554 * Note that all long-running system services shipped with systemd will
2555 now default to a system call allow list (rather than a deny list, as
2556 before). In particular, systemd-udevd will now enforce one too. For
2557 most cases this should be safe, however downstream distributions
2558 which disabled sandboxing of systemd-udevd (specifically the
2559 MountFlags= setting), might want to disable this security feature
2560 too, as the default allow-listing will prohibit all mount, swap,
2561 reboot and clock changing operations from udev rules.
2563 * sd-boot acquired new loader configuration settings to optionally turn
2564 off Windows and MacOS boot partition discovery as well as
2565 reboot-into-firmware menu items. It is also able to pick a better
2566 screen resolution for HiDPI systems, and now provides loader
2567 configuration settings to change the resolution explicitly.
2569 * systemd-resolved now supports DNS-over-TLS. It's still
2570 turned off by default, use DNSOverTLS=opportunistic to turn it on in
2571 resolved.conf. We intend to make this the default as soon as couple
2572 of additional techniques for optimizing the initial latency caused by
2573 establishing a TLS/TCP connection are implemented.
2575 * systemd-resolved.service and systemd-networkd.service now set
2576 DynamicUser=yes. The users systemd-resolve and systemd-network are
2577 not created by systemd-sysusers anymore.
2579 NOTE: This has a chance of breaking nss-ldap and similar NSS modules
2580 that embed a network facing module into any process using getpwuid()
2581 or related call: the dynamic allocation of the user ID for
2582 systemd-resolved.service means the service manager has to check NSS
2583 if the user name is already taken when forking off the service. Since
2584 the user in the common case won't be defined in /etc/passwd the
2585 lookup is likely to trigger nss-ldap which in turn might use NSS to
2586 ask systemd-resolved for hostname lookups. This will hence result in
2587 a deadlock: a user name lookup in order to start
2588 systemd-resolved.service will result in a hostname lookup for which
2589 systemd-resolved.service needs to be started already. There are
2590 multiple ways to work around this problem: pre-allocate the
2591 "systemd-resolve" user on such systems, so that nss-ldap won't be
2592 triggered; or use a different NSS package that doesn't do networking
2593 in-process but provides a local asynchronous name cache; or configure
2594 the NSS package to avoid lookups for UIDs in the range `pkg-config
2595 systemd --variable=dynamicuidmin` … `pkg-config systemd
2596 --variable=dynamicuidmax`, so that it does not consider itself
2597 authoritative for the same UID range systemd allocates dynamic users
2600 * The systemd-resolve tool has been renamed to resolvectl (it also
2601 remains available under the old name, for compatibility), and its
2602 interface is now verb-based, similar in style to the other <xyz>ctl
2603 tools, such as systemctl or loginctl.
2605 * The resolvectl/systemd-resolve tool also provides 'resolvconf'
2606 compatibility. It may be symlinked under the 'resolvconf' name, in
2607 which case it will take arguments and input compatible with the
2608 Debian and FreeBSD resolvconf tool.
2610 * Support for suspend-then-hibernate has been added, i.e. a sleep mode
2611 where the system initially suspends, and after a timeout resumes and
2614 * networkd's ClientIdentifier= now accepts a new option "duid-only". If
2615 set the client will only send a DUID as client identifier.
2617 * The nss-systemd glibc NSS module will now enumerate dynamic users and
2618 groups in effect. Previously, it could resolve UIDs/GIDs to user
2619 names/groups and vice versa, but did not support enumeration.
2621 * journald's Compress= configuration setting now optionally accepts a
2622 byte threshold value. All journal objects larger than this threshold
2623 will be compressed, smaller ones will not. Previously this threshold
2624 was not configurable and set to 512.
2626 * A new system.conf setting NoNewPrivileges= is now available which may
2627 be used to turn off acquisition of new privileges system-wide
2628 (i.e. set Linux' PR_SET_NO_NEW_PRIVS for PID 1 itself, and thus also
2629 for all its children). Note that turning this option on means setuid
2630 binaries and file system capabilities lose their special powers.
2631 While turning on this option is a big step towards a more secure
2632 system, doing so is likely to break numerous pre-existing UNIX tools,
2633 in particular su and sudo.
2635 * A new service systemd-time-sync-wait.service has been added. If
2636 enabled it will delay the time-sync.target unit at boot until time
2637 synchronization has been received from the network. This
2638 functionality is useful on systems lacking a local RTC or where it is
2639 acceptable that the boot process shall be delayed by external network
2642 * When hibernating, systemd will now inform the kernel of the image
2643 write offset, on kernels new enough to support this. This means swap
2644 files should work for hibernation now.
2646 * When loading unit files, systemd will now look for drop-in unit files
2647 extensions in additional places. Previously, for a unit file name
2648 "foo-bar-baz.service" it would look for dropin files in
2649 "foo-bar-baz.service.d/*.conf". Now, it will also look in
2650 "foo-bar-.service.d/*.conf" and "foo-.service.d/", i.e. at the
2651 service name truncated after all inner dashes. This scheme allows
2652 writing drop-ins easily that apply to a whole set of unit files at
2653 once. It's particularly useful for mount and slice units (as their
2654 naming is prefix based), but is also useful for service and other
2655 units, for packages that install multiple unit files at once,
2656 following a strict naming regime of beginning the unit file name with
2657 the package's name. Two new specifiers are now supported in unit
2658 files to match this: %j and %J are replaced by the part of the unit
2659 name following the last dash.
2661 * Unit files and other configuration files that support specifier
2662 expansion now understand another three new specifiers: %T and %V will
2663 resolve to /tmp and /var/tmp respectively, or whatever temporary
2664 directory has been set for the calling user. %E will expand to either
2665 /etc (for system units) or $XDG_CONFIG_HOME (for user units).
2667 * The ExecStart= lines of unit files are no longer required to
2668 reference absolute paths. If non-absolute paths are specified the
2669 specified binary name is searched within the service manager's
2670 built-in $PATH, which may be queried with 'systemd-path
2671 search-binaries-default'. It's generally recommended to continue to
2672 use absolute paths for all binaries specified in unit files.
2674 * Units gained a new load state "bad-setting", which is used when a
2675 unit file was loaded, but contained fatal errors which prevent it
2676 from being started (for example, a service unit has been defined
2677 lacking both ExecStart= and ExecStop= lines).
2679 * coredumpctl's "gdb" verb has been renamed to "debug", in order to
2680 support alternative debuggers, for example lldb. The old name
2681 continues to be available however, for compatibility reasons. Use the
2682 new --debugger= switch or the $SYSTEMD_DEBUGGER environment variable
2683 to pick an alternative debugger instead of the default gdb.
2685 * systemctl and the other tools will now output escape sequences that
2686 generate proper clickable hyperlinks in various terminal emulators
2687 where useful (for example, in the "systemctl status" output you can
2688 now click on the unit file name to quickly open it in the
2689 editor/viewer of your choice). Note that not all terminal emulators
2690 support this functionality yet, but many do. Unfortunately, the
2691 "less" pager doesn't support this yet, hence this functionality is
2692 currently automatically turned off when a pager is started (which
2693 happens quite often due to auto-paging). We hope to remove this
2694 limitation as soon as "less" learns these escape sequences. This new
2695 behaviour may also be turned off explicitly with the $SYSTEMD_URLIFY
2696 environment variable. For details on these escape sequences see:
2697 https://gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda
2699 * networkd's .network files now support a new IPv6MTUBytes= option for
2700 setting the MTU used by IPv6 explicitly as well as a new MTUBytes=
2701 option in the [Route] section to configure the MTU to use for
2702 specific routes. It also gained support for configuration of the DHCP
2703 "UserClass" option through the new UserClass= setting. It gained
2704 three new options in the new [CAN] section for configuring CAN
2705 networks. The MULTICAST and ALLMULTI interface flags may now be
2706 controlled explicitly with the new Multicast= and AllMulticast=
2709 * networkd will now automatically make use of the kernel's route
2710 expiration feature, if it is available.
2712 * udevd's .link files now support setting the number of receive and
2713 transmit channels, using the RxChannels=, TxChannels=,
2714 OtherChannels=, CombinedChannels= settings.
2716 * Support for UDPSegmentationOffload= has been removed, given its
2717 limited support in hardware, and waning software support.
2719 * networkd's .netdev files now support creating "netdevsim" interfaces.
2721 * PID 1 learnt a new bus call GetUnitByControlGroup() which may be used
2722 to query the unit belonging to a specific kernel control group.
2724 * systemd-analyze gained a new verb "cat-config", which may be used to
2725 dump the contents of any configuration file, with all its matching
2726 drop-in files added in, and honouring the usual search and masking
2727 logic applied to systemd configuration files. For example use
2728 "systemd-analyze cat-config systemd/system.conf" to get the complete
2729 system configuration file of systemd how it would be loaded by PID 1
2730 itself. Similar to this, various tools such as systemd-tmpfiles or
2731 systemd-sysusers, gained a new option "--cat-config", which does the
2732 corresponding operation for their own configuration settings. For
2733 example, "systemd-tmpfiles --cat-config" will now output the full
2734 list of tmpfiles.d/ lines in place.
2736 * timedatectl gained three new verbs: "show" shows bus properties of
2737 systemd-timedated, "timesync-status" shows the current NTP
2738 synchronization state of systemd-timesyncd, and "show-timesync"
2739 shows bus properties of systemd-timesyncd.
2741 * systemd-timesyncd gained a bus interface on which it exposes details
2744 * A new environment variable $SYSTEMD_TIMEDATED_NTP_SERVICES is now
2745 understood by systemd-timedated. It takes a colon-separated list of
2746 unit names of NTP client services. The list is used by
2747 "timedatectl set-ntp".
2749 * systemd-nspawn gained a new --rlimit= switch for setting initial
2750 resource limits for the container payload. There's a new switch
2751 --hostname= to explicitly override the container's hostname. A new
2752 --no-new-privileges= switch may be used to control the
2753 PR_SET_NO_NEW_PRIVS flag for the container payload. A new
2754 --oom-score-adjust= switch controls the OOM scoring adjustment value
2755 for the payload. The new --cpu-affinity= switch controls the CPU
2756 affinity of the container payload. The new --resolv-conf= switch
2757 allows more detailed control of /etc/resolv.conf handling of the
2758 container. Similarly, the new --timezone= switch allows more detailed
2759 control of /etc/localtime handling of the container.
2761 * systemd-detect-virt gained a new --list switch, which will print a
2762 list of all currently known VM and container environments.
2764 * Support for "Portable Services" has been added, see
2765 doc/PORTABLE_SERVICES.md for details. Currently, the support is still
2766 experimental, but this is expected to change soon. Reflecting this
2767 experimental state, the "portablectl" binary is not installed into
2768 /usr/bin yet. The binary has to be called with the full path
2769 /usr/lib/systemd/portablectl instead.
2771 * journalctl's and systemctl's -o switch now knows a new log output
2772 mode "with-unit". The output it generates is very similar to the
2773 regular "short" mode, but displays the unit name instead of the
2774 syslog tag for each log line. Also, the date is shown with timezone
2775 information. This mode is probably more useful than the classic
2776 "short" output mode for most purposes, except where pixel-perfect
2777 compatibility with classic /var/log/messages formatting is required.
2779 * A new --dump-bus-properties switch has been added to the systemd
2780 binary, which may be used to dump all supported D-Bus properties.
2781 (Options which are still supported, but are deprecated, are *not*
2784 * sd-bus gained a set of new calls:
2785 sd_bus_slot_set_floating()/sd_bus_slot_get_floating() may be used to
2786 enable/disable the "floating" state of a bus slot object,
2787 i.e. whether the slot object pins the bus it is allocated for into
2788 memory or if the bus slot object gets disconnected when the bus goes
2789 away. sd_bus_open_with_description(),
2790 sd_bus_open_user_with_description(),
2791 sd_bus_open_system_with_description() may be used to allocate bus
2792 objects and set their description string already during allocation.
2794 * sd-event gained support for watching inotify events from the event
2795 loop, in an efficient way, sharing inotify handles between multiple
2796 users. For this a new function sd_event_add_inotify() has been added.
2798 * sd-event and sd-bus gained support for calling special user-supplied
2799 destructor functions for userdata pointers associated with
2800 sd_event_source, sd_bus_slot, and sd_bus_track objects. For this new
2801 functions sd_bus_slot_set_destroy_callback,
2802 sd_bus_slot_get_destroy_callback, sd_bus_track_set_destroy_callback,
2803 sd_bus_track_get_destroy_callback,
2804 sd_event_source_set_destroy_callback,
2805 sd_event_source_get_destroy_callback have been added.
2807 * The "net.ipv4.tcp_ecn" sysctl will now be turned on by default.
2809 * PID 1 will now automatically reschedule .timer units whenever the
2810 local timezone changes. (They previously got rescheduled
2811 automatically when the system clock changed.)
2813 * New documentation has been added to document cgroups delegation,
2814 portable services and the various code quality tools we have set up:
2816 https://github.com/systemd/systemd/blob/master/docs/CGROUP_DELEGATION.md
2817 https://github.com/systemd/systemd/blob/master/docs/PORTABLE_SERVICES.md
2818 https://github.com/systemd/systemd/blob/master/docs/CODE_QUALITY.md
2820 * The Boot Loader Specification has been added to the source tree.
2822 https://github.com/systemd/systemd/blob/master/docs/BOOT_LOADER_SPECIFICATION.md
2824 While moving it into our source tree we have updated it and further
2825 changes are now accepted through the usual github PR workflow.
2827 * pam_systemd will now look for PAM userdata fields systemd.memory_max,
2828 systemd.tasks_max, systemd.cpu_weight, systemd.io_weight set by
2829 earlier PAM modules. The data in these fields is used to initialize
2830 the session scope's resource properties. Thus external PAM modules
2831 may now configure per-session limits, for example sourced from
2832 external user databases.
2834 * socket units with Accept=yes will now maintain a "refused" counter in
2835 addition to the existing "accepted" counter, counting connections
2836 refused due to the enforced limits.
2838 * The "systemd-path search-binaries-default" command may now be use to
2839 query the default, built-in $PATH PID 1 will pass to the services it
2842 * A new unit file setting PrivateMounts= has been added. It's a boolean
2843 option. If enabled the unit's processes are invoked in their own file
2844 system namespace. Note that this behaviour is also implied if any
2845 other file system namespacing options (such as PrivateTmp=,
2846 PrivateDevices=, ProtectSystem=, …) are used. This option is hence
2847 primarily useful for services that do not use any of the other file
2848 system namespacing options. One such service is systemd-udevd.service
2849 where this is now used by default.
2851 * ConditionSecurity= gained a new value "uefi-secureboot" that is true
2852 when the system is booted in UEFI "secure mode".
2854 * A new unit "system-update-pre.target" is added, which defines an
2855 optional synchronization point for offline system updates, as
2856 implemented by the pre-existing "system-update.target" unit. It
2857 allows ordering services before the service that executes the actual
2858 update process in a generic way.
2860 * Systemd now emits warnings whenever .include syntax is used.
2862 Contributions from: Adam Duskett, Alan Jenkins, Alessandro Casale,
2863 Alexander Kurtz, Alex Gartrell, Anssi Hannula, Arnaud Rebillout, Brian
2864 J. Murrell, Bruno Vernay, Chris Lamb, Chris Lesiak, Christian Brauner,
2865 Christian Hesse, Christian Rebischke, Colin Guthrie, Daniel Dao, Daniel
2866 Lin, Danylo Korostil, Davide Cavalca, David Tardon, Dimitri John
2867 Ledkov, Dmitriy Geels, Douglas Christman, Elia Geretto, emelenas, Emil
2868 Velikov, Evgeny Vereshchagin, Felipe Sateler, Feng Sun, Filipe
2869 Brandenburger, Franck Bui, futpib, Giuseppe Scrivano, Guillem Jover,
2870 guixxx, Hannes Reinecke, Hans de Goede, Harald Hoyer, Henrique Dante de
2871 Almeida, Hiram van Paassen, Ian Miell, Igor Gnatenko, Ivan Shapovalov,
2872 Iwan Timmer, James Cowgill, Jan Janssen, Jan Synacek, Jared Kazimir,
2873 Jérémy Rosen, João Paulo Rechi Vita, Joost Heitbrink, Jui-Chi Ricky
2874 Liang, Jürg Billeter, Kai-Heng Feng, Karol Augustin, Kay Sievers,
2875 Krzysztof Nowicki, Lauri Tirkkonen, Lennart Poettering, Leonard König,
2876 Long Li, Luca Boccassi, Lucas Werkmeister, Marcel Hoppe, Marc
2877 Kleine-Budde, Mario Limonciello, Martin Jansa, Martin Wilck, Mathieu
2878 Malaterre, Matteo F. Vescovi, Matthew McGinn, Matthias-Christian Ott,
2879 Michael Biebl, Michael Olbrich, Michael Prokop, Michal Koutný, Michal
2880 Sekletar, Mike Gilbert, Mikhail Kasimov, Milan Broz, Milan Pässler,
2881 Mladen Pejaković, Muhammet Kara, Nicolas Boichat, Omer Katz, Paride
2882 Legovini, Paul Menzel, Paul Milliken, Pavel Hrdina, Peter A. Bigot,
2883 Peter D'Hoye, Peter Hutterer, Peter Jones, Philip Sequeira, Philip
2884 Withnall, Piotr Drąg, Radostin Stoyanov, Ricardo Salveti de Araujo,
2885 Ronny Chevalier, Rosen Penev, Rubén Suárez Alvarez, Ryan Gonzalez,
2886 Salvo Tomaselli, Sebastian Reichel, Sergey Ptashnick, Sergio Lindo
2887 Mansilla, Stefan Schweter, Stephen Hemminger, Stuart Hayes, Susant
2888 Sahani, Sylvain Plantefève, Thomas H. P. Andersen, Tobias Jungel,
2889 Tomasz Torcz, Vito Caputo, Will Dietz, Will Thompson, Wim van Mourik,
2890 Yu Watanabe, Zbigniew Jędrzejewski-Szmek
2892 — Berlin, 2018-06-22
2896 * The MemoryAccounting= unit property now defaults to on. After
2897 discussions with the upstream control group maintainers we learnt
2898 that the negative impact of cgroup memory accounting on current
2899 kernels is finally relatively minimal, so that it should be safe to
2900 enable this by default without affecting system performance. Besides
2901 memory accounting only task accounting is turned on by default, all
2902 other forms of resource accounting (CPU, IO, IP) remain off for now,
2903 because it's not clear yet that their impact is small enough to move
2904 from opt-in to opt-out. We recommend downstreams to leave memory
2905 accounting on by default if kernel 4.14 or higher is primarily
2906 used. On very resource constrained systems or when support for old
2907 kernels is a necessity, -Dmemory-accounting-default=false can be used
2908 to revert this change.
2910 * rpm scriptlets to update the udev hwdb and rules (%udev_hwdb_update,
2911 %udev_rules_update) and the journal catalog (%journal_catalog_update)
2912 from the upgrade scriptlets of individual packages now do nothing.
2913 Transfiletriggers have been added which will perform those updates
2914 once at the end of the transaction.
2916 Similar transfiletriggers have been added to execute any sysctl.d
2917 and binfmt.d rules. Thus, it should be unnecessary to provide any
2918 scriptlets to execute this configuration from package installation
2921 * systemd-sysusers gained a mode where the configuration to execute is
2922 specified on the command line, but this configuration is not executed
2923 directly, but instead it is merged with the configuration on disk,
2924 and the result is executed. This is useful for package installation
2925 scripts which want to create the user before installing any files on
2926 disk (in case some of those files are owned by that user), while
2927 still allowing local admin overrides.
2929 This functionality is exposed to rpm scriptlets through a new
2930 %sysusers_create_package macro. Old %sysusers_create and
2931 %sysusers_create_inline macros are deprecated.
2933 A transfiletrigger for sysusers.d configuration is now installed,
2934 which means that it should be unnecessary to call systemd-sysusers from
2935 package installation scripts, unless the package installs any files
2936 owned by those newly-created users, in which case
2937 %sysusers_create_package should be used.
2939 * Analogous change has been done for systemd-tmpfiles: it gained a mode
2940 where the command-line configuration is merged with the configuration
2941 on disk. This is exposed as the new %tmpfiles_create_package macro,
2942 and %tmpfiles_create is deprecated. A transfiletrigger is installed
2943 for tmpfiles.d, hence it should be unnecessary to call systemd-tmpfiles
2944 from package installation scripts.
2946 * sysusers.d configuration for a user may now also specify the group
2947 number, in addition to the user number ("u username 123:456"), or
2948 without the user number ("u username -:456").
2950 * Configution items for systemd-sysusers can now be specified as
2951 positional arguments when the new --inline switch is used.
2953 * The login shell of users created through sysusers.d may now be
2954 specified (previously, it was always /bin/sh for root and
2955 /sbin/nologin for other users).
2957 * systemd-analyze gained a new --global switch to look at global user
2958 configuration. It also gained a unit-paths verb to list the unit load
2959 paths that are compiled into systemd (which can be used with
2960 --systemd, --user, or --global).
2962 * udevadm trigger gained a new --settle/-w option to wait for any
2963 triggered events to finish (but just those, and not any other events
2964 which are triggered meanwhile).
2966 * The action that systemd-logind takes when the lid is closed and the
2967 machine is connected to external power can now be configured using
2968 HandleLidSwitchExternalPower= in logind.conf. Previously, this action
2969 was determined by HandleLidSwitch=, and, for backwards compatibility,
2970 is still is, if HandleLidSwitchExternalPower= is not explicitly set.
2972 * journalctl will periodically call sd_journal_process() to make it
2973 resilient against inotify queue overruns when journal files are
2974 rotated very quickly.
2976 * Two new functions in libsystemd — sd_bus_get_n_queued_read and
2977 sd_bus_get_n_queued_write — may be used to check the number of
2978 pending bus messages.
2980 * systemd gained a new
2981 org.freedesktop.systemd1.Manager.AttachProcessesToUnit dbus call
2982 which can be used to migrate foreign processes to scope and service
2983 units. The primary user for this new API is systemd itself: the
2984 systemd --user instance uses this call of the systemd --system
2985 instance to migrate processes if it itself gets the request to
2986 migrate processes and the kernel refuses this due to access
2987 restrictions. Thanks to this "systemd-run --scope --user …" works
2988 again in pure cgroup v2 environments when invoked from the user
2991 * A new TemporaryFileSystem= setting can be used to mask out part of
2992 the real file system tree with tmpfs mounts. This may be combined
2993 with BindPaths= and BindReadOnlyPaths= to hide files or directories
2994 not relevant to the unit, while still allowing some paths lower in
2995 the tree to be accessed.
2997 ProtectHome=tmpfs may now be used to hide user home and runtime
2998 directories from units, in a way that is mostly equivalent to
2999 "TemporaryFileSystem=/home /run/user /root".
3001 * Non-service units are now started with KeyringMode=shared by default.
3002 This means that mount and swapon and other mount tools have access
3003 to keys in the main keyring.
3005 * /sys/fs/bpf is now mounted automatically.
3007 * QNX virtualization is now detected by systemd-detect-virt and may
3008 be used in ConditionVirtualization=.
3010 * IPAccounting= may now be enabled also for slice units.
3012 * A new -Dsplit-bin= build configuration switch may be used to specify
3013 whether bin and sbin directories are merged, or if they should be
3014 included separately in $PATH and various listings of executable
3015 directories. The build configuration scripts will try to autodetect
3016 the proper values of -Dsplit-usr= and -Dsplit-bin= based on build
3017 system, but distributions are encouraged to configure this
3020 * A new -Dok-color= build configuration switch may be used to change
3021 the colour of "OK" status messages.
3023 * UPGRADE ISSUE: serialization of units using JoinsNamespaceOf= with
3024 PrivateNetwork=yes was buggy in previous versions of systemd. This
3025 means that after the upgrade and daemon-reexec, any such units must
3028 * INCOMPATIBILITY: as announced in the NEWS for 237, systemd-tmpfiles
3029 will not exclude read-only files owned by root from cleanup.
3031 Contributions from: Alan Jenkins, Alexander F Rødseth, Alexis Jeandet,
3032 Andika Triwidada, Andrei Gherzan, Ansgar Burchardt, antizealot1337,
3033 Batuhan Osman Taşkaya, Beniamino Galvani, Bill Yodlowsky, Caio Marcelo
3034 de Oliveira Filho, CuBiC, Daniele Medri, Daniel Mouritzen, Daniel
3035 Rusek, Davide Cavalca, Dimitri John Ledkov, Douglas Christman, Evgeny
3036 Vereshchagin, Faalagorn, Filipe Brandenburger, Franck Bui, futpib,
3037 Giacomo Longo, Gunnar Hjalmarsson, Hans de Goede, Hermann Gausterer,
3038 Iago López Galeiras, Jakub Filak, Jan Synacek, Jason A. Donenfeld,
3039 Javier Martinez Canillas, Jérémy Rosen, Lennart Poettering, Lucas
3040 Werkmeister, Mao Huang, Marco Gulino, Michael Biebl, Michael Vogt,
3041 MilhouseVH, Neal Gompa (ニール・ゴンパ), Oleander Reis, Olof Mogren,
3042 Patrick Uiterwijk, Peter Hutterer, Peter Portante, Piotr Drąg, Robert
3043 Antoni Buj Gelonch, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
3044 Fowler, SjonHortensius, snorreflorre, Susant Sahani, Sylvain
3045 Plantefève, Thomas Blume, Thomas Haller, Vito Caputo, Yu Watanabe,
3046 Zbigniew Jędrzejewski-Szmek, Марко М. Костић (Marko M. Kostić)
3048 — Warsaw, 2018-03-05
3052 * Some keyboards come with a zoom see-saw or rocker which until now got
3053 mapped to the Linux "zoomin/out" keys in hwdb. However, these
3054 keycodes are not recognized by any major desktop. They now produce
3055 Up/Down key events so that they can be used for scrolling.
3057 * INCOMPATIBILITY: systemd-tmpfiles' "f" lines changed behaviour
3058 slightly: previously, if an argument was specified for lines of this
3059 type (i.e. the right-most column was set) this string was appended to
3060 existing files each time systemd-tmpfiles was run. This behaviour was
3061 different from what the documentation said, and not particularly
3062 useful, as repeated systemd-tmpfiles invocations would not be
3063 idempotent and grow such files without bounds. With this release
3064 behaviour has been altered to match what the documentation says:
3065 lines of this type only have an effect if the indicated files don't
3066 exist yet, and only then the argument string is written to the file.
3068 * FUTURE INCOMPATIBILITY: In systemd v238 we intend to slightly change
3069 systemd-tmpfiles behaviour: previously, read-only files owned by root
3070 were always excluded from the file "aging" algorithm (i.e. the
3071 automatic clean-up of directories like /tmp based on
3072 atime/mtime/ctime). We intend to drop this restriction, and age files
3073 by default even when owned by root and read-only. This behaviour was
3074 inherited from older tools, but there have been requests to remove
3075 it, and it's not obvious why this restriction was made in the first
3076 place. Please speak up now, if you are aware of software that reqires
3077 this behaviour, otherwise we'll remove the restriction in v238.
3079 * A new environment variable $SYSTEMD_OFFLINE is now understood by
3080 systemctl. It takes a boolean argument. If on, systemctl assumes it
3081 operates on an "offline" OS tree, and will not attempt to talk to the
3082 service manager. Previously, this mode was implicitly enabled if a
3083 chroot() environment was detected, and this new environment variable
3084 now provides explicit control.
3086 * .path and .socket units may now be created transiently, too.
3087 Previously only service, mount, automount and timer units were
3088 supported as transient units. The systemd-run tool has been updated
3089 to expose this new functionality, you may hence use it now to bind
3090 arbitrary commands to path or socket activation on-the-fly from the
3091 command line. Moreover, almost all properties are now exposed for the
3092 unit types that already supported transient operation.
3094 * The systemd-mount command gained support for a new --owner= parameter
3095 which takes a user name, which is then resolved and included in uid=
3096 and gid= mount options string of the file system to mount.
3098 * A new unit condition ConditionControlGroupController= has been added
3099 that checks whether a specific cgroup controller is available.
3101 * Unit files, udev's .link files, and systemd-networkd's .netdev and
3102 .network files all gained support for a new condition
3103 ConditionKernelVersion= for checking against specific kernel
3106 * In systemd-networkd, the [IPVLAN] section in .netdev files gained
3107 support for configuring device flags in the Flags= setting. In the
3108 same files, the [Tunnel] section gained support for configuring
3109 AllowLocalRemote=. The [Route] section in .network files gained
3110 support for configuring InitialCongestionWindow=,
3111 InitialAdvertisedReceiveWindow= and QuickAck=. The [DHCP] section now
3112 understands RapidCommit=.
3114 * systemd-networkd's DHCPv6 support gained support for Prefix
3117 * sd-bus gained support for a new "watch-bind" feature. When this
3118 feature is enabled, an sd_bus connection may be set up to connect to
3119 an AF_UNIX socket in the file system as soon as it is created. This
3120 functionality is useful for writing early-boot services that
3121 automatically connect to the system bus as soon as it is started,
3122 without ugly time-based polling. systemd-networkd and
3123 systemd-resolved have been updated to make use of this
3124 functionality. busctl exposes this functionality in a new
3125 --watch-bind= command line switch.
3127 * sd-bus will now optionally synthesize a local "Connected" signal as
3128 soon as a D-Bus connection is set up fully. This message mirrors the
3129 already existing "Disconnected" signal which is synthesized when the
3130 connection is terminated. This signal is generally useful but
3131 particularly handy in combination with the "watch-bind" feature
3132 described above. Synthesizing of this message has to be requested
3133 explicitly through the new API call sd_bus_set_connected_signal(). In
3134 addition a new call sd_bus_is_ready() has been added that checks
3135 whether a connection is fully set up (i.e. between the "Connected" and
3136 "Disconnected" signals).
3138 * sd-bus gained two new calls sd_bus_request_name_async() and
3139 sd_bus_release_name_async() for asynchronously registering bus
3140 names. Similar, there is now sd_bus_add_match_async() for installing
3141 a signal match asynchronously. All of systemd's own services have
3142 been updated to make use of these calls. Doing these operations
3143 asynchronously has two benefits: it reduces the risk of deadlocks in
3144 case of cyclic dependencies between bus services, and it speeds up
3145 service initialization since synchronization points for bus
3146 round-trips are removed.
3148 * sd-bus gained two new calls sd_bus_match_signal() and
3149 sd_bus_match_signal_async(), which are similar to sd_bus_add_match()
3150 and sd_bus_add_match_async() but instead of taking a D-Bus match
3151 string take match fields as normal function parameters.
3153 * sd-bus gained two new calls sd_bus_set_sender() and
3154 sd_bus_message_set_sender() for setting the sender name of outgoing
3155 messages (either for all outgoing messages or for just one specific
3156 one). These calls are only useful in direct connections as on
3157 brokered connections the broker fills in the sender anyway,
3158 overwriting whatever the client filled in.
3160 * sd-event gained a new pseudo-handle that may be specified on all API
3161 calls where an "sd_event*" object is expected: SD_EVENT_DEFAULT. When
3162 used this refers to the default event loop object of the calling
3163 thread. Note however that this does not implicitly allocate one —
3164 which has to be done prior by using sd_event_default(). Similarly
3165 sd-bus gained three new pseudo-handles SD_BUS_DEFAULT,
3166 SD_BUS_DEFAULT_USER, SD_BUS_DEFAULT_SYSTEM that may be used to refer
3167 to the default bus of the specified type of the calling thread. Here
3168 too this does not implicitly allocate bus connection objects, this
3169 has to be done prior with sd_bus_default() and friends.
3171 * sd-event gained a new call pair
3172 sd_event_source_{get|set}_io_fd_own(). This may be used to request
3173 automatic closure of the file descriptor an IO event source watches
3174 when the event source is destroyed.
3176 * systemd-networkd gained support for natively configuring WireGuard
3179 * In previous versions systemd synthesized user records both for the
3180 "nobody" (UID 65534) and "root" (UID 0) users in nss-systemd and
3181 internally. In order to simplify distribution-wide renames of the
3182 "nobody" user (like it is planned in Fedora: nfsnobody → nobody), a
3183 new transitional flag file has been added: if
3184 /etc/systemd/dont-synthesize-nobody exists synthesizing of the 65534
3185 user and group record within the systemd codebase is disabled.
3187 * systemd-notify gained a new --uid= option for selecting the source
3188 user/UID to use for notification messages sent to the service
3191 * journalctl gained a new --grep= option to list only entries in which
3192 the message matches a certain pattern. By default matching is case
3193 insensitive if the pattern is lowercase, and case sensitive
3194 otherwise. Option --case-sensitive=yes|no can be used to override
3195 this an specify case sensitivity or case insensitivity.
3197 * There's now a "systemd-analyze service-watchdogs" command for printing
3198 the current state of the service runtime watchdog, and optionally
3199 enabling or disabling the per-service watchdogs system-wide if given a
3200 boolean argument (i.e. the concept you configure in WatchdogSec=), for
3201 debugging purposes. There's also a kernel command line option
3202 systemd.service_watchdogs= for controlling the same.
3204 * Two new "log-level" and "log-target" options for systemd-analyze were
3205 added that merge the now deprecated get-log-level, set-log-level and
3206 get-log-target, set-log-target pairs. The deprecated options are still
3207 understood for backwards compatibility. The two new options print the
3208 current value when no arguments are given, and set them when a
3209 level/target is given as an argument.
3211 * sysusers.d's "u" lines now optionally accept both a UID and a GID
3212 specification, separated by a ":" character, in order to create users
3213 where UID and GID do not match.
3215 Contributions from: Adam Duskett, Alan Jenkins, Alexander Kuleshov,
3216 Alexis Deruelle, Andrew Jeddeloh, Armin Widegreen, Batuhan Osman
3217 Taşkaya, Björn Esser, bleep_blop, Bruce A. Johnson, Chris Down, Clinton
3218 Roy, Colin Walters, Daniel Rusek, Dimitri John Ledkov, Dmitry Rozhkov,
3219 Evgeny Vereshchagin, Ewout van Mansom, Felipe Sateler, Franck Bui,
3220 Frantisek Sumsal, George Gaydarov, Gianluca Boiano, Hans-Christian
3221 Noren Egtvedt, Hans de Goede, Henrik Grindal Bakken, Jan Alexander
3222 Steffens, Jan Klötzke, Jason A. Donenfeld, jdkbx, Jérémy Rosen,
3223 Jerónimo Borque, John Lin, John Paul Herold, Jonathan Rudenberg, Jörg
3224 Thalheim, Ken (Bitsko) MacLeod, Larry Bernstone, Lennart Poettering,
3225 Lucas Werkmeister, Maciej S. Szmigiero, Marek Čermák, Martin Pitt,
3226 Mathieu Malaterre, Matthew Thode, Matthias-Christian Ott, Max Harmathy,
3227 Michael Biebl, Michael Vogt, Michal Koutný, Michal Sekletar, Michał
3228 Szczepański, Mike Gilbert, Nathaniel McCallum, Nicolas Chauvet, Olaf
3229 Hering, Olivier Schwander, Patrik Flykt, Paul Cercueil, Peter Hutterer,
3230 Piotr Drąg, Raphael Vogelgsang, Reverend Homer, Robert Kolchmeyer,
3231 Samuel Dionne-Riel, Sergey Ptashnick, Shawn Landden, Susant Sahani,
3232 Sylvain Plantefève, Thomas H. P. Andersen, Thomas Huth, Tomasz
3233 Bachorski, Vladislav Vishnyakov, Wieland Hoffmann, Yu Watanabe, Zachary
3234 Winnerman, Zbigniew Jędrzejewski-Szmek, Дамјан Георгиевски, Дилян
3241 * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
3242 in v235 has been extended to also set the dummy.ko module option
3243 numdummies=0, preventing the kernel from automatically creating
3244 dummy0. All dummy interfaces must now be explicitly created.
3246 * Unknown '%' specifiers in configuration files are now rejected. This
3247 applies to units and tmpfiles.d configuration. Any percent characters
3248 that are followed by a letter or digit that are not supposed to be
3249 interpreted as the beginning of a specifier should be escaped by
3250 doubling ("%%"). (So "size=5%" is still accepted, as well as
3251 "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
3252 valid specifiers today.)
3254 * systemd-resolved now maintains a new dynamic
3255 /run/systemd/resolve/stub-resolv.conf compatibility file. It is
3256 recommended to make /etc/resolv.conf a symlink to it. This file
3257 points at the systemd-resolved stub DNS 127.0.0.53 resolver and
3258 includes dynamically acquired search domains, achieving more correct
3259 DNS resolution by software that bypasses local DNS APIs such as NSS.
3261 * The "uaccess" udev tag has been dropped from /dev/kvm and
3262 /dev/dri/renderD*. These devices now have the 0666 permissions by
3263 default (but this may be changed at build-time). /dev/dri/renderD*
3264 will now be owned by the "render" group along with /dev/kfd.
3266 * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
3267 systemd-journal-gatewayd.service and
3268 systemd-journal-upload.service. This means "nss-systemd" must be
3269 enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
3270 services are resolved properly.
3272 * In /etc/fstab two new mount options are now understood:
3273 x-systemd.makefs and x-systemd.growfs. The former has the effect that
3274 the configured file system is formatted before it is mounted, the
3275 latter that the file system is resized to the full block device size
3276 after it is mounted (i.e. if the file system is smaller than the
3277 partition it resides on, it's grown). This is similar to the fsck
3278 logic in /etc/fstab, and pulls in systemd-makefs@.service and
3279 systemd-growfs@.service as necessary, similar to
3280 systemd-fsck@.service. Resizing is currently only supported on ext4
3283 * In systemd-networkd, the IPv6 RA logic now optionally may announce
3284 DNS server and domain information.
3286 * Support for the LUKS2 on-disk format for encrypted partitions has
3287 been added. This requires libcryptsetup2 during compilation and
3290 * The systemd --user instance will now signal "readiness" when its
3291 basic.target unit has been reached, instead of when the run queue ran
3292 empty for the first time.
3294 * Tmpfiles.d with user configuration are now also supported.
3295 systemd-tmpfiles gained a new --user switch, and snippets placed in
3296 ~/.config/user-tmpfiles.d/ and corresponding directories will be
3297 executed by systemd-tmpfiles --user running in the new
3298 systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
3299 running in the user session.
3301 * Unit files and tmpfiles.d snippets learnt three new % specifiers:
3302 %S resolves to the top-level state directory (/var/lib for the system
3303 instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
3304 top-level cache directory (/var/cache for the system instance,
3305 $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
3306 logs directory (/var/log for the system instance,
3307 $XDG_CONFIG_HOME/log/ for the user instance). This matches the
3308 existing %t specifier, that resolves to the top-level runtime
3309 directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
3312 * journalctl learnt a new parameter --output-fields= for limiting the
3313 set of journal fields to output in verbose and JSON output modes.
3315 * systemd-timesyncd's configuration file gained a new option
3316 RootDistanceMaxSec= for setting the maximum root distance of servers
3317 it'll use, as well as the new options PollIntervalMinSec= and
3318 PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
3320 * bootctl gained a new command "list" for listing all available boot
3321 menu items on systems that follow the boot loader specification.
3323 * systemctl gained a new --dry-run switch that shows what would be done
3324 instead of doing it, and is currently supported by the shutdown and
3327 * ConditionSecurity= can now detect the TOMOYO security module.
3329 * Unit file [Install] sections are now also respected in unit drop-in
3330 files. This is intended to be used by drop-ins under /usr/lib/.
3332 * systemd-firstboot may now also set the initial keyboard mapping.
3334 * Udev "changed" events for devices which are exposed as systemd
3335 .device units are now propagated to units specified in
3336 ReloadPropagatedFrom= as reload requests.
3338 * If a udev device has a SYSTEMD_WANTS= property containing a systemd
3339 unit template name (i.e. a name in the form of 'foobar@.service',
3340 without the instance component between the '@' and - the '.'), then
3341 the escaped sysfs path of the device is automatically used as the
3344 * SystemCallFilter= in unit files has been extended so that an "errno"
3345 can be specified individually for each system call. Example:
3346 SystemCallFilter=~uname:EILSEQ.
3348 * The cgroup delegation logic has been substantially updated. Delegate=
3349 now optionally takes a list of controllers (instead of a boolean, as
3350 before), which lists the controllers to delegate at least.
3352 * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
3354 * A new LogLevelMax= setting configures the maximum log level any
3355 process of the service may log at (i.e. anything with a lesser
3356 priority than what is specified is automatically dropped). A new
3357 LogExtraFields= setting allows configuration of additional journal
3358 fields to attach to all log records generated by any of the unit's
3361 * New StandardInputData= and StandardInputText= settings along with the
3362 new option StandardInput=data may be used to configure textual or
3363 binary data that shall be passed to the executed service process via
3364 standard input, encoded in-line in the unit file.
3366 * StandardInput=, StandardOutput= and StandardError= may now be used to
3367 connect stdin/stdout/stderr of executed processes directly with a
3368 file or AF_UNIX socket in the file system, using the new "file:" option.
3370 * A new unit file option CollectMode= has been added, that allows
3371 tweaking the garbage collection logic for units. It may be used to
3372 tell systemd to garbage collect units that have failed automatically
3373 (normally it only GCs units that exited successfully). systemd-run
3374 and systemd-mount expose this new functionality with a new -G option.
3376 * "machinectl bind" may now be used to bind mount non-directories
3377 (i.e. regularfiles, devices, fifos, sockets).
3379 * systemd-analyze gained a new verb "calendar" for validating and
3380 testing calendar time specifications to use for OnCalendar= in timer
3381 units. Besides validating the expression it will calculate the next
3382 time the specified expression would elapse.
3384 * In addition to the pre-existing FailureAction= unit file setting
3385 there's now SuccessAction=, for configuring a shutdown action to
3386 execute when a unit completes successfully. This is useful in
3387 particular inside containers that shall terminate after some workload
3388 has been completed. Also, both options are now supported for all unit
3389 types, not just services.
3391 * networkds's IP rule support gained two new options
3392 IncomingInterface= and OutgoingInterface= for configuring the incoming
3393 and outgoing interfaces of configured rules. systemd-networkd also
3394 gained support for "vxcan" network devices.
3396 * networkd gained a new setting RequiredForOnline=, taking a
3397 boolean. If set, systemd-wait-online will take it into consideration
3398 when determining that the system is up, otherwise it will ignore the
3399 interface for this purpose.
3401 * The sd_notify() protocol gained support for a new operation: with
3402 FDSTOREREMOVE=1 file descriptors may be removed from the per-service
3403 store again, ahead of POLLHUP or POLLERR when they are removed
3406 * A new document doc/UIDS-GIDS.md has been added to the source tree,
3407 that documents the UID/GID range and assignment assumptions and
3408 requirements of systemd.
3410 * The watchdog device PID 1 will ping may now be configured through the
3411 WatchdogDevice= configuration file setting, or by setting the
3412 systemd.watchdog_service= kernel commandline option.
3414 * systemd-resolved's gained support for registering DNS-SD services on
3415 the local network using MulticastDNS. Services may either be
3416 registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
3417 the same dir below /run, /usr/lib), or through its D-Bus API.
3419 * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
3420 extend the effective start, runtime, and stop time. The service must
3421 continue to send EXTEND_TIMEOUT_USEC within the period specified to
3422 prevent the service manager from making the service as timedout.
3424 * systemd-resolved's DNSSEC support gained support for RFC 8080
3425 (Ed25519 keys and signatures).
3427 * The systemd-resolve command line tool gained a new set of options
3428 --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
3429 --set-nta= and --revert to configure per-interface DNS configuration
3430 dynamically during runtime. It's useful for pushing DNS information
3431 into systemd-resolved from DNS hook scripts that various interface
3432 managing software supports (such as pppd).
3434 * systemd-nspawn gained a new --network-namespace-path= command line
3435 option, which may be used to make a container join an existing
3436 network namespace, by specifying a path to a "netns" file.
3438 Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
3439 Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
3440 Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
3441 Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
3442 John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
3443 Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
3444 Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
3445 Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
3446 Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
3447 Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
3448 Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
3449 Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
3450 Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
3451 Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
3452 Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
3453 Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
3454 Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
3455 Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
3456 Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
3457 Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
3458 Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
3459 Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
3460 Jędrzejewski-Szmek, Zeal Jagannatha
3462 — Berlin, 2017-12-14
3466 * INCOMPATIBILITY: systemd-logind.service and other long-running
3467 services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
3468 communication with the outside. This generally improves security of
3469 the system, and is in almost all cases a safe and good choice, as
3470 these services do not and should not provide any network-facing
3471 functionality. However, systemd-logind uses the glibc NSS API to
3472 query the user database. This creates problems on systems where NSS
3473 is set up to directly consult network services for user database
3474 lookups. In particular, this creates incompatibilities with the
3475 "nss-nis" module, which attempts to directly contact the NIS/YP
3476 network servers it is configured for, and will now consistently
3477 fail. In such cases, it is possible to turn off IP sandboxing for
3478 systemd-logind.service (set IPAddressDeny= in its [Service] section
3479 to the empty string, via a .d/ unit file drop-in). Downstream
3480 distributions might want to update their nss-nis packaging to include
3481 such a drop-in snippet, accordingly, to hide this incompatibility
3482 from the user. Another option is to make use of glibc's nscd service
3483 to proxy such network requests through a privilege-separated, minimal
3484 local caching daemon, or to switch to more modern technologies such
3485 sssd, whose NSS hook-ups generally do not involve direct network
3486 access. In general, we think it's definitely time to question the
3487 implementation choices of nss-nis, i.e. whether it's a good idea
3488 today to embed a network-facing loadable module into all local
3489 processes that need to query the user database, including the most
3490 trivial and benign ones, such as "ls". For more details about
3491 IPAddressDeny= see below.
3493 * A new modprobe.d drop-in is now shipped by default that sets the
3494 bonding module option max_bonds=0. This overrides the kernel default,
3495 to avoid conflicts and ambiguity as to whether or not bond0 should be
3496 managed by systemd-networkd or not. This resolves multiple issues
3497 with bond0 properties not being applied, when bond0 is configured
3498 with systemd-networkd. Distributors may choose to not package this,
3499 however in that case users will be prevented from correctly managing
3500 bond0 interface using systemd-networkd.
3502 * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
3503 which print the logging level and target of the system manager. They
3504 complement the existing "set-log-level" and "set-log-target" verbs
3505 used to change those values.
3507 * journald.conf gained a new boolean setting ReadKMsg= which defaults
3508 to on. If turned off kernel log messages will not be read by
3509 systemd-journald or included in the logs. It also gained a new
3510 setting LineMax= for configuring the maximum line length in
3511 STDOUT/STDERR log streams. The new default for this value is 48K, up
3512 from the previous hardcoded 2048.
3514 * A new unit setting RuntimeDirectoryPreserve= has been added, which
3515 allows more detailed control of what to do with a runtime directory
3516 configured with RuntimeDirectory= (i.e. a directory below /run or
3517 $XDG_RUNTIME_DIR) after a unit is stopped.
3519 * The RuntimeDirectory= setting for units gained support for creating
3520 deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
3521 one top-level directory.
3523 * Units gained new options StateDirectory=, CacheDirectory=,
3524 LogsDirectory= and ConfigurationDirectory= which are closely related
3525 to RuntimeDirectory= but manage per-service directories below
3526 /var/lib, /var/cache, /var/log and /etc. By making use of them it is
3527 possible to write unit files which when activated automatically gain
3528 properly owned service specific directories in these locations, thus
3529 making unit files self-contained and increasing compatibility with
3530 stateless systems and factory reset where /etc or /var are
3531 unpopulated at boot. Matching these new settings there's also
3532 StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
3533 ConfigurationDirectoryMode= for configuring the access mode of these
3534 directories. These settings are particularly useful in combination
3535 with DynamicUser=yes as they provide secure, properly-owned,
3536 writable, and stateful locations for storage, excluded from the
3537 sandbox that such services live in otherwise.
3539 * Automake support has been removed from this release. systemd is now
3542 * systemd-journald will now aggressively cache client metadata during
3543 runtime, speeding up log write performance under pressure. This comes
3544 at a small price though: as much of the metadata is read
3545 asynchronously from /proc/ (and isn't implicitly attached to log
3546 datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
3547 metadata stored alongside a log entry might be slightly
3548 out-of-date. Previously it could only be slightly newer than the log
3549 message. The time window is small however, and given that the kernel
3550 is unlikely to be improved anytime soon in this regard, this appears
3553 * nss-myhostname/systemd-resolved will now by default synthesize an
3554 A/AAAA resource record for the "_gateway" hostname, pointing to the
3555 current default IP gateway. Previously it did that for the "gateway"
3556 name, hampering adoption, as some distributions wanted to leave that
3557 hostname open for local use. The old behaviour may still be
3558 requested at build time.
3560 * systemd-networkd's [Address] section in .network files gained a new
3561 Scope= setting for configuring the IP address scope. The [Network]
3562 section gained a new boolean setting ConfigureWithoutCarrier= that
3563 tells systemd-networkd to ignore link sensing when configuring the
3564 device. The [DHCP] section gained a new Anonymize= boolean option for
3565 turning on a number of options suggested in RFC 7844. A new
3566 [RoutingPolicyRule] section has been added for configuring the IP
3567 routing policy. The [Route] section has gained support for a new
3568 Type= setting which permits configuring
3569 blackhole/unreachable/prohibit routes.
3571 * The [VRF] section in .netdev files gained a new Table= setting for
3572 configuring the routing table to use. The [Tunnel] section gained a
3573 new Independent= boolean field for configuring tunnels independent of
3574 an underlying network interface. The [Bridge] section gained a new
3575 GroupForwardMask= option for configuration of propagation of link
3576 local frames between bridge ports.
3578 * The WakeOnLan= setting in .link files gained support for a number of
3579 new modes. A new TCP6SegmentationOffload= setting has been added for
3580 configuring TCP/IPv6 hardware segmentation offload.
3582 * The IPv6 RA sender implementation may now optionally send out RDNSS
3583 and RDNSSL records to supply DNS configuration to peers.
3585 * systemd-nspawn gained support for a new --system-call-filter= command
3586 line option for adding and removing entries in the default system
3587 call filter it applies. Moreover systemd-nspawn has been changed to
3588 implement a system call allow list instead of a deny list.
3590 * systemd-run gained support for a new --pipe command line option. If
3591 used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
3592 are directly passed on to the activated transient service
3593 executable. This allows invoking arbitrary processes as systemd
3594 services (for example to take benefit of dependency management,
3595 accounting management, resource management or log management that is
3596 done automatically for services) — while still allowing them to be
3597 integrated in a classic UNIX shell pipeline.
3599 * When a service sends RELOAD=1 via sd_notify() and reload propagation
3600 using ReloadPropagationTo= is configured, a reload is now propagated
3601 to configured units. (Previously this was only done on explicitly
3602 requested reloads, using "systemctl reload" or an equivalent
3605 * For each service unit a restart counter is now kept: it is increased
3606 each time the service is restarted due to Restart=, and may be
3607 queried using "systemctl show -p NRestarts …".
3609 * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
3610 @signal and @timer have been added, for usage with SystemCallFilter=
3611 in unit files and the new --system-call-filter= command line option
3612 of systemd-nspawn (see above).
3614 * ExecStart= lines in unit files gained two new modifiers: when a
3615 command line is prefixed with "!" the command will be executed as
3616 configured, except for the credentials applied by
3617 setuid()/setgid()/setgroups(). It is very similar to the pre-existing
3618 "+", but does still apply namespacing options unlike "+". There's
3619 also "!!" now, which is mostly identical, but becomes a NOP on
3620 systems that support ambient capabilities. This is useful to write
3621 unit files that work with ambient capabilities where possible but
3622 automatically fall back to traditional privilege dropping mechanisms
3623 on systems where this is not supported.
3625 * ListenNetlink= settings in socket units now support RDMA netlink
3628 * A new unit file setting LockPersonality= has been added which permits
3629 locking down the chosen execution domain ("personality") of a service
3632 * A new special target "getty-pre.target" has been added, which is
3633 ordered before all text logins, and may be used to order services
3634 before textual logins acquire access to the console.
3636 * systemd will now attempt to load the virtio-rng.ko kernel module very
3637 early on if a VM environment supporting this is detected. This should
3638 improve entropy during early boot in virtualized environments.
3640 * A _netdev option is now supported in /etc/crypttab that operates in a
3641 similar way as the same option in /etc/fstab: it permits configuring
3642 encrypted devices that need to be ordered after the network is up.
3643 Following this logic, two new special targets
3644 remote-cryptsetup-pre.target and remote-cryptsetup.target have been
3645 added that are to cryptsetup.target what remote-fs.target and
3646 remote-fs-pre.target are to local-fs.target.
3648 * Service units gained a new UnsetEnvironment= setting which permits
3649 unsetting specific environment variables for services that are
3650 normally passed to it (for example in order to mask out locale
3651 settings for specific services that can't deal with it).
3653 * Units acquired a new boolean option IPAccounting=. When turned on, IP
3654 traffic accounting (packet count as well as byte count) is done for
3655 the service, and shown as part of "systemctl status" or "systemd-run
3658 * Service units acquired two new options IPAddressAllow= and
3659 IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
3660 for configuring a simple IP access control list for all sockets of
3661 the unit. These options are available also on .slice and .socket
3662 units, permitting flexible access list configuration for individual
3663 services as well as groups of services (as defined by a slice unit),
3664 including system-wide. Note that IP ACLs configured this way are
3665 enforced on every single IPv4 and IPv6 socket created by any process
3666 of the service unit, and apply to ingress as well as egress traffic.
3668 * If CPUAccounting= or IPAccounting= is turned on for a unit a new
3669 structured log message is generated each time the unit is stopped,
3670 containing information about the consumed resources of this
3673 * A new setting KeyringMode= has been added to unit files, which may be
3674 used to control how the kernel keyring is set up for executed
3677 * "systemctl poweroff", "systemctl reboot", "systemctl halt",
3678 "systemctl kexec" and "systemctl exit" are now always asynchronous in
3679 behaviour (that is: these commands return immediately after the
3680 operation was enqueued instead of waiting for the operation to
3681 complete). Previously, "systemctl poweroff" and "systemctl reboot"
3682 were asynchronous on systems using systemd-logind (i.e. almost
3683 always, and like they were on sysvinit), and the other three commands
3684 were unconditionally synchronous. With this release this is cleaned
3685 up, and callers will see the same asynchronous behaviour on all
3686 systems for all five operations.
3688 * systemd-logind gained new Halt() and CanHalt() bus calls for halting
3691 * .timer units now accept calendar specifications in other timezones
3692 than UTC or the local timezone.
3694 * The tmpfiles snippet var.conf has been changed to create
3695 /var/log/btmp with access mode 0660 instead of 0600. It was owned by
3696 the "utmp" group already, and it appears to be generally understood
3697 that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
3698 databases. Previously this was implemented correctly for all these
3699 databases excepts btmp, which has been opened up like this now
3700 too. Note that while the other databases are world-readable
3701 (i.e. 0644), btmp is not and remains more restrictive.
3703 * The systemd-resolve tool gained a new --reset-server-features
3704 switch. When invoked like this systemd-resolved will forget
3705 everything it learnt about the features supported by the configured
3706 upstream DNS servers, and restarts the feature probing logic on the
3707 next resolver look-up for them at the highest feature level
3710 * The status dump systemd-resolved sends to the logs upon receiving
3711 SIGUSR1 now also includes information about all DNS servers it is
3712 configured to use, and the features levels it probed for them.
3714 Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
3715 Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
3716 Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
3717 Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
3718 Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
3719 Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
3720 ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
3721 Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
3722 Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
3723 John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
3724 Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
3725 Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
3726 Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
3727 Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
3728 Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
3729 Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
3730 Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
3731 Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
3732 Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
3733 Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
3735 — Berlin, 2017-10-06
3739 * Meson is now supported as build system in addition to Automake. It is
3740 our plan to remove Automake in one of our next releases, so that
3741 Meson becomes our exclusive build system. Hence, please start using
3742 the Meson build system in your downstream packaging. There's plenty
3743 of documentation around how to use Meson, the extremely brief
3746 ./autogen.sh && ./configure && make && sudo make install
3750 meson build && ninja -C build && sudo ninja -C build install
3752 * Unit files gained support for a new JobRunningTimeoutUSec= setting,
3753 which permits configuring a timeout on the time a job is
3754 running. This is particularly useful for setting timeouts on jobs for
3757 * Unit files gained two new options ConditionUser= and ConditionGroup=
3758 for conditionalizing units based on the identity of the user/group
3759 running a systemd user instance.
3761 * systemd-networkd now understands a new FlowLabel= setting in the
3762 [VXLAN] section of .network files, as well as a Priority= in
3763 [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
3764 and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
3765 gained support for configuration of GENEVE links, and IPv6 address
3766 labels. The [Network] section gained the new IPv6ProxyNDP= setting.
3768 * .link files now understand a new Port= setting.
3770 * systemd-networkd's DHCP support gained support for DHCP option 119
3771 (domain search list).
3773 * systemd-networkd gained support for serving IPv6 address ranges using
3774 the Router Advertisement protocol. The new .network configuration
3775 section [IPv6Prefix] may be used to configure the ranges to
3776 serve. This is implemented based on a new, minimal, native server
3777 implementation of RA.
3779 * journalctl's --output= switch gained support for a new parameter
3780 "short-iso-precise" for a mode where timestamps are shown as precise
3783 * systemd-udevd's "net_id" builtin may now generate stable network
3784 interface names from IBM PowerVM VIO devices as well as ACPI platform
3787 * MulticastDNS support in systemd-resolved may now be explicitly
3788 enabled/disabled using the new MulticastDNS= configuration file
3791 * systemd-resolved may now optionally use libidn2 instead of the libidn
3792 for processing internationalized domain names. Support for libidn2
3793 should be considered experimental and should not be enabled by
3796 * "machinectl pull-tar" and related call may now do verification of
3797 downloaded images using SUSE-style .sha256 checksum files in addition
3798 to the already existing support for validating using Ubuntu-style
3801 * sd-bus gained support for a new sd_bus_message_appendv() call which
3802 is va_list equivalent of sd_bus_message_append().
3804 * sd-boot gained support for validating images using SHIM/MOK.
3806 * The SMACK code learnt support for "onlycap".
3808 * systemd-mount --umount is now much smarter in figuring out how to
3809 properly unmount a device given its mount or device path.
3811 * The code to call libnss_dns as a fallback from libnss_resolve when
3812 the communication with systemd-resolved fails was removed. This
3813 fallback was redundant and interfered with the [!UNAVAIL=return]
3814 suffix. See nss-resolve(8) for the recommended configuration.
3816 * systemd-logind may now be restarted without losing state. It stores
3817 the file descriptors for devices it manages in the system manager
3818 using the FDSTORE= mechanism. Please note that further changes in
3819 other components may be required to make use of this (for example
3820 Xorg has code to listen for stops of systemd-logind and terminate
3821 itself when logind is stopped or restarted, in order to avoid using
3822 stale file descriptors for graphical devices, which is now
3823 counterproductive and must be reverted in order for restarts of
3824 systemd-logind to be safe. See
3825 https://cgit.freedesktop.org/xorg/xserver/commit/?id=dc48bd653c7e101.)
3827 * All kernel-install plugins are called with the environment variable
3828 KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
3829 /etc/machine-id. If the machine ID could not be determined,
3830 $KERNEL_INSTALL_MACHINE_ID will be empty. Plugins should not put
3831 anything in the entry directory (passed as the second argument) if
3832 $KERNEL_INSTALL_MACHINE_ID is empty. For backwards compatibility, a
3833 temporary directory is passed as the entry directory and removed
3834 after all the plugins exit.
3836 Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
3837 Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
3838 Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
3839 Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
3840 Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
3841 Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
3842 Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
3843 Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
3844 Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
3845 Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
3846 hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
3847 Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
3848 Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
3849 Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
3850 Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
3851 Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
3852 Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
3853 Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
3854 Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
3855 Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
3856 Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
3857 Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
3858 Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
3859 Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
3860 Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
3861 H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
3862 Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
3863 userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
3864 Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
3867 — Berlin, 2017-07-12
3871 * The "hybrid" control group mode has been modified to improve
3872 compatibility with "legacy" cgroups-v1 setups. Specifically, the
3873 "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
3874 "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
3875 cgroups-v1 hierarchy), the only externally visible change being that
3876 the cgroups-v2 hierarchy is also mounted, to
3877 /sys/fs/cgroup/unified. This should provide a large degree of
3878 compatibility with "legacy" cgroups-v1, while taking benefit of the
3879 better management capabilities of cgroups-v2.
3881 * The default control group setup mode may be selected both a boot-time
3882 via a set of kernel command line parameters (specifically:
3883 systemd.unified_cgroup_hierarchy= and
3884 systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
3885 default selected on the configure command line
3886 (--with-default-hierarchy=). The upstream default is "hybrid"
3887 (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
3888 this will change in a future systemd version to be "unified" (pure
3889 cgroups-v2 mode). The third option for the compile time option is
3890 "legacy", to enter pure cgroups-v1 mode. We recommend downstream
3891 distributions to default to "hybrid" mode for release distributions,
3892 starting with v233. We recommend "unified" for development
3893 distributions (specifically: distributions such as Fedora's rawhide)
3894 as that's where things are headed in the long run. Use "legacy" for
3895 greatest stability and compatibility only.
3897 * Note one current limitation of "unified" and "hybrid" control group
3898 setup modes: the kernel currently does not permit the systemd --user
3899 instance (i.e. unprivileged code) to migrate processes between two
3900 disconnected cgroup subtrees, even if both are managed and owned by
3901 the user. This effectively means "systemd-run --user --scope" doesn't
3902 work when invoked from outside of any "systemd --user" service or
3903 scope. Specifically, it is not supported from session scopes. We are
3904 working on fixing this in a future systemd version. (See #3388 for
3905 further details about this.)
3907 * DBus policy files are now installed into /usr rather than /etc. Make
3908 sure your system has dbus >= 1.9.18 running before upgrading to this
3909 version, or override the install path with --with-dbuspolicydir= .
3911 * All python scripts shipped with systemd (specifically: the various
3912 tests written in Python) now require Python 3.
3914 * systemd unit tests can now run standalone (without the source or
3915 build directories), and can be installed into /usr/lib/systemd/tests/
3916 with 'make install-tests'.
3918 * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
3919 CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
3922 * Support for the %c, %r, %R specifiers in unit files has been
3923 removed. Specifiers are not supposed to be dependent on configuration
3924 in the unit file itself (so that they resolve the same regardless
3925 where used in the unit files), but these specifiers were influenced
3926 by the Slice= option.
3928 * The shell invoked by debug-shell.service now defaults to /bin/sh in
3929 all cases. If distributions want to use a different shell for this
3930 purpose (for example Fedora's /sbin/sushell) they need to specify
3931 this explicitly at configure time using --with-debug-shell=.
3933 * The confirmation spawn prompt has been reworked to offer the
3936 (c)ontinue, proceed without asking anymore
3937 (D)ump, show the state of the unit
3938 (f)ail, don't execute the command and pretend it failed
3940 (i)nfo, show a short summary of the unit
3941 (j)obs, show jobs that are in progress
3942 (s)kip, don't execute the command and pretend it succeeded
3943 (y)es, execute the command
3945 The 'n' choice for the confirmation spawn prompt has been removed,
3946 because its meaning was confusing.
3948 The prompt may now also be redirected to an alternative console by
3949 specifying the console as parameter to systemd.confirm_spawn=.
3951 * Services of Type=notify require a READY=1 notification to be sent
3952 during startup. If no such message is sent, the service now fails,
3953 even if the main process exited with a successful exit code.
3955 * Services that fail to start up correctly now always have their
3956 ExecStopPost= commands executed. Previously, they'd enter "failed"
3957 state directly, without executing these commands.
3959 * The option MulticastDNS= of network configuration files has acquired
3960 an actual implementation. With MulticastDNS=yes a host can resolve
3961 names of remote hosts and reply to mDNS A and AAAA requests.
3963 * When units are about to be started an additional check is now done to
3964 ensure that all dependencies of type BindsTo= (when used in
3965 combination with After=) have been started.
3967 * systemd-analyze gained a new verb "syscall-filter" which shows which
3968 system call groups are defined for the SystemCallFilter= unit file
3969 setting, and which system calls they contain.
3971 * A new system call filter group "@filesystem" has been added,
3972 consisting of various file system related system calls. Group
3973 "@reboot" has been added, covering reboot, kexec and shutdown related
3974 calls. Finally, group "@swap" has been added covering swap
3975 configuration related calls.
3977 * A new unit file option RestrictNamespaces= has been added that may be
3978 used to restrict access to the various process namespace types the
3979 Linux kernel provides. Specifically, it may be used to take away the
3980 right for a service unit to create additional file system, network,
3981 user, and other namespaces. This sandboxing option is particularly
3982 relevant due to the high amount of recently discovered namespacing
3983 related vulnerabilities in the kernel.
3985 * systemd-udev's .link files gained support for a new AutoNegotiation=
3986 setting for configuring Ethernet auto-negotiation.
3988 * systemd-networkd's .network files gained support for a new
3989 ListenPort= setting in the [DHCP] section to explicitly configure the
3990 UDP client port the DHCP client shall listen on.
3992 * .network files gained a new Unmanaged= boolean setting for explicitly
3993 excluding one or more interfaces from management by systemd-networkd.
3995 * The systemd-networkd ProxyARP= option has been renamed to
3996 IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
3997 renamed to ReduceARPProxy=. The old names continue to be available
4000 * systemd-networkd gained support for configuring IPv6 Proxy NDP
4001 addresses via the new IPv6ProxyNDPAddress= .network file setting.
4003 * systemd-networkd's bonding device support gained support for two new
4004 configuration options ActiveSlave= and PrimarySlave=.
4006 * The various options in the [Match] section of .network files gained
4007 support for negative matching.
4009 * New systemd-specific mount options are now understood in /etc/fstab:
4011 x-systemd.mount-timeout= may be used to configure the maximum
4012 permitted runtime of the mount command.
4014 x-systemd.device-bound may be set to bind a mount point to its
4015 backing device unit, in order to automatically remove a mount point
4016 if its backing device is unplugged. This option may also be
4017 configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
4018 on the block device, which is now automatically set for all CDROM
4019 drives, so that mounted CDs are automatically unmounted when they are
4020 removed from the drive.
4022 x-systemd.after= and x-systemd.before= may be used to explicitly
4023 order a mount after or before another unit or mount point.
4025 * Enqueued start jobs for device units are now automatically garbage
4026 collected if there are no jobs waiting for them anymore.
4028 * systemctl list-jobs gained two new switches: with --after, for every
4029 queued job the jobs it's waiting for are shown; with --before the
4030 jobs which it's blocking are shown.
4032 * systemd-nspawn gained support for ephemeral boots from disk images
4033 (or in other words: --ephemeral and --image= may now be
4034 combined). Moreover, ephemeral boots are now supported for normal
4035 directories, even if the backing file system is not btrfs. Of course,
4036 if the file system does not support file system snapshots or
4037 reflinks, the initial copy operation will be relatively expensive, but
4038 this should still be suitable for many use cases.
4040 * Calendar time specifications in .timer units now support
4041 specifications relative to the end of a month by using "~" instead of
4042 "-" as separator between month and day. For example, "*-02~03" means
4043 "the third last day in February". In addition a new syntax for
4044 repeated events has been added using the "/" character. For example,
4045 "9..17/2:00" means "every two hours from 9am to 5pm".
4047 * systemd-socket-proxyd gained a new parameter --connections-max= for
4048 configuring the maximum number of concurrent connections.
4050 * sd-id128 gained a new API for generating unique IDs for the host in a
4051 way that does not leak the machine ID. Specifically,
4052 sd_id128_get_machine_app_specific() derives an ID based on the
4053 machine ID a in well-defined, non-reversible, stable way. This is
4054 useful whenever an identifier for the host is needed but where the
4055 identifier shall not be useful to identify the system beyond the
4056 scope of the application itself. (Internally this uses HMAC-SHA256 as
4057 keyed hash function using the machine ID as input.)
4059 * NotifyAccess= gained a new supported value "exec". When set
4060 notifications are accepted from all processes systemd itself invoked,
4061 including all control processes.
4063 * .nspawn files gained support for defining overlay mounts using the
4064 Overlay= and OverlayReadOnly= options. Previously this functionality
4065 was only available on the systemd-nspawn command line.
4067 * systemd-nspawn's --bind= and --overlay= options gained support for
4068 bind/overlay mounts whose source lies within the container tree by
4069 prefixing the source path with "+".
4071 * systemd-nspawn's --bind= and --overlay= options gained support for
4072 automatically allocating a temporary source directory in /var/tmp
4073 that is removed when the container dies. Specifically, if the source
4074 directory is specified as empty string this mechanism is selected. An
4075 example usage is --overlay=+/var::/var, which creates an overlay
4076 mount based on the original /var contained in the image, overlaid
4077 with a temporary directory in the host's /var/tmp. This way changes
4078 to /var are automatically flushed when the container shuts down.
4080 * systemd-nspawn --image= option does now permit raw file system block
4081 devices (in addition to images containing partition tables, as
4084 * The disk image dissection logic in systemd-nspawn gained support for
4085 automatically setting up LUKS encrypted as well as Verity protected
4086 partitions. When a container is booted from an encrypted image the
4087 passphrase is queried at start-up time. When a container with Verity
4088 data is started, the root hash is search in a ".roothash" file
4089 accompanying the disk image (alternatively, pass the root hash via
4090 the new --root-hash= command line option).
4092 * A new tool /usr/lib/systemd/systemd-dissect has been added that may
4093 be used to dissect disk images the same way as systemd-nspawn does
4094 it, following the Bootable Partition Specification. It may even be
4095 used to mount disk images with complex partition setups (including
4096 LUKS and Verity partitions) to a local host directory, in order to
4097 inspect them. This tool is not considered public API (yet), and is
4098 thus not installed into /usr/bin. Please do not rely on its
4099 existence, since it might go away or be changed in later systemd
4102 * A new generator "systemd-verity-generator" has been added, similar in
4103 style to "systemd-cryptsetup-generator", permitting automatic setup of
4104 Verity root partitions when systemd boots up. In order to make use of
4105 this your partition setup should follow the Discoverable Partitions
4106 Specification, and the GPT partition ID of the root file system
4107 partition should be identical to the upper 128bit of the Verity root
4108 hash. The GPT partition ID of the Verity partition protecting it
4109 should be the lower 128bit of the Verity root hash. If the partition
4110 image follows this model it is sufficient to specify a single
4111 "roothash=" kernel command line argument to both configure which root
4112 image and verity partition to use as well as the root hash for
4113 it. Note that systemd-nspawn's Verity support follows the same
4114 semantics, meaning that disk images with proper Verity data in place
4115 may be booted in containers with systemd-nspawn as well as on
4116 physical systems via the verity generator. Also note that the "mkosi"
4117 tool available at https://github.com/systemd/mkosi has been updated
4118 to generate Verity protected disk images following this scheme. In
4119 fact, it has been updated to generate disk images that optionally
4120 implement a complete UEFI SecureBoot trust chain, involving a signed
4121 kernel and initrd image that incorporates such a root hash as well as
4122 a Verity-enabled root partition.
4124 * The hardware database (hwdb) udev supports has been updated to carry
4125 accelerometer quirks.
4127 * All system services are now run with a fresh kernel keyring set up
4128 for them. The invocation ID is stored by default in it, thus
4129 providing a safe, non-overridable way to determine the invocation
4132 * Service unit files gained new BindPaths= and BindReadOnlyPaths=
4133 options for bind mounting arbitrary paths in a service-specific
4134 way. When these options are used, arbitrary host or service files and
4135 directories may be mounted to arbitrary locations in the service's
4138 * Documentation has been added that lists all of systemd's low-level
4139 environment variables:
4141 https://github.com/systemd/systemd/blob/master/docs/ENVIRONMENT.md
4143 * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
4144 whether a specific socket file descriptor matches a specified socket
4147 * systemd-firstboot has been updated to check for the
4148 systemd.firstboot= kernel command line option. It accepts a boolean
4149 and when set to false the first boot questions are skipped.
4151 * systemd-fstab-generator has been updated to check for the
4152 systemd.volatile= kernel command line option, which either takes an
4153 optional boolean parameter or the special value "state". If used the
4154 system may be booted in a "volatile" boot mode. Specifically,
4155 "systemd.volatile" is used, the root directory will be mounted as
4156 tmpfs, and only /usr is mounted from the actual root file system. If
4157 "systemd.volatile=state" is used, the root directory will be mounted
4158 as usual, but /var is mounted as tmpfs. This concept provides similar
4159 functionality as systemd-nspawn's --volatile= option, but provides it
4160 on physical boots. Use this option for implementing stateless
4161 systems, or testing systems with all state and/or configuration reset
4162 to the defaults. (Note though that many distributions are not
4163 prepared to boot up without a populated /etc or /var, though.)
4165 * systemd-gpt-auto-generator gained support for LUKS encrypted root
4166 partitions. Previously it only supported LUKS encrypted partitions
4167 for all other uses, except for the root partition itself.
4169 * Socket units gained support for listening on AF_VSOCK sockets for
4170 communication in virtualized QEMU environments.
4172 * The "configure" script gained a new option --with-fallback-hostname=
4173 for specifying the fallback hostname to use if none is configured in
4174 /etc/hostname. For example, by specifying
4175 --with-fallback-hostname=fedora it is possible to default to a
4176 hostname of "fedora" on pristine installations.
4178 * systemd-cgls gained support for a new --unit= switch for listing only
4179 the control groups of a specific unit. Similar --user-unit= has been
4180 added for listing only the control groups of a specific user unit.
4182 * systemd-mount gained a new --umount switch for unmounting a mount or
4183 automount point (and all mount/automount points below it).
4185 * systemd will now refuse full configuration reloads (via systemctl
4186 daemon-reload and related calls) unless at least 16MiB of free space
4187 are available in /run. This is a safety precaution in order to ensure
4188 that generators can safely operate after the reload completed.
4190 * A new unit file option RootImage= has been added, which has a similar
4191 effect as RootDirectory= but mounts the service's root directory from
4192 a disk image instead of plain directory. This logic reuses the same
4193 image dissection and mount logic that systemd-nspawn already uses,
4194 and hence supports any disk images systemd-nspawn supports, including
4195 those following the Discoverable Partition Specification, as well as
4196 Verity enabled images. This option enables systemd to run system
4197 services directly off disk images acting as resource bundles,
4198 possibly even including full integrity data.
4200 * A new MountAPIVFS= unit file option has been added, taking a boolean
4201 argument. If enabled /proc, /sys and /dev (collectively called the
4202 "API VFS") will be mounted for the service. This is only relevant if
4203 RootDirectory= or RootImage= is used for the service, as these mounts
4204 are of course in place in the host mount namespace anyway.
4206 * systemd-nspawn gained support for a new --pivot-root= switch. If
4207 specified the root directory within the container image is pivoted to
4208 the specified mount point, while the original root disk is moved to a
4209 different place. This option enables booting of ostree images
4210 directly with systemd-nspawn.
4212 * The systemd build scripts will no longer complain if the NTP server
4213 addresses are not changed from the defaults. Google now supports
4214 these NTP servers officially. We still recommend downstreams to
4215 properly register an NTP pool with the NTP pool project though.
4217 * coredumpctl gained a new "--reverse" option for printing the list
4218 of coredumps in reverse order.
4220 * coredumpctl will now show additional information about truncated and
4221 inaccessible coredumps, as well as coredumps that are still being
4222 processed. It also gained a new --quiet switch for suppressing
4223 additional informational message in its output.
4225 * coredumpctl gained support for only showing coredumps newer and/or
4226 older than specific timestamps, using the new --since= and --until=
4227 options, reminiscent of journalctl's options by the same name.
4229 * The systemd-coredump logic has been improved so that it may be reused
4230 to collect backtraces in non-compiled languages, for example in
4231 scripting languages such as Python.
4233 * machinectl will now show the UID shift of local containers, if user
4234 namespacing is enabled for them.
4236 * systemd will now optionally run "environment generator" binaries at
4237 configuration load time. They may be used to add environment
4238 variables to the environment block passed to services invoked. One
4239 user environment generator is shipped by default that sets up
4240 environment variables based on files dropped into /etc/environment.d
4241 and ~/.config/environment.d/.
4243 * systemd-resolved now includes the new, recently published 2017 DNSSEC
4246 * hostnamed has been updated to report a new chassis type of
4247 "convertible" to cover "foldable" laptops that can both act as a
4248 tablet and as a laptop, such as various Lenovo Yoga devices.
4250 Contributions from: Adrián López, Alexander Galanin, Alexander
4251 Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
4252 Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
4253 Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
4254 Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
4255 David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
4256 Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
4257 Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
4258 Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
4259 Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
4260 Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
4261 Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
4262 Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
4263 Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
4264 Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
4265 Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
4266 Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
4267 Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
4268 Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
4269 Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
4270 Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
4271 Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
4272 Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
4273 Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
4274 Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
4275 Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
4276 Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
4277 YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
4280 — Berlin, 2017-03-01
4284 * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
4285 RestrictAddressFamilies= enabled. These sandboxing options should
4286 generally be compatible with the various external udev call-out
4287 binaries we are aware of, however there may be exceptions, in
4288 particular when exotic languages for these call-outs are used. In
4289 this case, consider turning off these settings locally.
4291 * The new RemoveIPC= option can be used to remove IPC objects owned by
4292 the user or group of a service when that service exits.
4294 * The new ProtectKernelModules= option can be used to disable explicit
4295 load and unload operations of kernel modules by a service. In
4296 addition access to /usr/lib/modules is removed if this option is set.
4298 * ProtectSystem= option gained a new value "strict", which causes the
4299 whole file system tree with the exception of /dev, /proc, and /sys,
4300 to be remounted read-only for a service.
4302 * The new ProtectKernelTunables= option can be used to disable
4303 modification of configuration files in /sys and /proc by a service.
4304 Various directories and files are remounted read-only, so access is
4305 restricted even if the file permissions would allow it.
4307 * The new ProtectControlGroups= option can be used to disable write
4308 access by a service to /sys/fs/cgroup.
4310 * Various systemd services have been hardened with
4311 ProtectKernelTunables=yes, ProtectControlGroups=yes,
4312 RestrictAddressFamilies=.
4314 * Support for dynamically creating users for the lifetime of a service
4315 has been added. If DynamicUser=yes is specified, user and group IDs
4316 will be allocated from the range 61184..65519 for the lifetime of the
4317 service. They can be resolved using the new nss-systemd.so NSS
4318 module. The module must be enabled in /etc/nsswitch.conf. Services
4319 started in this way have PrivateTmp= and RemoveIPC= enabled, so that
4320 any resources allocated by the service will be cleaned up when the
4321 service exits. They also have ProtectHome=read-only and
4322 ProtectSystem=strict enabled, so they are not able to make any
4323 permanent modifications to the system.
4325 * The nss-systemd module also always resolves root and nobody, making
4326 it possible to have no /etc/passwd or /etc/group files in minimal
4327 container or chroot environments.
4329 * Services may be started with their own user namespace using the new
4330 boolean PrivateUsers= option. Only root, nobody, and the uid/gid
4331 under which the service is running are mapped. All other users are
4334 * Support for the cgroup namespace has been added to systemd-nspawn. If
4335 supported by kernel, the container system started by systemd-nspawn
4336 will have its own view of the cgroup hierarchy. This new behaviour
4337 can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
4339 * The new MemorySwapMax= option can be used to limit the maximum swap
4340 usage under the unified cgroup hierarchy.
4342 * Support for the CPU controller in the unified cgroup hierarchy has
4343 been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
4344 options. This controller requires out-of-tree patches for the kernel
4345 and the support is provisional.
4347 * Mount and automount units may now be created transiently
4348 (i.e. dynamically at runtime via the bus API, instead of requiring
4349 unit files in the file system).
4351 * systemd-mount is a new tool which may mount file systems – much like
4352 mount(8), optionally pulling in additional dependencies through
4353 transient .mount and .automount units. For example, this tool
4354 automatically runs fsck on a backing block device before mounting,
4355 and allows the automount logic to be used dynamically from the
4356 command line for establishing mount points. This tool is particularly
4357 useful when dealing with removable media, as it will ensure fsck is
4358 run – if necessary – before the first access and that the file system
4359 is quickly unmounted after each access by utilizing the automount
4360 logic. This maximizes the chance that the file system on the
4361 removable media stays in a clean state, and if it isn't in a clean
4362 state is fixed automatically.
4364 * LazyUnmount=yes option for mount units has been added to expose the
4365 umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
4368 * /efi will be used as the mount point of the EFI boot partition, if
4369 the directory is present, and the mount point was not configured
4370 through other means (e.g. fstab). If /efi directory does not exist,
4371 /boot will be used as before. This makes it easier to automatically
4372 mount the EFI partition on systems where /boot is used for something
4375 * When operating on GPT disk images for containers, systemd-nspawn will
4376 now mount the ESP to /boot or /efi according to the same rules as PID
4377 1 running on a host. This allows tools like "bootctl" to operate
4378 correctly within such containers, in order to make container images
4379 bootable on physical systems.
4381 * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
4383 * Two new user session targets have been added to support running
4384 graphical sessions under the systemd --user instance:
4385 graphical-session.target and graphical-session-pre.target. See
4386 systemd.special(7) for a description of how those targets should be
4389 * The vconsole initialization code has been significantly reworked to
4390 use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
4391 support unicode keymaps. Font and keymap configuration will now be
4392 copied to all allocated virtual consoles.
4394 * FreeBSD's bhyve virtualization is now detected.
4396 * Information recorded in the journal for core dumps now includes the
4397 contents of /proc/mountinfo and the command line of the process at
4398 the top of the process hierarchy (which is usually the init process
4401 * systemd-journal-gatewayd learned the --directory= option to serve
4402 files from the specified location.
4404 * journalctl --root=… can be used to peruse the journal in the
4405 /var/log/ directories inside of a container tree. This is similar to
4406 the existing --machine= option, but does not require the container to
4409 * The hardware database has been extended to support
4410 ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
4413 MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
4414 specify the click rate for mice which include a horizontal wheel with
4415 a click rate that is different than the one for the vertical wheel.
4417 * systemd-run gained a new --wait option that makes service execution
4418 synchronous. (Specifically, the command will not return until the
4419 specified service binary exited.)
4421 * systemctl gained a new --wait option that causes the start command to
4422 wait until the units being started have terminated again.
4424 * A new journal output mode "short-full" has been added which displays
4425 timestamps with abbreviated English day names and adds a timezone
4426 suffix. Those timestamps include more information than the default
4427 "short" output mode, and can be passed directly to journalctl's
4428 --since= and --until= options.
4430 * /etc/resolv.conf will be bind-mounted into containers started by
4431 systemd-nspawn, if possible, so any changes to resolv.conf contents
4432 are automatically propagated to the container.
4434 * The number of instances for socket-activated services originating
4435 from a single IP address can be limited with
4436 MaxConnectionsPerSource=, extending the existing setting of
4439 * systemd-networkd gained support for vcan ("Virtual CAN") interface
4442 * .netdev and .network configuration can now be extended through
4445 * UDP Segmentation Offload, TCP Segmentation Offload, Generic
4446 Segmentation Offload, Generic Receive Offload, Large Receive Offload
4447 can be enabled and disabled using the new UDPSegmentationOffload=,
4448 TCPSegmentationOffload=, GenericSegmentationOffload=,
4449 GenericReceiveOffload=, LargeReceiveOffload= options in the
4450 [Link] section of .link files.
4452 * The Spanning Tree Protocol, Priority, Aging Time, and the Default
4453 Port VLAN ID can be configured for bridge devices using the new STP=,
4454 Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
4455 section of .netdev files.
4457 * The route table to which routes received over DHCP or RA should be
4458 added can be configured with the new RouteTable= option in the [DHCP]
4459 and [IPv6AcceptRA] sections of .network files.
4461 * The Address Resolution Protocol can be disabled on links managed by
4462 systemd-networkd using the ARP=no setting in the [Link] section of
4465 * New environment variables $SERVICE_RESULT, $EXIT_CODE and
4466 $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
4467 encode information about the result and exit codes of the current
4468 service runtime cycle.
4470 * systemd-sysctl will now configure kernel parameters in the order
4471 they occur in the configuration files. This matches what sysctl
4472 has been traditionally doing.
4474 * kernel-install "plugins" that are executed to perform various
4475 tasks after a new kernel is added and before an old one is removed
4476 can now return a special value to terminate the procedure and
4477 prevent any later plugins from running.
4479 * Journald's SplitMode=login setting has been deprecated. It has been
4480 removed from documentation, and its use is discouraged. In a future
4481 release it will be completely removed, and made equivalent to current
4482 default of SplitMode=uid.
4484 * Storage=both option setting in /etc/systemd/coredump.conf has been
4485 removed. With fast LZ4 compression storing the core dump twice is not
4488 * The --share-system systemd-nspawn option has been replaced with an
4489 (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
4490 this functionality is discouraged. In addition the variables
4491 $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID,
4492 $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
4493 individual namespaces.
4495 * "machinectl list" now shows the IP address of running containers in
4496 the output, as well as OS release information.
4498 * "loginctl list" now shows the TTY of each session in the output.
4500 * sd-bus gained new API calls sd_bus_track_set_recursive(),
4501 sd_bus_track_get_recursive(), sd_bus_track_count_name(),
4502 sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
4503 tracking objects in a "recursive" mode, where a single client can be
4504 counted multiple times, if it takes multiple references.
4506 * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
4507 sd_bus_get_exit_on_disconnect(). They may be used to make a
4508 process using sd-bus automatically exit if the bus connection is
4511 * Bus clients of the service manager may now "pin" loaded units into
4512 memory, by taking an explicit reference on them. This is useful to
4513 ensure the client can retrieve runtime data about the service even
4514 after the service completed execution. Taking such a reference is
4515 available only for privileged clients and should be helpful to watch
4516 running services in a race-free manner, and in particular collect
4517 information about exit statuses and results.
4519 * The nss-resolve module has been changed to strictly return UNAVAIL
4520 when communication via D-Bus with resolved failed, and NOTFOUND when
4521 a lookup completed but was negative. This means it is now possible to
4522 neatly configure fallbacks using nsswitch.conf result checking
4523 expressions. Taking benefit of this, the new recommended
4524 configuration line for the "hosts" entry in /etc/nsswitch.conf is:
4526 hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
4528 * A new setting CtrlAltDelBurstAction= has been added to
4529 /etc/systemd/system.conf which may be used to configure the precise
4530 behaviour if the user on the console presses Ctrl-Alt-Del more often
4531 than 7 times in 2s. Previously this would unconditionally result in
4532 an expedited, immediate reboot. With this new setting the precise
4533 operation may be configured in more detail, and also turned off
4536 * In .netdev files two new settings RemoteChecksumTx= and
4537 RemoteChecksumRx= are now understood that permit configuring the
4538 remote checksumming logic for VXLAN networks.
4540 * The service manager learnt a new "invocation ID" concept for invoked
4541 services. Each runtime cycle of a service will get a new invocation
4542 ID (a 128bit random UUID) assigned that identifies the current
4543 run of the service uniquely and globally. A new invocation ID
4544 is generated each time a service starts up. The journal will store
4545 the invocation ID of a service along with any logged messages, thus
4546 making the invocation ID useful for matching the online runtime of a
4547 service with the offline log data it generated in a safe way without
4548 relying on synchronized timestamps. In many ways this new service
4549 invocation ID concept is similar to the kernel's boot ID concept that
4550 uniquely and globally identifies the runtime of each boot. The
4551 invocation ID of a service is passed to the service itself via an
4552 environment variable ($INVOCATION_ID). A new bus call
4553 GetUnitByInvocationID() has been added that is similar to GetUnit()
4554 but instead of retrieving the bus path for a unit by its name
4555 retrieves it by its invocation ID. The returned path is valid only as
4556 long as the passed invocation ID is current.
4558 * systemd-resolved gained a new "DNSStubListener" setting in
4559 resolved.conf. It either takes a boolean value or the special values
4560 "udp" and "tcp", and configures whether to enable the stub DNS
4561 listener on 127.0.0.53:53.
4563 * IP addresses configured via networkd may now carry additional
4564 configuration settings supported by the kernel. New options include:
4565 HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
4566 PrefixRoute=, AutoJoin=.
4568 * The PAM configuration fragment file for "user@.service" shipped with
4569 systemd (i.e. the --user instance of systemd) has been stripped to
4570 the minimum necessary to make the system boot. Previously, it
4571 contained Fedora-specific stanzas that did not apply to other
4572 distributions. It is expected that downstream distributions add
4573 additional configuration lines, matching their needs to this file,
4574 using it only as rough template of what systemd itself needs. Note
4575 that this reduced fragment does not even include an invocation of
4576 pam_limits which most distributions probably want to add, even though
4577 systemd itself does not need it. (There's also the new build time
4578 option --with-pamconfdir=no to disable installation of the PAM
4581 * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
4582 capability is now also dropped from its set (in addition to
4583 CAP_SYS_MKNOD as before).
4585 * In service unit files it is now possible to connect a specific named
4586 file descriptor with stdin/stdout/stdout of an executed service. The
4587 name may be specified in matching .socket units using the
4588 FileDescriptorName= setting.
4590 * A number of journal settings may now be configured on the kernel
4591 command line. Specifically, the following options are now understood:
4592 systemd.journald.max_level_console=,
4593 systemd.journald.max_level_store=,
4594 systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
4595 systemd.journald.max_level_wall=.
4597 * "systemctl is-enabled --full" will now show by which symlinks a unit
4598 file is enabled in the unit dependency tree.
4600 * Support for VeraCrypt encrypted partitions has been added to the
4601 "cryptsetup" logic and /etc/crypttab.
4603 * systemd-detect-virt gained support for a new --private-users switch
4604 that checks whether the invoking processes are running inside a user
4605 namespace. Similar, a new special value "private-users" for the
4606 existing ConditionVirtualization= setting has been added, permitting
4607 skipping of specific units in user namespace environments.
4609 Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
4610 Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
4611 Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
4612 Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
4613 Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
4614 Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
4615 Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
4616 Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
4617 Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
4618 Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
4619 Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
4620 Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
4621 Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
4622 Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
4623 Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
4624 Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
4625 Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
4626 Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
4627 Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
4628 Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
4629 Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
4630 Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
4631 Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
4632 Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
4633 E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
4634 Jędrzejewski-Szmek, Zeal Jagannatha
4636 — Santa Fe, 2016-11-03
4640 * In service units the various ExecXYZ= settings have been extended
4641 with an additional special character as first argument of the
4642 assigned value: if the character '+' is used the specified command
4643 line it will be run with full privileges, regardless of User=,
4644 Group=, CapabilityBoundingSet= and similar options. The effect is
4645 similar to the existing PermissionsStartOnly= option, but allows
4646 configuration of this concept for each executed command line
4649 * Services may now alter the service watchdog timeout at runtime by
4650 sending a WATCHDOG_USEC= message via sd_notify().
4652 * MemoryLimit= and related unit settings now optionally take percentage
4653 specifications. The percentage is taken relative to the amount of
4654 physical memory in the system (or in case of containers, the assigned
4655 amount of memory). This allows scaling service resources neatly with
4656 the amount of RAM available on the system. Similarly, systemd-logind's
4657 RuntimeDirectorySize= option now also optionally takes percentage
4660 * In similar fashion TasksMax= takes percentage values now, too. The
4661 value is taken relative to the configured maximum number of processes
4662 on the system. The per-service task maximum has been changed to 15%
4663 using this functionality. (Effectively this is an increase of 512 →
4664 4915 for service units, given the kernel's default pid_max setting.)
4666 * Calendar time specifications in .timer units now understand a ".."
4667 syntax for time ranges. Example: "4..7:10" may now be used for
4668 defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
4671 * The InaccessableDirectories=, ReadOnlyDirectories= and
4672 ReadWriteDirectories= unit file settings have been renamed to
4673 InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
4674 applied to all kinds of file nodes, and not just directories, with
4675 the exception of symlinks. Specifically these settings may now be
4676 used on block and character device nodes, UNIX sockets and FIFOS as
4677 well as regular files. The old names of these settings remain
4678 available for compatibility.
4680 * systemd will now log about all service processes it kills forcibly
4681 (using SIGKILL) because they remained after the clean shutdown phase
4682 of the service completed. This should help identifying services that
4683 shut down uncleanly. Moreover if KillUserProcesses= is enabled in
4684 systemd-logind's configuration a similar log message is generated for
4685 processes killed at the end of each session due to this setting.
4687 * systemd will now set the $JOURNAL_STREAM environment variable for all
4688 services whose stdout/stderr are connected to the Journal (which
4689 effectively means by default: all services). The variable contains
4690 the device and inode number of the file descriptor used for
4691 stdout/stderr. This may be used by invoked programs to detect whether
4692 their stdout/stderr is connected to the Journal, in which case they
4693 can switch over to direct Journal communication, thus being able to
4694 pass extended, structured metadata along with their log messages. As
4695 one example, this is now used by glib's logging primitives.
4697 * When using systemd's default tmp.mount unit for /tmp, the mount point
4698 will now be established with the "nosuid" and "nodev" options. This
4699 avoids privilege escalation attacks that put traps and exploits into
4700 /tmp. However, this might cause problems if you e. g. put container
4701 images or overlays into /tmp; if you need this, override tmp.mount's
4702 "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
4705 * systemd now supports the "memory" cgroup controller also on
4708 * The systemd-cgtop tool now optionally takes a control group path as
4709 command line argument. If specified, the control group list shown is
4710 limited to subgroups of that group.
4712 * The SystemCallFilter= unit file setting gained support for
4713 pre-defined, named system call filter sets. For example
4714 SystemCallFilter=@clock is now an effective way to make all clock
4715 changing-related system calls unavailable to a service. A number of
4716 similar pre-defined groups are defined. Writing system call filters
4717 for system services is simplified substantially with this new
4718 concept. Accordingly, all of systemd's own, long-running services now
4719 enable system call filtering based on this, by default.
4721 * A new service setting MemoryDenyWriteExecute= has been added, taking
4722 a boolean value. If turned on, a service may no longer create memory
4723 mappings that are writable and executable at the same time. This
4724 enhances security for services where this is enabled as it becomes
4725 harder to dynamically write and then execute memory in exploited
4726 service processes. This option has been enabled for all of systemd's
4727 own long-running services.
4729 * A new RestrictRealtime= service setting has been added, taking a
4730 boolean argument. If set the service's processes may no longer
4731 acquire realtime scheduling. This improves security as realtime
4732 scheduling may otherwise be used to easily freeze the system.
4734 * systemd-nspawn gained a new switch --notify-ready= taking a boolean
4735 value. This may be used for requesting that the system manager inside
4736 of the container reports start-up completion to nspawn which then
4737 propagates this notification further to the service manager
4738 supervising nspawn itself. A related option NotifyReady= in .nspawn
4739 files has been added too. This functionality allows ordering of the
4740 start-up of multiple containers using the usual systemd ordering
4743 * machinectl gained a new command "stop" that is an alias for
4746 * systemd-resolved gained support for contacting DNS servers on
4747 link-local IPv6 addresses.
4749 * If systemd-resolved receives the SIGUSR2 signal it will now flush all
4750 its caches. A method call for requesting the same operation has been
4751 added to the bus API too, and is made available via "systemd-resolve
4754 * systemd-resolve gained a new --status switch. If passed a brief
4755 summary of the used DNS configuration with per-interface information
4758 * resolved.conf gained a new Cache= boolean option, defaulting to
4759 on. If turned off local DNS caching is disabled. This comes with a
4760 performance penalty in particular when DNSSEC is enabled. Note that
4761 resolved disables its internal caching implicitly anyway, when the
4762 configured DNS server is on a host-local IP address such as ::1 or
4763 127.0.0.1, thus automatically avoiding double local caching.
4765 * systemd-resolved now listens on the local IP address 127.0.0.53:53
4766 for DNS requests. This improves compatibility with local programs
4767 that do not use the libc NSS or systemd-resolved's bus APIs for name
4768 resolution. This minimal DNS service is only available to local
4769 programs and does not implement the full DNS protocol, but enough to
4770 cover local DNS clients. A new, static resolv.conf file, listing just
4771 this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
4772 now recommended to make /etc/resolv.conf a symlink to this file in
4773 order to route all DNS lookups to systemd-resolved, regardless if
4774 done via NSS, the bus API or raw DNS packets. Note that this local
4775 DNS service is not as fully featured as the libc NSS or
4776 systemd-resolved's bus APIs. For example, as unicast DNS cannot be
4777 used to deliver link-local address information (as this implies
4778 sending a local interface index along), LLMNR/mDNS support via this
4779 interface is severely restricted. It is thus strongly recommended for
4780 all applications to use the libc NSS API or native systemd-resolved
4783 * systemd-networkd's bridge support learned a new setting
4784 VLANFiltering= for controlling VLAN filtering. Moreover a new section
4785 in .network files has been added for configuring VLAN bridging in
4786 more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
4788 * systemd-networkd's IPv6 Router Advertisement code now makes use of
4789 the DNSSL and RDNSS options. This means IPv6 DNS configuration may
4790 now be acquired without relying on DHCPv6. Two new options
4791 UseDomains= and UseDNS= have been added to configure this behaviour.
4793 * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
4794 renamed IPv6AcceptRA=, without altering its behaviour. The old
4795 setting name remains available for compatibility reasons.
4797 * The systemd-networkd VTI/VTI6 tunneling support gained new options
4798 Key=, InputKey= and OutputKey=.
4800 * systemd-networkd gained support for VRF ("Virtual Routing Function")
4801 interface configuration.
4803 * "systemctl edit" may now be used to create new unit files by
4804 specifying the --force switch.
4806 * sd-event gained a new function sd_event_get_iteration() for
4807 requesting the current iteration counter of the event loop. It starts
4808 at zero and is increased by one with each event loop iteration.
4810 * A new rpm macro %systemd_ordering is provided by the macros.systemd
4811 file. It can be used in lieu of %systemd_requires in packages which
4812 don't use any systemd functionality and are intended to be installed
4813 in minimal containers without systemd present. This macro provides
4814 ordering dependencies to ensure that if the package is installed in
4815 the same rpm transaction as systemd, systemd will be installed before
4816 the scriptlets for the package are executed, allowing unit presets
4819 New macros %_systemdgeneratordir and %_systemdusergeneratordir have
4820 been added to simplify packaging of generators.
4822 * The os-release file gained VERSION_CODENAME field for the
4823 distribution nickname (e.g. VERSION_CODENAME=woody).
4825 * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1
4826 can be set to disable parsing of metadata and the creation
4827 of persistent symlinks for that device.
4829 * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
4830 to make them available to logged-in users has been reverted.
4832 * Much of the common code of the various systemd components is now
4833 built into an internal shared library libsystemd-shared-231.so
4834 (incorporating the systemd version number in the name, to be updated
4835 with future releases) that the components link to. This should
4836 decrease systemd footprint both in memory during runtime and on
4837 disk. Note that the shared library is not for public use, and is
4838 neither API nor ABI stable, but is likely to change with every new
4839 released update. Packagers need to make sure that binaries
4840 linking to libsystemd-shared.so are updated in step with the
4843 * Configuration for "mkosi" is now part of the systemd
4844 repository. mkosi is a tool to easily build legacy-free OS images,
4845 and is available on github: https://github.com/systemd/mkosi. If
4846 "mkosi" is invoked in the build tree a new raw OS image is generated
4847 incorporating the systemd sources currently being worked on and a
4848 clean, fresh distribution installation. The generated OS image may be
4849 booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
4850 UEFI PC. This functionality is particularly useful to easily test
4851 local changes made to systemd in a pristine, defined environment. See
4852 doc/HACKING for details.
4854 * configure learned the --with-support-url= option to specify the
4855 distribution's bugtracker.
4857 Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
4858 Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
4859 Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
4860 Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
4861 Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
4862 Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
4863 Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
4864 Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
4865 Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
4866 Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
4867 Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
4868 Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
4869 Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
4870 Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
4871 Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
4872 Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
4873 Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
4874 Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
4875 WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
4877 — Berlin, 2016-07-25
4881 * DNSSEC is now turned on by default in systemd-resolved (in
4882 "allow-downgrade" mode), but may be turned off during compile time by
4883 passing "--with-default-dnssec=no" to "configure" (and of course,
4884 during runtime with DNSSEC= in resolved.conf). We recommend
4885 downstreams to leave this on at least during development cycles and
4886 report any issues with the DNSSEC logic upstream. We are very
4887 interested in collecting feedback about the DNSSEC validator and its
4888 limitations in the wild. Note however, that DNSSEC support is
4889 probably nothing downstreams should turn on in stable distros just
4890 yet, as it might create incompatibilities with a few DNS servers and
4891 networks. We tried hard to make sure we downgrade to non-DNSSEC mode
4892 automatically whenever we detect such incompatible setups, but there
4893 might be systems we do not cover yet. Hence: please help us testing
4894 the DNSSEC code, leave this on where you can, report back, but then
4895 again don't consider turning this on in your stable, LTS or
4896 production release just yet. (Note that you have to enable
4897 nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
4898 and its DNSSEC mode for hostname resolution from local
4901 * systemd-resolve conveniently resolves DANE records with the --tlsa
4902 option and OPENPGPKEY records with the --openpgp option. It also
4903 supports dumping raw DNS record data via the new --raw= switch.
4905 * systemd-logind will now by default terminate user processes that are
4906 part of the user session scope unit (session-XX.scope) when the user
4907 logs out. This behavior is controlled by the KillUserProcesses=
4908 setting in logind.conf, and the previous default of "no" is now
4909 changed to "yes". This means that user sessions will be properly
4910 cleaned up after, but additional steps are necessary to allow
4911 intentionally long-running processes to survive logout.
4913 While the user is logged in at least once, user@.service is running,
4914 and any service that should survive the end of any individual login
4915 session can be started at a user service or scope using systemd-run.
4916 systemd-run(1) man page has been extended with an example which shows
4917 how to run screen in a scope unit underneath user@.service. The same
4918 command works for tmux.
4920 After the user logs out of all sessions, user@.service will be
4921 terminated too, by default, unless the user has "lingering" enabled.
4922 To effectively allow users to run long-term tasks even if they are
4923 logged out, lingering must be enabled for them. See loginctl(1) for
4924 details. The default polkit policy was modified to allow users to
4925 set lingering for themselves without authentication.
4927 Previous defaults can be restored at compile time by the
4928 --without-kill-user-processes option to "configure".
4930 * systemd-logind gained new configuration settings SessionsMax= and
4931 InhibitorsMax=, both with a default of 8192. It will not register new
4932 user sessions or inhibitors above this limit.
4934 * systemd-logind will now reload configuration on SIGHUP.
4936 * The unified cgroup hierarchy added in Linux 4.5 is now supported.
4937 Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
4938 enable. Also, support for the "io" cgroup controller in the unified
4939 hierarchy has been added, so that the "memory", "pids" and "io" are
4940 now the controllers that are supported on the unified hierarchy.
4942 WARNING: it is not possible to use previous systemd versions with
4943 systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
4944 is necessary to also update systemd in the initramfs if using the
4945 unified hierarchy. An updated SELinux policy is also required.
4947 * LLDP support has been extended, and both passive (receive-only) and
4948 active (sender) modes are supported. Passive mode ("routers-only") is
4949 enabled by default in systemd-networkd. Active LLDP mode is enabled
4950 by default for containers on the internal network. The "networkctl
4951 lldp" command may be used to list information gathered. "networkctl
4952 status" will also show basic LLDP information on connected peers now.
4954 * The IAID and DUID unique identifier sent in DHCP requests may now be
4955 configured for the system and each .network file managed by
4956 systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
4958 * systemd-networkd gained support for configuring proxy ARP support for
4959 each interface, via the ProxyArp= setting in .network files. It also
4960 gained support for configuring the multicast querier feature of
4961 bridge devices, via the new MulticastQuerier= setting in .netdev
4962 files. Similarly, snooping on the IGMP traffic can be controlled
4963 via the new setting MulticastSnooping=.
4965 A new setting PreferredLifetime= has been added for addresses
4966 configured in .network file to configure the lifetime intended for an
4969 The systemd-networkd DHCP server gained the option EmitRouter=, which
4970 defaults to yes, to configure whether the DHCP Option 3 (Router)
4973 * The testing tool /usr/lib/systemd/systemd-activate is renamed to
4974 systemd-socket-activate and installed into /usr/bin. It is now fully
4977 * systemd-journald now uses separate threads to flush changes to disk
4978 when closing journal files, thus reducing impact of slow disk I/O on
4979 logging performance.
4981 * The sd-journal API gained two new calls
4982 sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
4983 can be used to open journal files using file descriptors instead of
4984 file or directory paths. sd_journal_open_container() has been
4985 deprecated, sd_journal_open_directory_fd() should be used instead
4986 with the flag SD_JOURNAL_OS_ROOT.
4988 * journalctl learned a new output mode "-o short-unix" that outputs log
4989 lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
4990 UTC). It also gained support for a new --no-hostname setting to
4991 suppress the hostname column in the family of "short" output modes.
4993 * systemd-ask-password now optionally skips printing of the password to
4994 stdout with --no-output which can be useful in scripts.
4996 * Framebuffer devices (/dev/fb*) and 3D printers and scanners
4997 (devices tagged with ID_MAKER_TOOL) are now tagged with
4998 "uaccess" and are available to logged in users.
5000 * The DeviceAllow= unit setting now supports specifiers (with "%").
5002 * "systemctl show" gained a new --value switch, which allows print a
5003 only the contents of a specific unit property, without also printing
5004 the property's name. Similar support was added to "show*" verbs
5005 of loginctl and machinectl that output "key=value" lists.
5007 * A new unit type "generated" was added for files dynamically generated
5008 by generator tools. Similarly, a new unit type "transient" is used
5009 for unit files created using the runtime API. "systemctl enable" will
5010 refuse to operate on such files.
5012 * A new command "systemctl revert" has been added that may be used to
5013 revert to the vendor version of a unit file, in case local changes
5014 have been made by adding drop-ins or overriding the unit file.
5016 * "machinectl clean" gained a new verb to automatically remove all or
5017 just hidden container images.
5019 * systemd-tmpfiles gained support for a new line type "e" for emptying
5020 directories, if they exist, without creating them if they don't.
5022 * systemd-nspawn gained support for automatically patching the UID/GIDs
5023 of the owners and the ACLs of all files and directories in a
5024 container tree to match the UID/GID user namespacing range selected
5025 for the container invocation. This mode is enabled via the new
5026 --private-users-chown switch. It also gained support for
5027 automatically choosing a free, previously unused UID/GID range when
5028 starting a container, via the new --private-users=pick setting (which
5029 implies --private-users-chown). Together, these options for the first
5030 time make user namespacing for nspawn containers fully automatic and
5031 thus deployable. The systemd-nspawn@.service template unit file has
5032 been changed to use this functionality by default.
5034 * systemd-nspawn gained a new --network-zone= switch, that allows
5035 creating ad-hoc virtual Ethernet links between multiple containers,
5036 that only exist as long as at least one container referencing them is
5037 running. This allows easy connecting of multiple containers with a
5038 common link that implements an Ethernet broadcast domain. Each of
5039 these network "zones" may be named relatively freely by the user, and
5040 may be referenced by any number of containers, but each container may
5041 only reference one of these "zones". On the lower level, this is
5042 implemented by an automatically managed bridge network interface for
5043 each zone, that is created when the first container referencing its
5044 zone is created and removed when the last one referencing its zone
5047 * The default start timeout may now be configured on the kernel command
5048 line via systemd.default_timeout_start_sec=. It was already
5049 configurable via the DefaultTimeoutStartSec= option in
5050 /etc/systemd/system.conf.
5052 * Socket units gained a new TriggerLimitIntervalSec= and
5053 TriggerLimitBurst= setting to configure a limit on the activation
5054 rate of the socket unit.
5056 * The LimitNICE= setting now optionally takes normal UNIX nice values
5057 in addition to the raw integer limit value. If the specified
5058 parameter is prefixed with "+" or "-" and is in the range -20..19 the
5059 value is understood as UNIX nice value. If not prefixed like this it
5060 is understood as raw RLIMIT_NICE limit.
5062 * Note that the effect of the PrivateDevices= unit file setting changed
5063 slightly with this release: the per-device /dev file system will be
5064 mounted read-only from this version on, and will have "noexec"
5065 set. This (minor) change of behavior might cause some (exceptional)
5066 legacy software to break, when PrivateDevices=yes is set for its
5067 service. Please leave PrivateDevices= off if you run into problems
5070 * systemd-bootchart has been split out to a separate repository:
5071 https://github.com/systemd/systemd-bootchart
5073 * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
5074 merged into the kernel in its current form.
5076 * The compatibility libraries libsystemd-daemon.so,
5077 libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so
5078 which have been deprecated since systemd-209 have been removed along
5079 with the corresponding pkg-config files. All symbols provided by
5080 those libraries are provided by libsystemd.so.
5082 * The Capabilities= unit file setting has been removed (it is ignored
5083 for backwards compatibility). AmbientCapabilities= and
5084 CapabilityBoundingSet= should be used instead.
5086 * A new special target has been added, initrd-root-device.target,
5087 which creates a synchronization point for dependencies of the root
5088 device in early userspace. Initramfs builders must ensure that this
5089 target is now included in early userspace.
5091 Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
5092 Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
5093 Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
5094 Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
5095 Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
5096 R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
5097 Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
5098 Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
5099 Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
5100 Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
5101 John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
5102 Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
5103 Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
5104 Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
5105 Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
5106 mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
5107 Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
5108 Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
5109 Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
5110 Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
5111 Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
5112 Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
5113 Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
5114 Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
5115 Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
5118 — Fairfax, 2016-05-21
5122 * The systemd-resolved DNS resolver service has gained a substantial
5123 set of new features, most prominently it may now act as a DNSSEC
5124 validating stub resolver. DNSSEC mode is currently turned off by
5125 default, but is expected to be turned on by default in one of the
5126 next releases. For now, we invite everybody to test the DNSSEC logic
5127 by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
5128 service also gained a full set of D-Bus interfaces, including calls
5129 to configure DNS and DNSSEC settings per link (for use by external
5130 network management software). systemd-resolved and systemd-networkd
5131 now distinguish between "search" and "routing" domains. The former
5132 are used to qualify single-label names, the latter are used purely
5133 for routing lookups within certain domains to specific links.
5134 resolved now also synthesizes RRs for all entries from /etc/hosts.
5136 * The systemd-resolve tool (which is a client utility for
5137 systemd-resolved) has been improved considerably and is now fully
5138 supported and documented. Hence it has moved from /usr/lib/systemd to
5141 * /dev/disk/by-path/ symlink support has been (re-)added for virtio
5144 * The coredump collection logic has been reworked: when a coredump is
5145 collected it is now written to disk, compressed and processed
5146 (including stacktrace extraction) from a new instantiated service
5147 systemd-coredump@.service, instead of directly from the
5148 /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
5149 processing large coredumps can take up a substantial amount of
5150 resources and time, and this previously happened entirely outside of
5151 systemd's service supervision. With the new logic the core_pattern
5152 hook only does minimal metadata collection before passing off control
5153 to the new instantiated service, which is configured with a time
5154 limit, a nice level and other settings to minimize negative impact on
5155 the rest of the system. Also note that the new logic will honour the
5156 RLIMIT_CORE setting of the crashed process, which now allows users
5157 and processes to turn off coredumping for their processes by setting
5160 * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
5161 and all forked processes by default. Previously, PID 1 would leave
5162 the setting at "0" for all processes, as set by the kernel. Note that
5163 the resource limit traditionally has no effect on the generated
5164 coredumps on the system if the /proc/sys/kernel/core_pattern hook
5165 logic is used. Since the limit is now honoured (see above) its
5166 default has been changed so that the coredumping logic is enabled by
5167 default for all processes, while allowing specific opt-out.
5169 * When the stacktrace is extracted from processes of system users, this
5170 is now done as "systemd-coredump" user, in order to sandbox this
5171 potentially security sensitive parsing operation. (Note that when
5172 processing coredumps of normal users this is done under the user ID
5173 of process that crashed, as before.) Packagers should take notice
5174 that it is now necessary to create the "systemd-coredump" system user
5175 and group at package installation time.
5177 * The systemd-activate socket activation testing tool gained support
5178 for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
5179 and --seqpacket switches. It also has been extended to support both
5180 new-style and inetd-style file descriptor passing. Use the new
5181 --inetd switch to request inetd-style file descriptor passing.
5183 * Most systemd tools now honor a new $SYSTEMD_COLORS environment
5184 variable, which takes a boolean value. If set to false, ANSI color
5185 output is disabled in the tools even when run on a terminal that
5188 * The VXLAN support in networkd now supports two new settings
5189 DestinationPort= and PortRange=.
5191 * A new systemd.machine_id= kernel command line switch has been added,
5192 that may be used to set the machine ID in /etc/machine-id if it is
5193 not initialized yet. This command line option has no effect if the
5194 file is already initialized.
5196 * systemd-nspawn gained a new --as-pid2 switch that invokes any
5197 specified command line as PID 2 rather than PID 1 in the
5198 container. In this mode PID 1 is a minimal stub init process that
5199 implements the special POSIX and Linux semantics of PID 1 regarding
5200 signal and child process management. Note that this stub init process
5201 is implemented in nspawn itself and requires no support from the
5202 container image. This new logic is useful to support running
5203 arbitrary commands in the container, as normal processes are
5204 generally not prepared to run as PID 1.
5206 * systemd-nspawn gained a new --chdir= switch for setting the current
5207 working directory for the process started in the container.
5209 * "journalctl /dev/sda" will now output all kernel log messages for
5210 specified device from the current boot, in addition to all devices
5211 that are parents of it. This should make log output about devices
5212 pretty useful, as long as kernel drivers attach enough metadata to
5213 the log messages. (The usual SATA drivers do.)
5215 * The sd-journal API gained two new calls
5216 sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
5217 that report whether log data from /run or /var has been found.
5219 * journalctl gained a new switch "--fields" that prints all journal
5220 record field names currently in use in the journal. This is backed
5221 by two new sd-journal API calls sd_journal_enumerate_fields() and
5222 sd_journal_restart_fields().
5224 * Most configurable timeouts in systemd now expect an argument of
5225 "infinity" to turn them off, instead of "0" as before. The semantics
5226 from now on is that a timeout of "0" means "now", and "infinity"
5227 means "never". To maintain backwards compatibility, "0" continues to
5228 turn off previously existing timeout settings.
5230 * "systemctl reload-or-try-restart" has been renamed to "systemctl
5231 try-reload-or-restart" to clarify what it actually does: the "try"
5232 logic applies to both reloading and restarting, not just restarting.
5233 The old name continues to be accepted for compatibility.
5235 * On boot-up, when PID 1 detects that the system clock is behind the
5236 release date of the systemd version in use, the clock is now set
5237 to the latter. Previously, this was already done in timesyncd, in order
5238 to avoid running with clocks set to the various clock epochs such as
5239 1902, 1938 or 1970. With this change the logic is now done in PID 1
5240 in addition to timesyncd during early boot-up, so that it is enforced
5241 before the first process is spawned by systemd. Note that the logic
5242 in timesyncd remains, as it is more comprehensive and ensures
5243 clock monotonicity by maintaining a persistent timestamp file in
5244 /var. Since /var is generally not available in earliest boot or the
5245 initrd, this part of the logic remains in timesyncd, and is not done
5248 * Support for tweaking details in net_cls.class_id through the
5249 NetClass= configuration directive has been removed, as the kernel
5250 people have decided to deprecate that controller in cgroup v2.
5251 Userspace tools such as nftables are moving over to setting rules
5252 that are specific to the full cgroup path of a task, which obsoletes
5253 these controllers anyway. The NetClass= directive is kept around for
5254 legacy compatibility reasons. For a more in-depth description of the
5255 kernel change, please refer to the respective upstream commit:
5257 https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
5259 * A new service setting RuntimeMaxSec= has been added that may be used
5260 to specify a maximum runtime for a service. If the timeout is hit, the
5261 service is terminated and put into a failure state.
5263 * A new service setting AmbientCapabilities= has been added. It allows
5264 configuration of additional Linux process capabilities that are
5265 passed to the activated processes. This is only available on very
5268 * The process resource limit settings in service units may now be used
5269 to configure hard and soft limits individually.
5271 * The various libsystemd APIs such as sd-bus or sd-event now publicly
5272 expose support for gcc's __attribute__((cleanup())) C extension.
5273 Specifically, for many object destructor functions alternative
5274 versions have been added that have names suffixed with "p" and take a
5275 pointer to a pointer to the object to destroy, instead of just a
5276 pointer to the object itself. This is useful because these destructor
5277 functions may be used directly as parameters to the cleanup
5278 construct. Internally, systemd has been a heavy user of this GCC
5279 extension for a long time, and with this change similar support is
5280 now available to consumers of the library outside of systemd. Note
5281 that by using this extension in your sources compatibility with old
5282 and strictly ANSI compatible C compilers is lost. However, all gcc or
5283 LLVM versions of recent years support this extension.
5285 * Timer units gained support for a new setting RandomizedDelaySec= that
5286 allows configuring some additional randomized delay to the configured
5287 time. This is useful to spread out timer events to avoid load peaks in
5288 clusters or larger setups.
5290 * Calendar time specifications now support sub-second accuracy.
5292 * Socket units now support listening on SCTP and UDP-lite protocol
5295 * The sd-event API now comes with a full set of man pages.
5297 * Older versions of systemd contained experimental support for
5298 compressing journal files and coredumps with the LZ4 compressor that
5299 was not compatible with the lz4 binary (due to API limitations of the
5300 lz4 library). This support has been removed; only support for files
5301 compatible with the lz4 binary remains. This LZ4 logic is now
5302 officially supported and no longer considered experimental.
5304 * The dkr image import logic has been removed again from importd. dkr's
5305 micro-services focus doesn't fit into the machine image focus of
5306 importd, and quickly got out of date with the upstream dkr API.
5308 * Creation of the /run/lock/lockdev/ directory was dropped from
5309 tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
5310 been available for many years. If you still need this, you need to
5311 create your own tmpfiles.d config file with:
5313 d /run/lock/lockdev 0775 root lock -
5315 * The settings StartLimitBurst=, StartLimitInterval=, StartLimitAction=
5316 and RebootArgument= have been moved from the [Service] section of
5317 unit files to [Unit], and they are now supported on all unit types,
5318 not just service units. Of course, systemd will continue to
5319 understand these settings also at the old location, in order to
5320 maintain compatibility.
5322 Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
5323 Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
5324 Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
5325 Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
5326 Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
5327 David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
5328 Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
5329 Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
5330 Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
5331 Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
5332 Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
5333 lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
5334 Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
5335 Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
5336 Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
5337 Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
5338 Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
5339 Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
5340 Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
5342 — Berlin, 2016-02-11
5346 * A number of properties previously only settable in unit
5347 files are now also available as properties to set when
5348 creating transient units programmatically via the bus, as it
5349 is exposed with systemd-run's --property=
5350 setting. Specifically, these are: SyslogIdentifier=,
5351 SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
5352 EnvironmentFile=, ReadWriteDirectories=,
5353 ReadOnlyDirectories=, InaccessibleDirectories=,
5354 ProtectSystem=, ProtectHome=, RuntimeDirectory=.
5356 * When creating transient services via the bus API it is now
5357 possible to pass in a set of file descriptors to use as
5358 STDIN/STDOUT/STDERR for the invoked process.
5360 * Slice units may now be created transiently via the bus APIs,
5361 similar to the way service and scope units may already be
5362 created transiently.
5364 * Wherever systemd expects a calendar timestamp specification
5365 (like in journalctl's --since= and --until= switches) UTC
5366 timestamps are now supported. Timestamps suffixed with "UTC"
5367 are now considered to be in Universal Time Coordinated
5368 instead of the local timezone. Also, timestamps may now
5369 optionally be specified with sub-second accuracy. Both of
5370 these additions also apply to recurring calendar event
5371 specification, such as OnCalendar= in timer units.
5373 * journalctl gained a new "--sync" switch that asks the
5374 journal daemon to write all so far unwritten log messages to
5375 disk and sync the files, before returning.
5377 * systemd-tmpfiles learned two new line types "q" and "Q" that
5378 operate like "v", but also set up a basic btrfs quota
5379 hierarchy when used on a btrfs file system with quota
5382 * tmpfiles' "v", "q" and "Q" will now create a plain directory
5383 instead of a subvolume (even on a btrfs file system) if the
5384 root directory is a plain directory, and not a
5385 subvolume. This should simplify things with certain chroot()
5386 environments which are not aware of the concept of btrfs
5389 * systemd-detect-virt gained a new --chroot switch to detect
5390 whether execution takes place in a chroot() environment.
5392 * CPUAffinity= now takes CPU index ranges in addition to
5395 * The various memory-related resource limit settings (such as
5396 LimitAS=) now understand the usual K, M, G, ... suffixes to
5397 the base of 1024 (IEC). Similar, the time-related resource
5398 limit settings understand the usual min, h, day, ...
5401 * There's a new system.conf setting DefaultTasksMax= to
5402 control the default TasksMax= setting for services and
5403 scopes running on the system. (TasksMax= is the primary
5404 setting that exposes the "pids" cgroup controller on systemd
5405 and was introduced in the previous systemd release.) The
5406 setting now defaults to 512, which means services that are
5407 not explicitly configured otherwise will only be able to
5408 create 512 processes or threads at maximum, from this
5409 version on. Note that this means that thread- or
5410 process-heavy services might need to be reconfigured to set
5411 TasksMax= to a higher value. It is sufficient to set
5412 TasksMax= in these specific unit files to a higher value, or
5413 even "infinity". Similar, there's now a logind.conf setting
5414 UserTasksMax= that defaults to 4096 and limits the total
5415 number of processes or tasks each user may own
5416 concurrently. nspawn containers also have the TasksMax=
5417 value set by default now, to 8192. Note that all of this
5418 only has an effect if the "pids" cgroup controller is
5419 enabled in the kernel. The general benefit of these changes
5420 should be a more robust and safer system, that provides a
5421 certain amount of per-service fork() bomb protection.
5423 * systemd-nspawn gained the new --network-veth-extra= switch
5424 to define additional and arbitrarily-named virtual Ethernet
5425 links between the host and the container.
5427 * A new service execution setting PassEnvironment= has been
5428 added that allows importing select environment variables
5429 from PID1's environment block into the environment block of
5432 * Timer units gained support for a new RemainAfterElapse=
5433 setting which takes a boolean argument. It defaults to on,
5434 exposing behaviour unchanged to previous releases. If set to
5435 off, timer units are unloaded after they elapsed if they
5436 cannot elapse again. This is particularly useful for
5437 transient timer units, which shall not stay around longer
5438 than until they first elapse.
5440 * systemd will now bump the net.unix.max_dgram_qlen to 512 by
5441 default now (the kernel default is 16). This is beneficial
5442 for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
5443 allows substantially larger numbers of queued
5444 datagrams. This should increase the capability of systemd to
5445 parallelize boot-up, as logging and sd_notify() are unlikely
5446 to stall execution anymore. If you need to change the value
5447 from the new defaults, use the usual sysctl.d/ snippets.
5449 * The compression framing format used by the journal or
5450 coredump processing has changed to be in line with what the
5451 official LZ4 tools generate. LZ4 compression support in
5452 systemd was considered unsupported previously, as the format
5453 was not compatible with the normal tools. With this release
5454 this has changed now, and it is hence safe for downstream
5455 distributions to turn it on. While not compressing as well
5456 as the XZ, LZ4 is substantially faster, which makes
5457 it a good default choice for the compression logic in the
5458 journal and in coredump handling.
5460 * Any reference to /etc/mtab has been dropped from
5461 systemd. The file has been obsolete since a while, but
5462 systemd refused to work on systems where it was incorrectly
5463 set up (it should be a symlink or non-existent). Please make
5464 sure to update to util-linux 2.27.1 or newer in conjunction
5465 with this systemd release, which also drops any reference to
5466 /etc/mtab. If you maintain a distribution make sure that no
5467 software you package still references it, as this is a
5468 likely source of bugs. There's also a glibc bug pending,
5469 asking for removal of any reference to this obsolete file:
5471 https://sourceware.org/bugzilla/show_bug.cgi?id=19108
5473 Note that only util-linux versions built with
5474 --enable-libmount-force-mountinfo are supported.
5476 * Support for the ".snapshot" unit type has been removed. This
5477 feature turned out to be little useful and little used, and
5478 has now been removed from the core and from systemctl.
5480 * The dependency types RequiresOverridable= and
5481 RequisiteOverridable= have been removed from systemd. They
5482 have been used only very sparingly to our knowledge and
5483 other options that provide a similar effect (such as
5484 systemctl --mode=ignore-dependencies) are much more useful
5485 and commonly used. Moreover, they were only half-way
5486 implemented as the option to control behaviour regarding
5487 these dependencies was never added to systemctl. By removing
5488 these dependency types the execution engine becomes a bit
5489 simpler. Unit files that use these dependencies should be
5490 changed to use the non-Overridable dependency types
5491 instead. In fact, when parsing unit files with these
5492 options, that's what systemd will automatically convert them
5493 too, but it will also warn, asking users to fix the unit
5494 files accordingly. Removal of these dependency types should
5495 only affect a negligible number of unit files in the wild.
5497 * Behaviour of networkd's IPForward= option changed
5498 (again). It will no longer maintain a per-interface setting,
5499 but propagate one way from interfaces where this is enabled
5500 to the global kernel setting. The global setting will be
5501 enabled when requested by a network that is set up, but
5502 never be disabled again. This change was made to make sure
5503 IPv4 and IPv6 behaviour regarding packet forwarding is
5504 similar (as the Linux IPv6 stack does not support
5505 per-interface control of this setting) and to minimize
5508 * In unit files the behaviour of %u, %U, %h, %s has
5509 changed. These specifiers will now unconditionally resolve
5510 to the various user database fields of the user that the
5511 systemd instance is running as, instead of the user
5512 configured in the specific unit via User=. Note that this
5513 effectively doesn't change much, as resolving of these
5514 specifiers was already turned off in the --system instance
5515 of systemd, as we cannot do NSS lookups from PID 1. In the
5516 --user instance of systemd these specifiers where correctly
5517 resolved, but hardly made any sense, since the user instance
5518 lacks privileges to do user switches anyway, and User= is
5519 hence useless. Moreover, even in the --user instance of
5520 systemd behaviour was awkward as it would only take settings
5521 from User= assignment placed before the specifier into
5522 account. In order to unify and simplify the logic around
5523 this the specifiers will now always resolve to the
5524 credentials of the user invoking the manager (which in case
5525 of PID 1 is the root user).
5527 Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
5528 Yang, Daniel Machon, Daniel Mack, David Herrmann, David
5529 Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
5530 Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
5531 Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
5532 Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
5533 Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
5534 Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
5535 Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
5536 Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
5537 Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
5538 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
5539 Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
5540 Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
5543 — Berlin, 2015-11-18
5547 * systemd now depends on util-linux v2.27. More specifically,
5548 the newly added mount monitor feature in libmount now
5549 replaces systemd's former own implementation.
5551 * libmount mandates /etc/mtab not to be regular file, and
5552 systemd now enforces this condition at early boot.
5553 /etc/mtab has been deprecated and warned about for a very
5554 long time, so systems running systemd should already have
5555 stopped having this file around as anything else than a
5556 symlink to /proc/self/mounts.
5558 * Support for the "pids" cgroup controller has been added. It
5559 allows accounting the number of tasks in a cgroup and
5560 enforcing limits on it. This adds two new setting
5561 TasksAccounting= and TasksMax= to each unit, as well as a
5562 global option DefaultTasksAccounting=.
5564 * Support for the "net_cls" cgroup controller has been added.
5565 It allows assigning a net class ID to each task in the
5566 cgroup, which can then be used in firewall rules and traffic
5567 shaping configurations. Note that the kernel netfilter net
5568 class code does not currently work reliably for ingress
5569 packets on unestablished sockets.
5571 This adds a new config directive called NetClass= to CGroup
5572 enabled units. Allowed values are positive numbers for fixed
5573 assignments and "auto" for picking a free value
5576 * 'systemctl is-system-running' now returns 'offline' if the
5577 system is not booted with systemd. This command can now be
5578 used as a substitute for 'systemd-notify --booted'.
5580 * Watchdog timeouts have been increased to 3 minutes for all
5581 in-tree service files. Apparently, disk IO issues are more
5582 frequent than we hoped, and user reported >1 minute waiting
5585 * 'machine-id-commit' functionality has been merged into
5586 'machine-id-setup --commit'. The separate binary has been
5589 * The WorkingDirectory= directive in unit files may now be set
5590 to the special value '~'. In this case, the working
5591 directory is set to the home directory of the user
5592 configured in User=.
5594 * "machinectl shell" will now open the shell in the home
5595 directory of the selected user by default.
5597 * The CrashChVT= configuration file setting is renamed to
5598 CrashChangeVT=, following our usual logic of not
5599 abbreviating unnecessarily. The old directive is still
5600 supported for compat reasons. Also, this directive now takes
5601 an integer value between 1 and 63, or a boolean value. The
5602 formerly supported '-1' value for disabling stays around for
5605 * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
5606 NoNewPrivileges=, TTYPath=, WorkingDirectory= and
5607 RootDirectory= properties can now be set for transient
5610 * The systemd-analyze tool gained a new "set-log-target" verb
5611 to change the logging target the system manager logs to
5612 dynamically during runtime. This is similar to how
5613 "systemd-analyze set-log-level" already changes the log
5616 * In nspawn /sys is now mounted as tmpfs, with only a selected
5617 set of subdirectories mounted in from the real sysfs. This
5618 enhances security slightly, and is useful for ensuring user
5619 namespaces work correctly.
5621 * Support for USB FunctionFS activation has been added. This
5622 allows implementation of USB gadget services that are
5623 activated as soon as they are requested, so that they don't
5624 have to run continuously, similar to classic socket
5627 * The "systemctl exit" command now optionally takes an
5628 additional parameter that sets the exit code to return from
5629 the systemd manager when exiting. This is only relevant when
5630 running the systemd user instance, or when running the
5631 system instance in a container.
5633 * sd-bus gained the new API calls sd_bus_path_encode_many()
5634 and sd_bus_path_decode_many() that allow easy encoding and
5635 decoding of multiple identifier strings inside a D-Bus
5636 object path. Another new call sd_bus_default_flush_close()
5637 has been added to flush and close per-thread default
5640 * systemd-cgtop gained support for a -M/--machine= switch to
5641 show the control groups within a certain container only.
5643 * "systemctl kill" gained support for an optional --fail
5644 switch. If specified the requested operation will fail of no
5645 processes have been killed, because the unit had no
5646 processes attached, or similar.
5648 * A new systemd.crash_reboot=1 kernel command line option has
5649 been added that triggers a reboot after crashing. This can
5650 also be set through CrashReboot= in systemd.conf.
5652 * The RuntimeDirectory= setting now understands unit
5653 specifiers like %i or %f.
5655 * A new (still internal) library API sd-ipv4acd has been added,
5656 that implements address conflict detection for IPv4. It's
5657 based on code from sd-ipv4ll, and will be useful for
5658 detecting DHCP address conflicts.
5660 * File descriptors passed during socket activation may now be
5661 named. A new API sd_listen_fds_with_names() is added to
5662 access the names. The default names may be overridden,
5663 either in the .socket file using the FileDescriptorName=
5664 parameter, or by passing FDNAME= when storing the file
5665 descriptors using sd_notify().
5667 * systemd-networkd gained support for:
5669 - Setting the IPv6 Router Advertisement settings via
5670 IPv6AcceptRouterAdvertisements= in .network files.
5672 - Configuring the HelloTimeSec=, MaxAgeSec= and
5673 ForwardDelaySec= bridge parameters in .netdev files.
5675 - Configuring PreferredSource= for static routes in
5678 * The "ask-password" framework used to query for LUKS harddisk
5679 passwords or SSL passwords during boot gained support for
5680 caching passwords in the kernel keyring, if it is
5681 available. This makes sure that the user only has to type in
5682 a passphrase once if there are multiple objects to unlock
5683 with the same one. Previously, such password caching was
5684 available only when Plymouth was used; this moves the
5685 caching logic into the systemd codebase itself. The
5686 "systemd-ask-password" utility gained a new --keyname=
5687 switch to control which kernel keyring key to use for
5688 caching a password in. This functionality is also useful for
5689 enabling display managers such as gdm to automatically
5690 unlock the user's GNOME keyring if its passphrase, the
5691 user's password and the harddisk password are the same, if
5692 gdm-autologin is used.
5694 * When downloading tar or raw images using "machinectl
5695 pull-tar" or "machinectl pull-raw", a matching ".nspawn"
5696 file is now also downloaded, if it is available and stored
5697 next to the image file.
5699 * Units of type ".socket" gained a new boolean setting
5700 Writable= which is only useful in conjunction with
5701 ListenSpecial=. If true, enables opening the specified
5702 special file in O_RDWR mode rather than O_RDONLY mode.
5704 * systemd-rfkill has been reworked to become a singleton
5705 service that is activated through /dev/rfkill on each rfkill
5706 state change and saves the settings to disk. This way,
5707 systemd-rfkill is now compatible with devices that exist
5708 only intermittendly, and even restores state if the previous
5709 system shutdown was abrupt rather than clean.
5711 * The journal daemon gained support for vacuuming old journal
5712 files controlled by the number of files that shall remain,
5713 in addition to the already existing control by size and by
5714 date. This is useful as journal interleaving performance
5715 degrades with too many separate journal files, and allows
5716 putting an effective limit on them. The new setting defaults
5717 to 100, but this may be changed by setting SystemMaxFiles=
5718 and RuntimeMaxFiles= in journald.conf. Also, the
5719 "journalctl" tool gained the new --vacuum-files= switch to
5720 manually vacuum journal files to leave only the specified
5721 number of files in place.
5723 * udev will now create /dev/disk/by-path links for ATA devices
5724 on kernels where that is supported.
5726 * Galician, Serbian, Turkish and Korean translations were added.
5728 Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
5729 Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
5730 (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
5731 Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
5732 Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
5733 de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
5734 Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
5735 Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
5736 Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
5737 Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
5738 Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
5739 Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
5740 Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
5741 Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
5742 Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
5743 Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
5744 Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
5745 Zbigniew Jędrzejewski-Szmek, Марко М. Костић
5747 — Berlin, 2015-10-07
5751 * The DHCP implementation of systemd-networkd gained a set of
5754 - The DHCP server now supports emitting DNS and NTP
5755 information. It may be enabled and configured via
5756 EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
5757 and NTP information is enabled, but no servers are
5758 configured, the corresponding uplink information (if there
5759 is any) is propagated.
5761 - Server and client now support transmission and reception
5762 of timezone information. It can be configured via the
5763 newly introduced network options UseTimezone=,
5764 EmitTimezone=, and Timezone=. Transmission of timezone
5765 information is enabled between host and containers by
5766 default now: the container will change its local timezone
5767 to what the host has set.
5769 - Lease timeouts can now be configured via
5770 MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
5772 - The DHCP server improved on the stability of
5773 leases. Clients are more likely to get the same lease
5774 information back, even if the server loses state.
5776 - The DHCP server supports two new configuration options to
5777 control the lease address pool metrics, PoolOffset= and
5780 * The encapsulation limit of tunnels in systemd-networkd may
5781 now be configured via 'EncapsulationLimit='. It allows
5782 modifying the maximum additional levels of encapsulation
5783 that are permitted to be prepended to a packet.
5785 * systemd now supports the concept of user buses replacing
5786 session buses, if used with dbus-1.10 (and enabled via dbus
5787 --enable-user-session). It previously only supported this on
5788 kdbus-enabled systems, and this release expands this to
5789 'dbus-daemon' systems.
5791 * systemd-networkd now supports predictable interface names
5794 * systemd now optionally supports the new Linux kernel
5795 "unified" control group hierarchy. If enabled via the kernel
5796 command-line option 'systemd.unified_cgroup_hierarchy=1',
5797 systemd will try to mount the unified cgroup hierarchy
5798 directly on /sys/fs/cgroup. If not enabled, or not
5799 available, systemd will fall back to the legacy cgroup
5800 hierarchy setup, as before. Host system and containers can
5801 mix and match legacy and unified hierarchies as they
5802 wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
5803 environment variable to individually select the hierarchy to
5804 use for executed containers. By default, nspawn will use the
5805 unified hierarchy for the containers if the host uses the
5806 unified hierarchy, and the legacy hierarchy otherwise.
5807 Please note that at this point the unified hierarchy is an
5808 experimental kernel feature and is likely to change in one
5809 of the next kernel releases. Therefore, it should not be
5810 enabled by default in downstream distributions yet. The
5811 minimum required kernel version for the unified hierarchy to
5812 work is 4.2. Note that when the unified hierarchy is used
5813 for the first time delegated access to controllers is
5814 safe. Because of this systemd-nspawn containers will get
5815 access to controllers now, as will systemd user
5816 sessions. This means containers and user sessions may now
5817 manage their own resources, partitioning up what the system
5820 * A new special scope unit "init.scope" has been introduced
5821 that encapsulates PID 1 of the system. It may be used to
5822 determine resource usage and enforce resource limits on PID
5823 1 itself. PID 1 hence moved out of the root of the control
5826 * The cgtop tool gained support for filtering out kernel
5827 threads when counting tasks in a control group. Also, the
5828 count of processes is now recursively summed up by
5829 default. Two options -k and --recursive= have been added to
5830 revert to old behaviour. The tool has also been updated to
5831 work correctly in containers now.
5833 * systemd-nspawn's --bind= and --bind-ro= options have been
5834 extended to allow creation of non-recursive bind mounts.
5836 * libsystemd gained two new calls sd_pid_get_cgroup() and
5837 sd_peer_get_cgroup() which return the control group path of
5838 a process or peer of a connected AF_UNIX socket. This
5839 function call is particularly useful when implementing
5840 delegated subtrees support in the control group hierarchy.
5842 * The "sd-event" event loop API of libsystemd now supports
5843 correct dequeuing of real-time signals, without losing
5846 * When systemd requests a polkit decision when managing units it
5847 will now add additional fields to the request, including unit
5848 name and desired operation. This enables more powerful polkit
5849 policies, that make decisions depending on these parameters.
5851 * nspawn learnt support for .nspawn settings files, that may
5852 accompany the image files or directories of containers, and
5853 may contain additional settings for the container. This is
5854 an alternative to configuring container parameters via the
5855 nspawn command line.
5857 Contributions from: Cristian Rodríguez, Daniel Mack, David
5858 Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
5859 Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
5860 Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
5861 Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
5862 Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
5863 Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
5864 Andersen, Tom Gundersen, Torstein Husebø
5866 — Berlin, 2015-09-08
5870 * machinectl gained a new verb 'shell' which opens a fresh
5871 shell on the target container or the host. It is similar to
5872 the existing 'login' command of machinectl, but spawns the
5873 shell directly without prompting for username or
5874 password. The pseudo machine '.host' now refers to the local
5875 host and is used by default. Hence, 'machinectl shell' can
5876 be used as replacement for 'su -' which spawns a session as
5877 a fresh systemd unit in a way that is fully isolated from
5878 the originating session.
5880 * systemd-networkd learned to cope with private-zone DHCP
5881 options and allows other programs to query the values.
5883 * SELinux access control when enabling/disabling units is no
5884 longer enforced with this release. The previous implementation
5885 was incorrect, and a new corrected implementation is not yet
5886 available. As unit file operations are still protected via
5887 polkit and D-Bus policy this is not a security problem. Yet,
5888 distributions which care about optimal SELinux support should
5889 probably not stabilize on this release.
5891 * sd-bus gained support for matches of type "arg0has=", that
5892 test for membership of strings in string arrays sent in bus
5895 * systemd-resolved now dumps the contents of its DNS and LLMNR
5896 caches to the logs on reception of the SIGUSR1 signal. This
5897 is useful to debug DNS behaviour.
5899 * The coredumpctl tool gained a new --directory= option to
5900 operate on journal files in a specific directory.
5902 * "systemctl reboot" and related commands gained a new
5903 "--message=" option which may be used to set a free-text
5904 wall message when shutting down or rebooting the
5905 system. This message is also logged, which is useful for
5906 figuring out the reason for a reboot or shutdown a
5909 * The "systemd-resolve-host" tool's -i switch now takes
5910 network interface numbers as alternative to interface names.
5912 * A new unit file setting for services has been introduced:
5913 UtmpMode= allows configuration of how precisely systemd
5914 handles utmp and wtmp entries for the service if this is
5915 enabled. This allows writing services that appear similar to
5916 user sessions in the output of the "w", "who", "last" and
5919 * systemd-resolved will now locally synthesize DNS resource
5920 records for the "localhost" and "gateway" domains as well as
5921 the local hostname. This should ensure that clients querying
5922 RRs via resolved will get similar results as those going via
5923 NSS, if nss-myhostname is enabled.
5925 Contributions from: Alastair Hughes, Alex Crawford, Daniel
5926 Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
5927 Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
5928 Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
5929 Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
5930 Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
5931 Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
5932 Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
5933 reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
5934 Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
5935 Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
5936 WaLyong Cho, Zbigniew Jędrzejewski-Szmek
5938 — Berlin, 2015-08-27
5942 * The systemd-efi-boot-generator functionality was merged into
5943 systemd-gpt-auto-generator.
5945 * systemd-networkd now supports Group Policy for vxlan
5946 devices. It can be enabled via the new boolean configuration
5947 option called 'GroupPolicyExtension='.
5949 Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
5950 Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
5951 Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
5953 — Berlin, 2015-07-31
5957 * The python-systemd code has been removed from the systemd repository.
5958 A new repository has been created which accommodates the code from
5959 now on, and we kindly ask distributions to create a separate package
5960 for this: https://github.com/systemd/python-systemd
5962 * The systemd daemon will now reload its main configuration
5963 (/etc/systemd/system.conf) on daemon-reload.
5965 * sd-dhcp now exposes vendor specific extensions via
5966 sd_dhcp_lease_get_vendor_specific().
5968 * systemd-networkd gained a number of new configuration options.
5970 - A new boolean configuration option for TAP devices called
5971 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
5972 device, thus allowing to send and receive GSO packets.
5974 - A new tunnel configuration option called 'CopyDSCP='.
5975 If enabled, the DSCP field of ip6 tunnels is copied into the
5976 decapsulated packet.
5978 - A set of boolean bridge configuration options were added.
5979 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
5980 and 'UnicastFlood=' are now parsed by networkd and applied to the
5981 respective bridge link device via the respective IFLA_BRPORT_*
5984 - A new string configuration option to override the hostname sent
5985 to a DHCP server, called 'Hostname='. If set and 'SendHostname='
5986 is true, networkd will use the configured hostname instead of the
5987 system hostname when sending DHCP requests.
5989 - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
5990 networkd will configure the IPv6 flow-label of the tunnel device
5991 according to RFC2460.
5993 - The 'macvtap' virtual network devices are now supported, similar to
5994 the already supported 'macvlan' devices.
5996 * systemd-resolved now implements RFC5452 to improve resilience against
5997 cache poisoning. Additionally, source port randomization is enabled
5998 by default to further protect against DNS spoofing attacks.
6000 * nss-mymachines now supports translating UIDs and GIDs of running
6001 containers with user-namespaces enabled. If a container 'foo'
6002 translates a host uid 'UID' to the container uid 'TUID', then
6003 nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
6004 (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
6005 mapped as 'vg-foo-TGID'.
6007 Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
6008 Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
6009 HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
6010 Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
6011 Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
6012 Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
6013 Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
6014 Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
6015 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
6016 Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
6018 — Berlin, 2015-07-29
6022 * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
6023 There are no known issues with current sysfs, and udev does not need
6024 or should be used to work around such bugs.
6026 * udev does no longer enable USB HID power management. Several reports
6027 indicate, that some devices cannot handle that setting.
6029 * The udev accelerometer helper was removed. The functionality
6030 is now fully included in iio-sensor-proxy. But this means,
6031 older iio-sensor-proxy versions will no longer provide
6032 accelerometer/orientation data with this systemd version.
6033 Please upgrade iio-sensor-proxy to version 1.0.
6035 * networkd gained a new configuration option IPv6PrivacyExtensions=
6036 which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
6037 for Stateless Address") on selected networks.
6039 * For the sake of fewer build-time dependencies and less code in the
6040 main repository, the python bindings are about to be removed in the
6041 next release. A new repository has been created which accommodates
6042 the code from now on, and we kindly ask distributions to create a
6043 separate package for this. The removal will take place in v223.
6045 https://github.com/systemd/python-systemd
6047 Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
6048 Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
6049 daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
6050 Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
6051 Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
6052 (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
6053 Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
6054 Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
6055 Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
6056 Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
6058 — Berlin, 2015-07-07
6062 * The sd-bus.h and sd-event.h APIs have now been declared
6063 stable and have been added to the official interface of
6064 libsystemd.so. sd-bus implements an alternative D-Bus client
6065 library, that is relatively easy to use, very efficient and
6066 supports both classic D-Bus as well as kdbus as transport
6067 backend. sd-event is a generic event loop abstraction that
6068 is built around Linux epoll, but adds features such as event
6069 prioritization or efficient timer handling. Both APIs are good
6070 choices for C programs looking for a bus and/or event loop
6071 implementation that is minimal and does not have to be
6072 portable to other kernels.
6074 * kdbus support is no longer compile-time optional. It is now
6075 always built-in. However, it can still be disabled at
6076 runtime using the kdbus=0 kernel command line setting, and
6077 that setting may be changed to default to off, by specifying
6078 --disable-kdbus at build-time. Note though that the kernel
6079 command line setting has no effect if the kdbus.ko kernel
6080 module is not installed, in which case kdbus is (obviously)
6081 also disabled. We encourage all downstream distributions to
6082 begin testing kdbus by adding it to the kernel images in the
6083 development distributions, and leaving kdbus support in
6086 * The minimal required util-linux version has been bumped to
6089 * Support for chkconfig (--enable-chkconfig) was removed in
6090 favor of calling an abstraction tool
6091 /lib/systemd/systemd-sysv-install. This needs to be
6092 implemented for your distribution. See "SYSV INIT.D SCRIPTS"
6093 in README for details.
6095 * If there's a systemd unit and a SysV init script for the
6096 same service name, and the user executes "systemctl enable"
6097 for it (or a related call), then this will now enable both
6098 (or execute the related operation on both), not just the
6101 * The libudev API documentation has been converted from gtkdoc
6104 * gudev has been removed from the systemd tree, it is now an
6107 * The systemd-cgtop tool learnt a new --raw switch to generate
6108 "raw" (machine parsable) output.
6110 * networkd's IPForwarding= .network file setting learnt the
6111 new setting "kernel", which ensures that networkd does not
6112 change the IP forwarding sysctl from the default kernel
6115 * The systemd-logind bus API now exposes a new boolean
6116 property "Docked" that reports whether logind considers the
6117 system "docked", i.e. connected to a docking station or not.
6119 Contributions from: Alex Crawford, Andreas Pokorny, Andrei
6120 Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
6121 Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
6122 David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
6123 Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
6124 Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
6125 Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
6126 Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
6127 Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
6128 Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
6129 Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
6130 Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
6131 Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
6132 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
6133 Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
6134 Fink, Zbigniew Jędrzejewski-Szmek
6136 — Berlin, 2015-06-19
6140 * The gudev library has been extracted into a separate repository
6141 available at: https://git.gnome.org/browse/libgudev/
6142 It is now managed as part of the Gnome project. Distributions
6143 are recommended to pass --disable-gudev to systemd and use
6144 gudev from the Gnome project instead. gudev is still included
6145 in systemd, for now. It will be removed soon, though. Please
6146 also see the announcement-thread on systemd-devel:
6147 https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html
6149 * systemd now exposes a CPUUsageNSec= property for each
6150 service unit on the bus, that contains the overall consumed
6151 CPU time of a service (the sum of what each process of the
6152 service consumed). This value is only available if
6153 CPUAccounting= is turned on for a service, and is then shown
6154 in the "systemctl status" output.
6156 * Support for configuring alternative mappings of the old SysV
6157 runlevels to systemd targets has been removed. They are now
6158 hardcoded in a way that runlevels 2, 3, 4 all map to
6159 multi-user.target and 5 to graphical.target (which
6160 previously was already the default behaviour).
6162 * The auto-mounter logic gained support for mount point
6163 expiry, using a new TimeoutIdleSec= setting in .automount
6164 units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
6166 * The EFI System Partition (ESP) as mounted to /boot by
6167 systemd-efi-boot-generator will now be unmounted
6168 automatically after 2 minutes of not being used. This should
6169 minimize the risk of ESP corruptions.
6171 * New /etc/fstab options x-systemd.requires= and
6172 x-systemd.requires-mounts-for= are now supported to express
6173 additional dependencies for mounts. This is useful for
6174 journalling file systems that support external journal
6175 devices or overlay file systems that require underlying file
6176 systems to be mounted.
6178 * systemd does not support direct live-upgrades (via systemctl
6179 daemon-reexec) from versions older than v44 anymore. As no
6180 distribution we are aware of shipped such old versions in a
6181 stable release this should not be problematic.
6183 * When systemd forks off a new per-connection service instance
6184 it will now set the $REMOTE_ADDR environment variable to the
6185 remote IP address, and $REMOTE_PORT environment variable to
6186 the remote IP port. This behaviour is similar to the
6187 corresponding environment variables defined by CGI.
6189 * systemd-networkd gained support for uplink failure
6190 detection. The BindCarrier= option allows binding interface
6191 configuration dynamically to the link sense of other
6192 interfaces. This is useful to achieve behaviour like in
6195 * systemd-networkd gained support for configuring the DHCP
6196 client identifier to use when requesting leases.
6198 * systemd-networkd now has a per-network UseNTP= option to
6199 configure whether NTP server information acquired via DHCP
6200 is passed on to services like systemd-timesyncd.
6202 * systemd-networkd gained support for vti6 tunnels.
6204 * Note that systemd-networkd manages the sysctl variable
6205 /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
6206 it is configured for since v219. The variable controls IP
6207 forwarding, and is a per-interface alternative to the global
6208 /proc/sys/net/ipv[46]/ip_forward. This setting is
6209 configurable in the IPForward= option, which defaults to
6210 "no". This means if networkd is used for an interface it is
6211 no longer sufficient to set the global sysctl option to turn
6212 on IP forwarding! Instead, the .network file option
6213 IPForward= needs to be turned on! Note that the
6214 implementation of this behaviour was broken in v219 and has
6217 * Many bonding and vxlan options are now configurable in
6220 * systemd-nspawn gained a new --property= setting to set unit
6221 properties for the container scope. This is useful for
6222 setting resource parameters (e.g. "CPUShares=500") on
6223 containers started from the command line.
6225 * systemd-nspawn gained a new --private-users= switch to make
6226 use of user namespacing available on recent Linux kernels.
6228 * systemd-nspawn may now be called as part of a shell pipeline
6229 in which case the pipes used for stdin and stdout are passed
6230 directly to the process invoked in the container, without
6231 indirection via a pseudo tty.
6233 * systemd-nspawn gained a new switch to control the UNIX
6234 signal to use when killing the init process of the container
6237 * systemd-nspawn gained a new --overlay= switch for mounting
6238 overlay file systems into the container using the new kernel
6241 * When a container image is imported via systemd-importd and
6242 the host file system is not btrfs, a loopback block device
6243 file is created in /var/lib/machines.raw with a btrfs file
6244 system inside. It is then mounted to /var/lib/machines to
6245 enable btrfs features for container management. The loopback
6246 file and btrfs file system is grown as needed when container
6247 images are imported via systemd-importd.
6249 * systemd-machined/systemd-importd gained support for btrfs
6250 quota, to enforce container disk space limits on disk. This
6251 is exposed in "machinectl set-limit".
6253 * systemd-importd now can import containers from local .tar,
6254 .raw and .qcow2 images, and export them to .tar and .raw. It
6255 can also import dkr v2 images now from the network (on top
6258 * systemd-importd gained support for verifying downloaded
6259 images with gpg2 (previously only gpg1 was supported).
6261 * systemd-machined, systemd-logind, systemd: most bus calls are
6262 now accessible to unprivileged processes via polkit. Also,
6263 systemd-logind will now allow users to kill their own sessions
6264 without further privileges or authorization.
6266 * systemd-shutdownd has been removed. This service was
6267 previously responsible for implementing scheduled shutdowns
6268 as exposed in /usr/bin/shutdown's time parameter. This
6269 functionality has now been moved into systemd-logind and is
6270 accessible via a bus interface.
6272 * "systemctl reboot" gained a new switch --firmware-setup that
6273 can be used to reboot into the EFI firmware setup, if that
6274 is available. systemd-logind now exposes an API on the bus
6275 to trigger such reboots, in case graphical desktop UIs want
6276 to cover this functionality.
6278 * "systemctl enable", "systemctl disable" and "systemctl mask"
6279 now support a new "--now" switch. If specified the units
6280 that are enabled will also be started, and the ones
6281 disabled/masked also stopped.
6283 * The Gummiboot EFI boot loader tool has been merged into
6284 systemd, and renamed to "systemd-boot". The bootctl tool has been
6285 updated to support systemd-boot.
6287 * An EFI kernel stub has been added that may be used to create
6288 kernel EFI binaries that contain not only the actual kernel,
6289 but also an initrd, boot splash, command line and OS release
6290 information. This combined binary can then be signed as a
6291 single image, so that the firmware can verify it all in one
6292 step. systemd-boot has special support for EFI binaries created
6293 like this and can extract OS release information from them
6294 and show them in the boot menu. This functionality is useful
6295 to implement cryptographically verified boot schemes.
6297 * Optional support has been added to systemd-fsck to pass
6298 fsck's progress report to an AF_UNIX socket in the file
6301 * udev will no longer create device symlinks for all block devices by
6302 default. A deny list for excluding special block devices from this
6303 logic has been turned into a allow list that requires picking block
6304 devices explicitly that require device symlinks.
6306 * A new (currently still internal) API sd-device.h has been
6307 added to libsystemd. This modernized API is supposed to
6308 replace libudev eventually. In fact, already much of libudev
6309 is now just a wrapper around sd-device.h.
6311 * A new hwdb database for storing metadata about pointing
6312 stick devices has been added.
6314 * systemd-tmpfiles gained support for setting file attributes
6315 similar to the "chattr" tool with new 'h' and 'H' lines.
6317 * systemd-journald will no longer unconditionally set the
6318 btrfs NOCOW flag on new journal files. This is instead done
6319 with tmpfiles snippet using the new 'h' line type. This
6320 allows easy disabling of this logic, by masking the
6321 journal-nocow.conf tmpfiles file.
6323 * systemd-journald will now translate audit message types to
6324 human readable identifiers when writing them to the
6325 journal. This should improve readability of audit messages.
6327 * The LUKS logic gained support for the offset= and skip=
6328 options in /etc/crypttab, as previously implemented by
6331 * /usr/lib/os-release gained a new optional field VARIANT= for
6332 distributions that support multiple variants (such as a
6333 desktop edition, a server edition, ...)
6335 Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
6336 Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
6337 Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
6338 Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
6339 Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
6340 Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
6341 Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
6342 Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
6343 Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
6344 Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
6345 Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
6346 Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
6347 Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
6348 Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
6349 De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
6350 Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
6351 Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
6352 Michael Biebl, Michael Marineau, Michael Olbrich, Michal
6353 Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
6354 Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
6355 Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
6356 Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
6357 Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
6358 Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
6359 Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
6360 Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
6361 Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
6363 — Berlin, 2015-05-22
6367 * Introduce a new API "sd-hwdb.h" for querying the hardware
6368 metadata database. With this minimal interface one can query
6369 and enumerate the udev hwdb, decoupled from the old libudev
6370 library. libudev's interface for this is now only a wrapper
6371 around sd-hwdb. A new tool systemd-hwdb has been added to
6372 interface with and update the database.
6374 * When any of systemd's tools copies files (for example due to
6375 tmpfiles' C lines) a btrfs reflink will attempted first,
6376 before bytewise copying is done.
6378 * systemd-nspawn gained a new --ephemeral switch. When
6379 specified a btrfs snapshot is taken of the container's root
6380 directory, and immediately removed when the container
6381 terminates again. Thus, a container can be started whose
6382 changes never alter the container's root directory, and are
6383 lost on container termination. This switch can also be used
6384 for starting a container off the root file system of the
6385 host without affecting the host OS. This switch is only
6386 available on btrfs file systems.
6388 * systemd-nspawn gained a new --template= switch. It takes the
6389 path to a container tree to use as template for the tree
6390 specified via --directory=, should that directory be
6391 missing. This allows instantiating containers dynamically,
6392 on first run. This switch is only available on btrfs file
6395 * When a .mount unit refers to a mount point on which multiple
6396 mounts are stacked, and the .mount unit is stopped all of
6397 the stacked mount points will now be unmounted until no
6398 mount point remains.
6400 * systemd now has an explicit notion of supported and
6401 unsupported unit types. Jobs enqueued for unsupported unit
6402 types will now fail with an "unsupported" error code. More
6403 specifically .swap, .automount and .device units are not
6404 supported in containers, .busname units are not supported on
6405 non-kdbus systems. .swap and .automount are also not
6406 supported if their respective kernel compile time options
6409 * machinectl gained support for two new "copy-from" and
6410 "copy-to" commands for copying files from a running
6411 container to the host or vice versa.
6413 * machinectl gained support for a new "bind" command to bind
6414 mount host directories into local containers. This is
6415 currently only supported for nspawn containers.
6417 * networkd gained support for configuring bridge forwarding
6418 database entries (fdb) from .network files.
6420 * A new tiny daemon "systemd-importd" has been added that can
6421 download container images in tar, raw, qcow2 or dkr formats,
6422 and make them available locally in /var/lib/machines, so
6423 that they can run as nspawn containers. The daemon can GPG
6424 verify the downloads (not supported for dkr, since it has no
6425 provisions for verifying downloads). It will transparently
6426 decompress bz2, xz, gzip compressed downloads if necessary,
6427 and restore sparse files on disk. The daemon uses privilege
6428 separation to ensure the actual download logic runs with
6429 fewer privileges than the daemon itself. machinectl has
6430 gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
6431 make the functionality of importd available to the
6432 user. With this in place the Fedora and Ubuntu "Cloud"
6433 images can be downloaded and booted as containers unmodified
6434 (the Fedora images lack the appropriate GPG signature files
6435 currently, so they cannot be verified, but this will change
6436 soon, hopefully). Note that downloading images is currently
6437 only fully supported on btrfs.
6439 * machinectl is now able to list container images found in
6440 /var/lib/machines, along with some metadata about sizes of
6441 disk and similar. If the directory is located on btrfs and
6442 quota is enabled, this includes quota display. A new command
6443 "image-status" has been added that shows additional
6444 information about images.
6446 * machinectl is now able to clone container images
6447 efficiently, if the underlying file system (btrfs) supports
6448 it, with the new "machinectl clone" command. It also
6449 gained commands for renaming and removing images, as well as
6450 marking them read-only or read-write (supported also on
6451 legacy file systems).
6453 * networkd gained support for collecting LLDP network
6454 announcements, from hardware that supports this. This is
6455 shown in networkctl output.
6457 * systemd-run gained support for a new -t (--pty) switch for
6458 invoking a binary on a pty whose input and output is
6459 connected to the invoking terminal. This allows executing
6460 processes as system services while interactively
6461 communicating with them via the terminal. Most interestingly
6462 this is supported across container boundaries. Invoking
6463 "systemd-run -t /bin/bash" is an alternative to running a
6464 full login session, the difference being that the former
6465 will not register a session, nor go through the PAM session
6468 * tmpfiles gained support for a new "v" line type for creating
6469 btrfs subvolumes. If the underlying file system is a legacy
6470 file system, this automatically degrades to creating a
6471 normal directory. Among others /var/lib/machines is now
6472 created like this at boot, should it be missing.
6474 * The directory /var/lib/containers/ has been deprecated and
6475 been replaced by /var/lib/machines. The term "machines" has
6476 been used in the systemd context as generic term for both
6477 VMs and containers, and hence appears more appropriate for
6478 this, as the directory can also contain raw images bootable
6481 * systemd-nspawn when invoked with -M but without --directory=
6482 or --image= is now capable of searching for the container
6483 root directory, subvolume or disk image automatically, in
6484 /var/lib/machines. systemd-nspawn@.service has been updated
6485 to make use of this, thus allowing it to be used for raw
6488 * A new machines.target unit has been introduced that is
6489 supposed to group all containers/VMs invoked as services on
6490 the system. systemd-nspawn@.service has been updated to
6491 integrate with that.
6493 * machinectl gained a new "start" command, for invoking a
6494 container as a service. "machinectl start foo" is mostly
6495 equivalent to "systemctl start systemd-nspawn@foo.service",
6496 but handles escaping in a nicer way.
6498 * systemd-nspawn will now mount most of the cgroupfs tree
6499 read-only into each container, with the exception of the
6500 container's own subtree in the name=systemd hierarchy.
6502 * journald now sets the special FS_NOCOW file flag for its
6503 journal files. This should improve performance on btrfs, by
6504 avoiding heavy fragmentation when journald's write-pattern
6505 is used on COW file systems. It degrades btrfs' data
6506 integrity guarantees for the files to the same levels as for
6507 ext3/ext4 however. This should be OK though as journald does
6508 its own data integrity checks and all its objects are
6509 checksummed on disk. Also, journald should handle btrfs disk
6510 full events a lot more gracefully now, by processing SIGBUS
6511 errors, and not relying on fallocate() anymore.
6513 * When journald detects that journal files it is writing to
6514 have been deleted it will immediately start new journal
6517 * systemd now provides a way to store file descriptors
6518 per-service in PID 1. This is useful for daemons to ensure
6519 that fds they require are not lost during a daemon
6520 restart. The fds are passed to the daemon on the next
6521 invocation in the same way socket activation fds are
6522 passed. This is now used by journald to ensure that the
6523 various sockets connected to all the system's stdout/stderr
6524 are not lost when journald is restarted. File descriptors
6525 may be stored in PID 1 via the sd_pid_notify_with_fds() API,
6526 an extension to sd_notify(). Note that a limit is enforced
6527 on the number of fds a service can store in PID 1, and it
6528 defaults to 0, so that no fds may be stored, unless this is
6529 explicitly turned on.
6531 * The default TERM variable to use for units connected to a
6532 terminal, when no other value is explicitly is set is now
6533 vt220 rather than vt102. This should be fairly safe still,
6534 but allows PgUp/PgDn work.
6536 * The /etc/crypttab option header= as known from Debian is now
6539 * "loginctl user-status" and "loginctl session-status" will
6540 now show the last 10 lines of log messages of the
6541 user/session following the status output. Similar,
6542 "machinectl status" will show the last 10 log lines
6543 associated with a virtual machine or container
6544 service. (Note that this is usually not the log messages
6545 done in the VM/container itself, but simply what the
6546 container manager logs. For nspawn this includes all console
6549 * "loginctl session-status" without further argument will now
6550 show the status of the session of the caller. Similar,
6551 "lock-session", "unlock-session", "activate",
6552 "enable-linger", "disable-linger" may now be called without
6553 session/user parameter in which case they apply to the
6554 caller's session/user.
6556 * An X11 session scriptlet is now shipped that uploads
6557 $DISPLAY and $XAUTHORITY into the environment of the systemd
6558 --user daemon if a session begins. This should improve
6559 compatibility with X11 enabled applications run as systemd
6562 * Generators are now subject to masking via /etc and /run, the
6563 same way as unit files.
6565 * networkd .network files gained support for configuring
6566 per-link IPv4/IPv6 packet forwarding as well as IPv4
6567 masquerading. This is by default turned on for veth links to
6568 containers, as registered by systemd-nspawn. This means that
6569 nspawn containers run with --network-veth will now get
6570 automatic routed access to the host's networks without any
6571 further configuration or setup, as long as networkd runs on
6574 * systemd-nspawn gained the --port= (-p) switch to expose TCP
6575 or UDP posts of a container on the host. With this in place
6576 it is possible to run containers with private veth links
6577 (--network-veth), and have their functionality exposed on
6578 the host as if their services were running directly on the
6581 * systemd-nspawn's --network-veth switch now gained a short
6582 version "-n", since with the changes above it is now truly
6583 useful out-of-the-box. The systemd-nspawn@.service has been
6584 updated to make use of it too by default.
6586 * systemd-nspawn will now maintain a per-image R/W lock, to
6587 ensure that the same image is not started more than once
6588 writable. (It's OK to run an image multiple times
6589 simultaneously in read-only mode.)
6591 * systemd-nspawn's --image= option is now capable of
6592 dissecting and booting MBR and GPT disk images that contain
6593 only a single active Linux partition. Previously it
6594 supported only GPT disk images with proper GPT type
6595 IDs. This allows running cloud images from major
6596 distributions directly with systemd-nspawn, without
6599 * In addition to collecting mouse dpi data in the udev
6600 hardware database, there's now support for collecting angle
6601 information for mouse scroll wheels. The database is
6602 supposed to guarantee similar scrolling behavior on mice
6603 that it knows about. There's also support for collecting
6604 information about Touchpad types.
6606 * udev's input_id built-in will now also collect touch screen
6607 dimension data and attach it to probed devices.
6609 * /etc/os-release gained support for a Distribution Privacy
6612 * networkd gained support for creating "ipvlan", "gretap",
6613 "ip6gre", "ip6gretap" and "ip6tnl" network devices.
6615 * systemd-tmpfiles gained support for "a" lines for setting
6618 * systemd-nspawn will now mount /tmp in the container to
6619 tmpfs, automatically.
6621 * systemd now exposes the memory.usage_in_bytes cgroup
6622 attribute and shows it for each service in the "systemctl
6623 status" output, if available.
6625 * When the user presses Ctrl-Alt-Del more than 7x within 2s an
6626 immediate reboot is triggered. This useful if shutdown is
6627 hung and is unable to complete, to expedite the
6628 operation. Note that this kind of reboot will still unmount
6629 all file systems, and hence should not result in fsck being
6632 * A .device unit for an optical block device will now be
6633 considered active only when a medium is in the drive. Also,
6634 mount units are now bound to their backing devices thus
6635 triggering automatic unmounting when devices become
6636 unavailable. With this in place systemd will now
6637 automatically unmount left-over mounts when a CD-ROM is
6638 ejected or an USB stick is yanked from the system.
6640 * networkd-wait-online now has support for waiting for
6641 specific interfaces only (with globbing), and for giving up
6642 after a configurable timeout.
6644 * networkd now exits when idle. It will be automatically
6645 restarted as soon as interfaces show up, are removed or
6646 change state. networkd will stay around as long as there is
6647 at least one DHCP state machine or similar around, that keep
6650 * networkd may now configure IPv6 link-local addressing in
6651 addition to IPv4 link-local addressing.
6653 * The IPv6 "token" for use in SLAAC may now be configured for
6654 each .network interface in networkd.
6656 * Routes configured with networkd may now be assigned a scope
6659 * networkd's [Match] sections now support globbing and lists
6660 of multiple space-separated matches per item.
6662 Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
6663 Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
6664 Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
6665 Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
6666 Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
6667 Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
6668 Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
6669 Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
6670 Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
6671 Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
6672 Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
6673 Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
6674 Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
6675 Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
6676 Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
6677 Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
6678 Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
6679 Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
6680 Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
6681 Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
6682 Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
6683 Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
6684 Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
6685 Hoffmann, Zbigniew Jędrzejewski-Szmek
6687 — Berlin, 2015-02-16
6691 * When querying unit file enablement status (for example via
6692 "systemctl is-enabled"), a new state "indirect" is now known
6693 which indicates that a unit might not be enabled itself, but
6694 another unit listed in its Also= setting might be.
6696 * Similar to the various existing ConditionXYZ= settings for
6697 units, there are now matching AssertXYZ= settings. While
6698 failing conditions cause a unit to be skipped, but its job
6699 to succeed, failing assertions declared like this will cause
6700 a unit start operation and its job to fail.
6702 * hostnamed now knows a new chassis type "embedded".
6704 * systemctl gained a new "edit" command. When used on a unit
6705 file, this allows extending unit files with .d/ drop-in
6706 configuration snippets or editing the full file (after
6707 copying it from /usr/lib to /etc). This will invoke the
6708 user's editor (as configured with $EDITOR), and reload the
6709 modified configuration after editing.
6711 * "systemctl status" now shows the suggested enablement state
6712 for a unit, as declared in the (usually vendor-supplied)
6713 system preset files.
6715 * nss-myhostname will now resolve the single-label hostname
6716 "gateway" to the locally configured default IP routing
6717 gateways, ordered by their metrics. This assigns a stable
6718 name to the used gateways, regardless which ones are
6719 currently configured. Note that the name will only be
6720 resolved after all other name sources (if nss-myhostname is
6721 configured properly) and should hence not negatively impact
6722 systems that use the single-label hostname "gateway" in
6725 * systemd-inhibit now allows filtering by mode when listing
6728 * Scope and service units gained a new "Delegate" boolean
6729 property, which, when set, allows processes running inside the
6730 unit to further partition resources. This is primarily
6731 useful for systemd user instances as well as container
6734 * journald will now pick up audit messages directly from
6735 the kernel, and log them like any other log message. The
6736 audit fields are split up and fully indexed. This means that
6737 journalctl in many ways is now a (nicer!) alternative to
6738 ausearch, the traditional audit client. Note that this
6739 implements only a minimal audit client. If you want the
6740 special audit modes like reboot-on-log-overflow, please use
6741 the traditional auditd instead, which can be used in
6742 parallel to journald.
6744 * The ConditionSecurity= unit file option now understands the
6745 special string "audit" to check whether auditing is
6748 * journalctl gained two new commands --vacuum-size= and
6749 --vacuum-time= to delete old journal files until the
6750 remaining ones take up no more than the specified size on disk,
6751 or are not older than the specified time.
6753 * A new, native PPPoE library has been added to sd-network,
6754 systemd's library of light-weight networking protocols. This
6755 library will be used in a future version of networkd to
6756 enable PPPoE communication without an external pppd daemon.
6758 * The busctl tool now understands a new "capture" verb that
6759 works similar to "monitor", but writes a packet capture
6760 trace to STDOUT that can be redirected to a file which is
6761 compatible with libcap's capture file format. This can then
6762 be loaded in Wireshark and similar tools to inspect bus
6765 * The busctl tool now understands a new "tree" verb that shows
6766 the object trees of a specific service on the bus, or of all
6769 * The busctl tool now understands a new "introspect" verb that
6770 shows all interfaces and members of objects on the bus,
6771 including their signature and values. This is particularly
6772 useful to get more information about bus objects shown by
6773 the new "busctl tree" command.
6775 * The busctl tool now understands new verbs "call",
6776 "set-property" and "get-property" for invoking bus method
6777 calls, setting and getting bus object properties in a
6780 * busctl gained a new --augment-creds= argument that controls
6781 whether the tool shall augment credential information it
6782 gets from the bus with data from /proc, in a possibly
6785 * nspawn's --link-journal= switch gained two new values
6786 "try-guest" and "try-host" that work like "guest" and
6787 "host", but do not fail if the host has no persistent
6788 journalling enabled. -j is now equivalent to
6789 --link-journal=try-guest.
6791 * macvlan network devices created by nspawn will now have
6792 stable MAC addresses.
6794 * A new SmackProcessLabel= unit setting has been added, which
6795 controls the SMACK security label processes forked off by
6796 the respective unit shall use.
6798 * If compiled with --enable-xkbcommon, systemd-localed will
6799 verify x11 keymap settings by compiling the given keymap. It
6800 will spew out warnings if the compilation fails. This
6801 requires libxkbcommon to be installed.
6803 * When a coredump is collected, a larger number of metadata
6804 fields is now collected and included in the journal records
6805 created for it. More specifically, control group membership,
6806 environment variables, memory maps, working directory,
6807 chroot directory, /proc/$PID/status, and a list of open file
6808 descriptors is now stored in the log entry.
6810 * The udev hwdb now contains DPI information for mice. For
6813 http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html
6815 * All systemd programs that read standalone configuration
6816 files in /etc now also support a corresponding series of
6817 .conf.d configuration directories in /etc/, /run/,
6818 /usr/local/lib/, /usr/lib/, and (if configured with
6819 --enable-split-usr) /lib/. In particular, the following
6820 configuration files now have corresponding configuration
6821 directories: system.conf user.conf, logind.conf,
6822 journald.conf, sleep.conf, bootchart.conf, coredump.conf,
6823 resolved.conf, timesyncd.conf, journal-remote.conf, and
6824 journal-upload.conf. Note that distributions should use the
6825 configuration directories in /usr/lib/; the directories in
6826 /etc/ are reserved for the system administrator.
6828 * systemd-rfkill will no longer take the rfkill device name
6829 into account when storing rfkill state on disk, as the name
6830 might be dynamically assigned and not stable. Instead, the
6831 ID_PATH udev variable combined with the rfkill type (wlan,
6832 bluetooth, ...) is used.
6834 * A new service systemd-machine-id-commit.service has been
6835 added. When used on systems where /etc is read-only during
6836 boot, and /etc/machine-id is not initialized (but an empty
6837 file), this service will copy the temporary machine ID
6838 created as replacement into /etc after the system is fully
6839 booted up. This is useful for systems that are freshly
6840 installed with a non-initialized machine ID, but should get
6841 a fixed machine ID for subsequent boots.
6843 * networkd's .netdev files now provide a large set of
6844 configuration parameters for VXLAN devices. Similarly, the
6845 bridge port cost parameter is now configurable in .network
6846 files. There's also new support for configuring IP source
6847 routing. networkd .link files gained support for a new
6848 OriginalName= match that is useful to match against the
6849 original interface name the kernel assigned. .network files
6850 may include MTU= and MACAddress= fields for altering the MTU
6851 and MAC address while being connected to a specific network
6854 * The LUKS logic gained supported for configuring
6855 UUID-specific key files. There's also new support for naming
6856 LUKS device from the kernel command line, using the new
6857 luks.name= argument.
6859 * Timer units may now be transiently created via the bus API
6860 (this was previously already available for scope and service
6861 units). In addition it is now possible to create multiple
6862 transient units at the same time with a single bus call. The
6863 "systemd-run" tool has been updated to make use of this for
6864 running commands on a specified time, in at(1)-style.
6866 * tmpfiles gained support for "t" lines, for assigning
6867 extended attributes to files. Among other uses this may be
6868 used to assign SMACK labels to files.
6870 Contributions from: Alin Rauta, Alison Chaiken, Andrej
6871 Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
6872 Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
6873 Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
6874 Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
6875 Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
6876 Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
6877 Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
6878 Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
6879 Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
6880 Michael Biebl, Michael Chapman, Michael Marineau, Michal
6881 Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
6882 Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
6883 Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
6884 Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
6885 Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
6886 Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
6887 Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
6889 — Berlin, 2014-12-10
6893 * journalctl gained the new options -t/--identifier= to match
6894 on the syslog identifier (aka "tag"), as well as --utc to
6895 show log timestamps in the UTC timezone. journalctl now also
6896 accepts -n/--lines=all to disable line capping in a pager.
6898 * journalctl gained a new switch, --flush, that synchronously
6899 flushes logs from /run/log/journal to /var/log/journal if
6900 persistent storage is enabled. systemd-journal-flush.service
6901 now waits until the operation is complete.
6903 * Services can notify the manager before they start a reload
6904 (by sending RELOADING=1) or shutdown (by sending
6905 STOPPING=1). This allows the manager to track and show the
6906 internal state of daemons and closes a race condition when
6907 the process is still running but has closed its D-Bus
6910 * Services with Type=oneshot do not have to have any ExecStart
6913 * User units are now loaded also from
6914 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
6915 /run/systemd/user directory that was already previously
6916 supported, but is under the control of the user.
6918 * Job timeouts (i.e. timeouts on the time a job that is
6919 queued stays in the run queue) can now optionally result in
6920 immediate reboot or power-off actions (JobTimeoutAction= and
6921 JobTimeoutRebootArgument=). This is useful on ".target"
6922 units, to limit the maximum time a target remains
6923 undispatched in the run queue, and to trigger an emergency
6924 operation in such a case. This is now used by default to
6925 turn off the system if boot-up (as defined by everything in
6926 basic.target) hangs and does not complete for at least
6927 15min. Also, if power-off or reboot hang for at least 30min
6928 an immediate power-off/reboot operation is triggered. This
6929 functionality is particularly useful to increase reliability
6930 on embedded devices, but also on laptops which might
6931 accidentally get powered on when carried in a backpack and
6932 whose boot stays stuck in a hard disk encryption passphrase
6935 * systemd-logind can be configured to also handle lid switch
6936 events even when the machine is docked or multiple displays
6937 are attached (HandleLidSwitchDocked= option).
6939 * A helper binary and a service have been added which can be
6940 used to resume from hibernation in the initramfs. A
6941 generator will parse the resume= option on the kernel
6942 command line to trigger resume.
6944 * A user console daemon systemd-consoled has been
6945 added. Currently, it is a preview, and will so far open a
6946 single terminal on each session of the user marked as
6947 Desktop=systemd-console.
6949 * Route metrics can be specified for DHCP routes added by
6952 * The SELinux context of socket-activated services can be set
6953 from the information provided by the networking stack
6954 (SELinuxContextFromNet= option).
6956 * Userspace firmware loading support has been removed and
6957 the minimum supported kernel version is thus bumped to 3.7.
6959 * Timeout for udev workers has been increased from 1 to 3
6960 minutes, but a warning will be printed after 1 minute to
6961 help diagnose kernel modules that take a long time to load.
6963 * Udev rules can now remove tags on devices with TAG-="foobar".
6965 * systemd's readahead implementation has been removed. In many
6966 circumstances it didn't give expected benefits even for
6967 rotational disk drives and was becoming less relevant in the
6968 age of SSDs. As none of the developers has been using
6969 rotating media anymore, and nobody stepped up to actively
6970 maintain this component of systemd it has now been removed.
6972 * Swap units can use Options= to specify discard options.
6973 Discard options specified for swaps in /etc/fstab are now
6976 * Docker containers are now detected as a separate type of
6979 * The Password Agent protocol gained support for queries where
6980 the user input is shown, useful e.g. for user names.
6981 systemd-ask-password gained a new --echo option to turn that
6984 * The default sysctl.d/ snippets will now set:
6986 net.core.default_qdisc = fq_codel
6988 This selects Fair Queuing Controlled Delay as the default
6989 queuing discipline for network interfaces. fq_codel helps
6990 fight the network bufferbloat problem. It is believed to be
6991 a good default with no tuning required for most workloads.
6992 Downstream distributions may override this choice. On 10Gbit
6993 servers that do not do forwarding, "fq" may perform better.
6994 Systems without a good clocksource should use "pfifo_fast".
6996 * If kdbus is enabled during build a new option BusPolicy= is
6997 available for service units, that allows locking all service
6998 processes into a stricter bus policy, in order to limit
6999 access to various bus services, or even hide most of them
7000 from the service's view entirely.
7002 * networkctl will now show the .network and .link file
7003 networkd has applied to a specific interface.
7005 * sd-login gained a new API call sd_session_get_desktop() to
7006 query which desktop environment has been selected for a
7009 * UNIX utmp support is now compile-time optional to support
7010 legacy-free systems.
7012 * systemctl gained two new commands "add-wants" and
7013 "add-requires" for pulling in units from specific targets
7016 * If the word "rescue" is specified on the kernel command line
7017 the system will now boot into rescue mode (aka
7018 rescue.target), which was previously available only by
7019 specifying "1" or "systemd.unit=rescue.target" on the kernel
7020 command line. This new kernel command line option nicely
7021 mirrors the already existing "emergency" kernel command line
7024 * New kernel command line options mount.usr=, mount.usrflags=,
7025 mount.usrfstype= have been added that match root=, rootflags=,
7026 rootfstype= but allow mounting a specific file system to
7029 * The $NOTIFY_SOCKET is now also passed to control processes of
7030 services, not only the main process.
7032 * This version reenables support for fsck's -l switch. This
7033 means at least version v2.25 of util-linux is required for
7034 operation, otherwise dead-locks on device nodes may
7035 occur. Again: you need to update util-linux to at least
7036 v2.25 when updating systemd to v217.
7038 * The "multi-seat-x" tool has been removed from systemd, as
7039 its functionality has been integrated into X servers 1.16,
7040 and the tool is hence redundant. It is recommended to update
7041 display managers invoking this tool to simply invoke X
7042 directly from now on, again.
7044 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
7045 message flag has been added for all of systemd's polkit
7046 authenticated method calls has been added. In particular this
7047 now allows optional interactive authorization via polkit for
7048 many of PID1's privileged operations such as unit file
7049 enabling and disabling.
7051 * "udevadm hwdb --update" learnt a new switch "--usr" for
7052 placing the rebuilt hardware database in /usr instead of
7053 /etc. When used only hardware database entries stored in
7054 /usr will be used, and any user database entries in /etc are
7055 ignored. This functionality is useful for vendors to ship a
7056 pre-built database on systems where local configuration is
7057 unnecessary or unlikely.
7059 * Calendar time specifications in .timer units now also
7060 understand the strings "semi-annually", "quarterly" and
7061 "minutely" as shortcuts (in addition to the preexisting
7062 "annually", "hourly", ...).
7064 * systemd-tmpfiles will now correctly create files in /dev
7065 at boot which are marked for creation only at boot. It is
7066 recommended to always create static device nodes with 'c!'
7067 and 'b!', so that they are created only at boot and not
7068 overwritten at runtime.
7070 * When the watchdog logic is used for a service (WatchdogSec=)
7071 and the watchdog timeout is hit the service will now be
7072 terminated with SIGABRT (instead of just SIGTERM), in order
7073 to make sure a proper coredump and backtrace is
7074 generated. This ensures that hanging services will result in
7075 similar coredump/backtrace behaviour as services that hit a
7078 Contributions from: Andreas Henriksson, Andrei Borzenkov,
7079 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
7080 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
7081 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
7082 Herrmann, David Sommerseth, David Strauss, Emil Renner
7083 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
7084 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
7085 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
7086 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
7087 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
7088 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
7089 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
7090 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
7091 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
7092 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
7093 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
7094 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
7095 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
7096 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
7097 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
7098 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
7101 — Berlin, 2014-10-28
7105 * timedated no longer reads NTP implementation unit names from
7106 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
7107 implementations should add a
7109 Conflicts=systemd-timesyncd.service
7111 to their unit files to take over and replace systemd's NTP
7112 default functionality.
7114 * systemd-sysusers gained a new line type "r" for configuring
7115 which UID/GID ranges to allocate system users/groups
7116 from. Lines of type "u" may now add an additional column
7117 that specifies the home directory for the system user to be
7118 created. Also, systemd-sysusers may now optionally read user
7119 information from STDIN instead of a file. This is useful for
7120 invoking it from RPM preinst scriptlets that need to create
7121 users before the first RPM file is installed since these
7122 files might need to be owned by them. A new
7123 %sysusers_create_inline RPM macro has been introduced to do
7124 just that. systemd-sysusers now updates the shadow files as
7125 well as the user/group databases, which should enhance
7126 compatibility with certain tools like grpck.
7128 * A number of bus APIs of PID 1 now optionally consult polkit to
7129 permit access for otherwise unprivileged clients under certain
7130 conditions. Note that this currently doesn't support
7131 interactive authentication yet, but this is expected to be
7132 added eventually, too.
7134 * /etc/machine-info now has new fields for configuring the
7135 deployment environment of the machine, as well as the
7136 location of the machine. hostnamectl has been updated with
7137 new command to update these fields.
7139 * systemd-timesyncd has been updated to automatically acquire
7140 NTP server information from systemd-networkd, which might
7141 have been discovered via DHCP.
7143 * systemd-resolved now includes a caching DNS stub resolver
7144 and a complete LLMNR name resolution implementation. A new
7145 NSS module "nss-resolve" has been added which can be used
7146 instead of glibc's own "nss-dns" to resolve hostnames via
7147 systemd-resolved. Hostnames, addresses and arbitrary RRs may
7148 be resolved via systemd-resolved D-Bus APIs. In contrast to
7149 the glibc internal resolver systemd-resolved is aware of
7150 multi-homed system, and keeps DNS server and caches separate
7151 and per-interface. Queries are sent simultaneously on all
7152 interfaces that have DNS servers configured, in order to
7153 properly handle VPNs and local LANs which might resolve
7154 separate sets of domain names. systemd-resolved may acquire
7155 DNS server information from systemd-networkd automatically,
7156 which in turn might have discovered them via DHCP. A tool
7157 "systemd-resolve-host" has been added that may be used to
7158 query the DNS logic in resolved. systemd-resolved implements
7159 IDNA and automatically uses IDNA or UTF-8 encoding depending
7160 on whether classic DNS or LLMNR is used as transport. In the
7161 next releases we intend to add a DNSSEC and mDNS/DNS-SD
7162 implementation to systemd-resolved.
7164 * A new NSS module nss-mymachines has been added, that
7165 automatically resolves the names of all local registered
7166 containers to their respective IP addresses.
7168 * A new client tool "networkctl" for systemd-networkd has been
7169 added. It currently is entirely passive and will query
7170 networking configuration from udev, rtnetlink and networkd,
7171 and present it to the user in a very friendly
7172 way. Eventually, we hope to extend it to become a full
7173 control utility for networkd.
7175 * .socket units gained a new DeferAcceptSec= setting that
7176 controls the kernels' TCP_DEFER_ACCEPT sockopt for
7177 TCP. Similarly, support for controlling TCP keep-alive
7178 settings has been added (KeepAliveTimeSec=,
7179 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
7180 turning off Nagle's algorithm on TCP has been added
7183 * logind learned a new session type "web", for use in projects
7184 like Cockpit which register web clients as PAM sessions.
7186 * timer units with at least one OnCalendar= setting will now
7187 be started only after time-sync.target has been
7188 reached. This way they will not elapse before the system
7189 clock has been corrected by a local NTP client or
7190 similar. This is particular useful on RTC-less embedded
7191 machines, that come up with an invalid system clock.
7193 * systemd-nspawn's --network-veth= switch should now result in
7194 stable MAC addresses for both the outer and the inner side
7197 * systemd-nspawn gained a new --volatile= switch for running
7198 container instances with /etc or /var unpopulated.
7200 * The kdbus client code has been updated to use the new Linux
7201 3.17 memfd subsystem instead of the old kdbus-specific one.
7203 * systemd-networkd's DHCP client and server now support
7204 FORCERENEW. There are also new configuration options to
7205 configure the vendor client identifier and broadcast mode
7208 * systemd will no longer inform the kernel about the current
7209 timezone, as this is necessarily incorrect and racy as the
7210 kernel has no understanding of DST and similar
7211 concepts. This hence means FAT timestamps will be always
7212 considered UTC, similar to what Android is already
7213 doing. Also, when the RTC is configured to the local time
7214 (rather than UTC) systemd will never synchronize back to it,
7215 as this might confuse Windows at a later boot.
7217 * systemd-analyze gained a new command "verify" for offline
7218 validation of unit files.
7220 * systemd-networkd gained support for a couple of additional
7221 settings for bonding networking setups. Also, the metric for
7222 statically configured routes may now be configured. For
7223 network interfaces where this is appropriate the peer IP
7224 address may now be configured.
7226 * systemd-networkd's DHCP client will no longer request
7227 broadcasting by default, as this tripped up some networks.
7228 For hardware where broadcast is required the feature should
7229 be switched back on using RequestBroadcast=yes.
7231 * systemd-networkd will now set up IPv4LL addresses (when
7232 enabled) even if DHCP is configured successfully.
7234 * udev will now default to respect network device names given
7235 by the kernel when the kernel indicates that these are
7236 predictable. This behavior can be tweaked by changing
7237 NamePolicy= in the relevant .link file.
7239 * A new library systemd-terminal has been added that
7240 implements full TTY stream parsing and rendering. This
7241 library is supposed to be used later on for implementing a
7242 full userspace VT subsystem, replacing the current kernel
7245 * A new tool systemd-journal-upload has been added to push
7246 journal data to a remote system running
7247 systemd-journal-remote.
7249 * journald will no longer forward all local data to another
7250 running syslog daemon. This change has been made because
7251 rsyslog (which appears to be the most commonly used syslog
7252 implementation these days) no longer makes use of this, and
7253 instead pulls the data out of the journal on its own. Since
7254 forwarding the messages to a non-existent syslog server is
7255 more expensive than we assumed we have now turned this
7256 off. If you run a syslog server that is not a recent rsyslog
7257 version, you have to turn this option on again
7258 (ForwardToSyslog= in journald.conf).
7260 * journald now optionally supports the LZ4 compressor for
7261 larger journal fields. This compressor should perform much
7262 better than XZ which was the previous default.
7264 * machinectl now shows the IP addresses of local containers,
7265 if it knows them, plus the interface name of the container.
7267 * A new tool "systemd-escape" has been added that makes it
7268 easy to escape strings to build unit names and similar.
7270 * sd_notify() messages may now include a new ERRNO= field
7271 which is parsed and collected by systemd and shown among the
7272 "systemctl status" output for a service.
7274 * A new component "systemd-firstboot" has been added that
7275 queries the most basic systemd information (timezone,
7276 hostname, root password) interactively on first
7277 boot. Alternatively it may also be used to provision these
7278 things offline on OS images installed into directories.
7280 * The default sysctl.d/ snippets will now set
7282 net.ipv4.conf.default.promote_secondaries=1
7284 This has the benefit of no flushing secondary IP addresses
7285 when primary addresses are removed.
7287 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
7288 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
7289 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
7290 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
7291 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
7292 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
7293 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
7294 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
7295 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
7296 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
7297 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
7298 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
7299 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
7300 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
7301 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
7303 — Berlin, 2014-08-19
7307 * A new tool systemd-sysusers has been added. This tool
7308 creates system users and groups in /etc/passwd and
7309 /etc/group, based on static declarative system user/group
7310 definitions in /usr/lib/sysusers.d/. This is useful to
7311 enable factory resets and volatile systems that boot up with
7312 an empty /etc directory, and thus need system users and
7313 groups created during early boot. systemd now also ships
7314 with two default sysusers.d/ files for the most basic
7315 users and groups systemd and the core operating system
7318 * A new tmpfiles snippet has been added that rebuilds the
7319 essential files in /etc on boot, should they be missing.
7321 * A directive for ensuring automatic clean-up of
7322 /var/cache/man/ has been removed from the default
7323 configuration. This line should now be shipped by the man
7324 implementation. The necessary change has been made to the
7325 man-db implementation. Note that you need to update your man
7326 implementation to one that ships this line, otherwise no
7327 automatic clean-up of /var/cache/man will take place.
7329 * A new condition ConditionNeedsUpdate= has been added that
7330 may conditionalize services to only run when /etc or /var
7331 are "older" than the vendor operating system resources in
7332 /usr. This is useful for reconstructing or updating /etc
7333 after an offline update of /usr or a factory reset, on the
7334 next reboot. Services that want to run once after such an
7335 update or reset should use this condition and order
7336 themselves before the new systemd-update-done.service, which
7337 will mark the two directories as fully updated. A number of
7338 service files have been added making use of this, to rebuild
7339 the udev hardware database, the journald message catalog and
7340 dynamic loader cache (ldconfig). The systemd-sysusers tool
7341 described above also makes use of this now. With this in
7342 place it is now possible to start up a minimal operating
7343 system with /etc empty cleanly. For more information on the
7344 concepts involved see this recent blog story:
7346 http://0pointer.de/blog/projects/stateless.html
7348 * A new system group "input" has been introduced, and all
7349 input device nodes get this group assigned. This is useful
7350 for system-level software to get access to input devices. It
7351 complements what is already done for "audio" and "video".
7353 * systemd-networkd learnt minimal DHCPv4 server support in
7354 addition to the existing DHCPv4 client support. It also
7355 learnt DHCPv6 client and IPv6 Router Solicitation client
7356 support. The DHCPv4 client gained support for static routes
7357 passed in from the server. Note that the [DHCPv4] section
7358 known in older systemd-networkd versions has been renamed to
7359 [DHCP] and is now also used by the DHCPv6 client. Existing
7360 .network files using settings of this section should be
7361 updated, though compatibility is maintained. Optionally, the
7362 client hostname may now be sent to the DHCP server.
7364 * networkd gained support for vxlan virtual networks as well
7365 as tun/tap and dummy devices.
7367 * networkd gained support for automatic allocation of address
7368 ranges for interfaces from a system-wide pool of
7369 addresses. This is useful for dynamically managing a large
7370 number of interfaces with a single network configuration
7371 file. In particular this is useful to easily assign
7372 appropriate IP addresses to the veth links of a large number
7373 of nspawn instances.
7375 * RPM macros for processing sysusers, sysctl and binfmt
7376 drop-in snippets at package installation time have been
7379 * The /etc/os-release file should now be placed in
7380 /usr/lib/os-release. The old location is automatically
7381 created as symlink. /usr/lib is the more appropriate
7382 location of this file, since it shall actually describe the
7383 vendor operating system shipped in /usr, and not the
7384 configuration stored in /etc.
7386 * .mount units gained a new boolean SloppyOptions= setting
7387 that maps to mount(8)'s -s option which enables permissive
7388 parsing of unknown mount options.
7390 * tmpfiles learnt a new "L+" directive which creates a symlink
7391 but (unlike "L") deletes a pre-existing file first, should
7392 it already exist and not already be the correct
7393 symlink. Similarly, "b+", "c+" and "p+" directives have been
7394 added as well, which create block and character devices, as
7395 well as fifos in the filesystem, possibly removing any
7396 pre-existing files of different types.
7398 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
7399 'argument' field (which so far specified the source to
7400 symlink/copy the files from) is now optional. If omitted the
7401 same file os copied from /usr/share/factory/ suffixed by the
7402 full destination path. This is useful for populating /etc
7403 with essential files, by copying them from vendor defaults
7404 shipped in /usr/share/factory/etc.
7406 * A new command "systemctl preset-all" has been added that
7407 applies the service preset settings to all installed unit
7408 files. A new switch --preset-mode= has been added that
7409 controls whether only enable or only disable operations
7412 * A new command "systemctl is-system-running" has been added
7413 that allows checking the overall state of the system, for
7414 example whether it is fully up and running.
7416 * When the system boots up with an empty /etc, the equivalent
7417 to "systemctl preset-all" is executed during early boot, to
7418 make sure all default services are enabled after a factory
7421 * systemd now contains a minimal preset file that enables the
7422 most basic services systemd ships by default.
7424 * Unit files' [Install] section gained a new DefaultInstance=
7425 field for defining the default instance to create if a
7426 template unit is enabled with no instance specified.
7428 * A new passive target cryptsetup-pre.target has been added
7429 that may be used by services that need to make they run and
7430 finish before the first LUKS cryptographic device is set up.
7432 * The /dev/loop-control and /dev/btrfs-control device nodes
7433 are now owned by the "disk" group by default, opening up
7434 access to this group.
7436 * systemd-coredump will now automatically generate a
7437 stack trace of all core dumps taking place on the system,
7438 based on elfutils' libdw library. This stack trace is logged
7441 * systemd-coredump may now optionally store coredumps directly
7442 on disk (in /var/lib/systemd/coredump, possibly compressed),
7443 instead of storing them unconditionally in the journal. This
7444 mode is the new default. A new configuration file
7445 /etc/systemd/coredump.conf has been added to configure this
7446 and other parameters of systemd-coredump.
7448 * coredumpctl gained a new "info" verb to show details about a
7449 specific coredump. A new switch "-1" has also been added
7450 that makes sure to only show information about the most
7451 recent entry instead of all entries. Also, as the tool is
7452 generally useful now the "systemd-" prefix of the binary
7453 name has been removed. Distributions that want to maintain
7454 compatibility with the old name should add a symlink from
7455 the old name to the new name.
7457 * journald's SplitMode= now defaults to "uid". This makes sure
7458 that unprivileged users can access their own coredumps with
7459 coredumpctl without restrictions.
7461 * New kernel command line options "systemd.wants=" (for
7462 pulling an additional unit during boot), "systemd.mask="
7463 (for masking a specific unit for the boot), and
7464 "systemd.debug-shell" (for enabling the debug shell on tty9)
7465 have been added. This is implemented in the new generator
7466 "systemd-debug-generator".
7468 * systemd-nspawn will now by default filter a couple of
7469 syscalls for containers, among them those required for
7470 kernel module loading, direct x86 IO port access, swap
7471 management, and kexec. Most importantly though
7472 open_by_handle_at() is now prohibited for containers,
7473 closing a hole similar to a recently discussed vulnerability
7474 in docker regarding access to files on file hierarchies the
7475 container should normally not have access to. Note that, for
7476 nspawn, we generally make no security claims anyway (and
7477 this is explicitly documented in the man page), so this is
7478 just a fix for one of the most obvious problems.
7480 * A new man page file-hierarchy(7) has been added that
7481 contains a minimized, modernized version of the file system
7482 layout systemd expects, similar in style to the FHS
7483 specification or hier(5). A new tool systemd-path(1) has
7484 been added to query many of these paths for the local
7487 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
7488 longer done. Since the directory now has a per-user size
7489 limit, and is cleaned on logout this appears unnecessary,
7490 in particular since this now brings the lifecycle of this
7491 directory closer in line with how IPC objects are handled.
7493 * systemd.pc now exports a number of additional directories,
7494 including $libdir (which is useful to identify the library
7495 path for the primary architecture of the system), and a
7496 couple of drop-in directories.
7498 * udev's predictable network interface names now use the dev_port
7499 sysfs attribute, introduced in linux 3.15 instead of dev_id to
7500 distinguish between ports of the same PCI function. dev_id should
7501 only be used for ports using the same HW address, hence the need
7504 * machined has been updated to export the OS version of a
7505 container (read from /etc/os-release and
7506 /usr/lib/os-release) on the bus. This is now shown in
7507 "machinectl status" for a machine.
7509 * A new service setting RestartForceExitStatus= has been
7510 added. If configured to a set of exit signals or process
7511 return values, the service will be restarted when the main
7512 daemon process exits with any of them, regardless of the
7515 * systemctl's -H switch for connecting to remote systemd
7516 machines has been extended so that it may be used to
7517 directly connect to a specific container on the
7518 host. "systemctl -H root@foobar:waldi" will now connect as
7519 user "root" to host "foobar", and then proceed directly to
7520 the container named "waldi". Note that currently you have to
7521 authenticate as user "root" for this to work, as entering
7522 containers is a privileged operation.
7524 Contributions from: Andreas Henriksson, Benjamin Steinwender,
7525 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
7526 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
7527 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
7528 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
7529 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
7530 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
7531 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
7532 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
7533 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
7534 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
7535 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
7537 — Berlin, 2014-07-03
7541 * As an experimental feature, udev now tries to lock the
7542 disk device node (flock(LOCK_SH|LOCK_NB)) while it
7543 executes events for the disk or any of its partitions.
7544 Applications like partitioning programs can lock the
7545 disk device node (flock(LOCK_EX)) and claim temporary
7546 device ownership that way; udev will entirely skip all event
7547 handling for this disk and its partitions. If the disk
7548 was opened for writing, the close will trigger a partition
7549 table rescan in udev's "watch" facility, and if needed
7550 synthesize "change" events for the disk and all its partitions.
7551 This is now unconditionally enabled, and if it turns out to
7552 cause major problems, we might turn it on only for specific
7553 devices, or might need to disable it entirely. Device Mapper
7554 devices are excluded from this logic.
7556 * We temporarily dropped the "-l" switch for fsck invocations,
7557 since they collide with the flock() logic above. util-linux
7558 upstream has been changed already to avoid this conflict,
7559 and we will re-add "-l" as soon as util-linux with this
7560 change has been released.
7562 * The dependency on libattr has been removed. Since a long
7563 time, the extended attribute calls have moved to glibc, and
7564 libattr is thus unnecessary.
7566 * Virtualization detection works without privileges now. This
7567 means the systemd-detect-virt binary no longer requires
7568 CAP_SYS_PTRACE file capabilities, and our daemons can run
7569 with fewer privileges.
7571 * systemd-networkd now runs under its own "systemd-network"
7572 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
7573 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
7574 loses the ability to write to files owned by root this way.
7576 * Similarly, systemd-resolved now runs under its own
7577 "systemd-resolve" user with no capabilities remaining.
7579 * Similarly, systemd-bus-proxyd now runs under its own
7580 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
7582 * systemd-networkd gained support for setting up "veth"
7583 virtual Ethernet devices for container connectivity, as well
7584 as GRE and VTI tunnels.
7586 * systemd-networkd will no longer automatically attempt to
7587 manually load kernel modules necessary for certain tunnel
7588 transports. Instead, it is assumed the kernel loads them
7589 automatically when required. This only works correctly on
7590 very new kernels. On older kernels, please consider adding
7591 the kernel modules to /etc/modules-load.d/ as a work-around.
7593 * The resolv.conf file systemd-resolved generates has been
7594 moved to /run/systemd/resolve/. If you have a symlink from
7595 /etc/resolv.conf, it might be necessary to correct it.
7597 * Two new service settings, ProtectHome= and ProtectSystem=,
7598 have been added. When enabled, they will make the user data
7599 (such as /home) inaccessible or read-only and the system
7600 (such as /usr) read-only, for specific services. This allows
7601 very light-weight per-service sandboxing to avoid
7602 modifications of user data or system files from
7603 services. These two new switches have been enabled for all
7604 of systemd's long-running services, where appropriate.
7606 * Socket units gained new SocketUser= and SocketGroup=
7607 settings to set the owner user and group of AF_UNIX sockets
7608 and FIFOs in the file system.
7610 * Socket units gained a new RemoveOnStop= setting. If enabled,
7611 all FIFOS and sockets in the file system will be removed
7612 when the specific socket unit is stopped.
7614 * Socket units gained a new Symlinks= setting. It takes a list
7615 of symlinks to create to file system sockets or FIFOs
7616 created by the specific Unix sockets. This is useful to
7617 manage symlinks to socket nodes with the same lifecycle as
7620 * The /dev/log socket and /dev/initctl FIFO have been moved to
7621 /run, and have been replaced by symlinks. This allows
7622 connecting to these facilities even if PrivateDevices=yes is
7623 used for a service (which makes /dev/log itself unavailable,
7624 but /run is left). This also has the benefit of ensuring
7625 that /dev only contains device nodes, directories and
7626 symlinks, and nothing else.
7628 * sd-daemon gained two new calls sd_pid_notify() and
7629 sd_pid_notifyf(). They are similar to sd_notify() and
7630 sd_notifyf(), but allow overriding of the source PID of
7631 notification messages if permissions permit this. This is
7632 useful to send notify messages on behalf of a different
7633 process (for example, the parent process). The
7634 systemd-notify tool has been updated to make use of this
7635 when sending messages (so that notification messages now
7636 originate from the shell script invoking systemd-notify and
7637 not the systemd-notify process itself. This should minimize
7638 a race where systemd fails to associate notification
7639 messages to services when the originating process already
7642 * A new "on-abnormal" setting for Restart= has been added. If
7643 set, it will result in automatic restarts on all "abnormal"
7644 reasons for a process to exit, which includes unclean
7645 signals, core dumps, timeouts and watchdog timeouts, but
7646 does not include clean and unclean exit codes or clean
7647 signals. Restart=on-abnormal is an alternative for
7648 Restart=on-failure for services that shall be able to
7649 terminate and avoid restarts on certain errors, by
7650 indicating so with an unclean exit code. Restart=on-failure
7651 or Restart=on-abnormal is now the recommended setting for
7652 all long-running services.
7654 * If the InaccessibleDirectories= service setting points to a
7655 mount point (or if there are any submounts contained within
7656 it), it is now attempted to completely unmount it, to make
7657 the file systems truly unavailable for the respective
7660 * The ReadOnlyDirectories= service setting and
7661 systemd-nspawn's --read-only parameter are now recursively
7662 applied to all submounts, too.
7664 * Mount units may now be created transiently via the bus APIs.
7666 * The support for SysV and LSB init scripts has been removed
7667 from the systemd daemon itself. Instead, it is now
7668 implemented as a generator that creates native systemd units
7669 from these scripts when needed. This enables us to remove a
7670 substantial amount of legacy code from PID 1, following the
7671 fact that many distributions only ship a very small number
7672 of LSB/SysV init scripts nowadays.
7674 * Privileged Xen (dom0) domains are not considered
7675 virtualization anymore by the virtualization detection
7676 logic. After all, they generally have unrestricted access to
7677 the hardware and usually are used to manage the unprivileged
7680 * systemd-tmpfiles gained a new "C" line type, for copying
7681 files or entire directories.
7683 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
7684 lines. So far, they have been non-globbing versions of the
7685 latter, and have thus been redundant. In future, it is
7686 recommended to only use "z". "m" has hence been removed
7687 from the documentation, even though it stays supported.
7689 * A tmpfiles snippet to recreate the most basic structure in
7690 /var has been added. This is enough to create the /var/run →
7691 /run symlink and create a couple of structural
7692 directories. This allows systems to boot up with an empty or
7693 volatile /var. Of course, while with this change, the core OS
7694 now is capable with dealing with a volatile /var, not all
7695 user services are ready for it. However, we hope that sooner
7696 or later, many service daemons will be changed upstream so
7697 that they are able to automatically create their necessary
7698 directories in /var at boot, should they be missing. This is
7699 the first step to allow state-less systems that only require
7700 the vendor image for /usr to boot.
7702 * systemd-nspawn has gained a new --tmpfs= switch to mount an
7703 empty tmpfs instance to a specific directory. This is
7704 particularly useful for making use of the automatic
7705 reconstruction of /var (see above), by passing --tmpfs=/var.
7707 * Access modes specified in tmpfiles snippets may now be
7708 prefixed with "~", which indicates that they shall be masked
7709 by whether the existing file or directory is currently
7710 writable, readable or executable at all. Also, if specified,
7711 the sgid/suid/sticky bits will be masked for all
7714 * A new passive target unit "network-pre.target" has been
7715 added which is useful for services that shall run before any
7716 network is configured, for example firewall scripts.
7718 * The "floppy" group that previously owned the /dev/fd*
7719 devices is no longer used. The "disk" group is now used
7720 instead. Distributions should probably deprecate usage of
7723 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
7724 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
7725 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
7726 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
7727 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
7728 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
7731 — Berlin, 2014-06-11
7735 * A new "systemd-timesyncd" daemon has been added for
7736 synchronizing the system clock across the network. It
7737 implements an SNTP client. In contrast to NTP
7738 implementations such as chrony or the NTP reference server,
7739 this only implements a client side, and does not bother with
7740 the full NTP complexity, focusing only on querying time from
7741 one remote server and synchronizing the local clock to
7742 it. Unless you intend to serve NTP to networked clients or
7743 want to connect to local hardware clocks, this simple NTP
7744 client should be more than appropriate for most
7745 installations. The daemon runs with minimal privileges, and
7746 has been hooked up with networkd to only operate when
7747 network connectivity is available. The daemon saves the
7748 current clock to disk every time a new NTP sync has been
7749 acquired, and uses this to possibly correct the system clock
7750 early at bootup, in order to accommodate for systems that
7751 lack an RTC such as the Raspberry Pi and embedded devices,
7752 and to make sure that time monotonically progresses on these
7753 systems, even if it is not always correct. To make use of
7754 this daemon, a new system user and group "systemd-timesync"
7755 needs to be created on installation of systemd.
7757 * The queue "seqnum" interface of libudev has been disabled, as
7758 it was generally incompatible with device namespacing as
7759 sequence numbers of devices go "missing" if the devices are
7760 part of a different namespace.
7762 * "systemctl list-timers" and "systemctl list-sockets" gained
7763 a --recursive switch for showing units of these types also
7764 for all local containers, similar in style to the already
7765 supported --recursive switch for "systemctl list-units".
7767 * A new RebootArgument= setting has been added for service
7768 units, which may be used to specify a kernel reboot argument
7769 to use when triggering reboots with StartLimitAction=.
7771 * A new FailureAction= setting has been added for service
7772 units which may be used to specify an operation to trigger
7773 when a service fails. This works similarly to
7774 StartLimitAction=, but unlike it, controls what is done
7775 immediately rather than only after several attempts to
7776 restart the service in question.
7778 * hostnamed got updated to also expose the kernel name,
7779 release, and version on the bus. This is useful for
7780 executing commands like hostnamectl with the -H switch.
7781 systemd-analyze makes use of this to properly display
7782 details when running non-locally.
7784 * The bootchart tool can now show cgroup information in the
7785 graphs it generates.
7787 * The CFS CPU quota cgroup attribute is now exposed for
7788 services. The new CPUQuota= switch has been added for this
7789 which takes a percentage value. Setting this will have the
7790 result that a service may never get more CPU time than the
7791 specified percentage, even if the machine is otherwise idle.
7793 * systemd-networkd learned IPIP and SIT tunnel support.
7795 * LSB init scripts exposing a dependency on $network will now
7796 get a dependency on network-online.target rather than simply
7797 network.target. This should bring LSB handling closer to
7798 what it was on SysV systems.
7800 * A new fsck.repair= kernel option has been added to control
7801 how fsck shall deal with unclean file systems at boot.
7803 * The (.ini) configuration file parser will now silently ignore
7804 sections whose names begin with "X-". This may be used to maintain
7805 application-specific extension sections in unit files.
7807 * machined gained a new API to query the IP addresses of
7808 registered containers. "machinectl status" has been updated
7809 to show these addresses in its output.
7811 * A new call sd_uid_get_display() has been added to the
7812 sd-login APIs for querying the "primary" session of a
7813 user. The "primary" session of the user is elected from the
7814 user's sessions and generally a graphical session is
7815 preferred over a text one.
7817 * A minimal systemd-resolved daemon has been added. It
7818 currently simply acts as a companion to systemd-networkd and
7819 manages resolv.conf based on per-interface DNS
7820 configuration, possibly supplied via DHCP. In the long run
7821 we hope to extend this into a local DNSSEC enabled DNS and
7824 * The systemd-networkd-wait-online tool is now enabled by
7825 default. It will delay network-online.target until a network
7826 connection has been configured. The tool primarily integrates
7827 with networkd, but will also make a best effort to make sense
7828 of network configuration performed in some other way.
7830 * Two new service options StartupCPUShares= and
7831 StartupBlockIOWeight= have been added that work similarly to
7832 CPUShares= and BlockIOWeight= however only apply during
7833 system startup. This is useful to prioritize certain services
7834 differently during bootup than during normal runtime.
7836 * hostnamed has been changed to prefer the statically
7837 configured hostname in /etc/hostname (unless set to
7838 'localhost' or empty) over any dynamic one supplied by
7839 dhcp. With this change, the rules for picking the hostname
7840 match more closely the rules of other configuration settings
7841 where the local administrator's configuration in /etc always
7842 overrides any other settings.
7844 Contributions from: Ali H. Caliskan, Alison Chaiken, Bas van
7845 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
7846 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
7847 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
7848 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
7849 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
7850 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
7851 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
7852 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
7853 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
7854 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
7855 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
7856 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
7857 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
7858 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
7859 Lindskog, WaLyong Cho, Will Woods, Zbigniew
7862 — Beijing, 2014-05-28
7866 * When restoring the screen brightness at boot, stay away from
7867 the darkest setting or from the lowest 5% of the available
7868 range, depending on which is the larger value of both. This
7869 should effectively protect the user from rebooting into a
7870 black screen, should the brightness have been set to minimum
7873 * sd-login gained a new sd_machine_get_class() call to
7874 determine the class ("vm" or "container") of a machine
7875 registered with machined.
7877 * sd-login gained new calls
7878 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
7879 to query the identity of the peer of a local AF_UNIX
7880 connection. They operate similarly to their sd_pid_get_xyz()
7883 * PID 1 will now maintain a system-wide system state engine
7884 with the states "starting", "running", "degraded",
7885 "maintenance", "stopping". These states are bound to system
7886 startup, normal runtime, runtime with at least one failed
7887 service, rescue/emergency mode and system shutdown. This
7888 state is shown in the "systemctl status" output when no unit
7889 name is passed. It is useful to determine system state, in
7890 particularly when doing so for many systems or containers at
7893 * A new command "list-machines" has been added to "systemctl"
7894 that lists all local OS containers and shows their system
7895 state (see above), if systemd runs inside of them.
7897 * systemctl gained a new "-r" switch to recursively enumerate
7898 units on all local containers, when used with the
7899 "list-unit" command (which is the default one that is
7900 executed when no parameters are specified).
7902 * The GPT automatic partition discovery logic will now honour
7903 two GPT partition flags: one may be set on a partition to
7904 cause it to be mounted read-only, and the other may be set
7905 on a partition to ignore it during automatic discovery.
7907 * Two new GPT type UUIDs have been added for automatic root
7908 partition discovery, for 32-bit and 64-bit ARM. This is not
7909 particularly useful for discovering the root directory on
7910 these architectures during bare-metal boots (since UEFI is
7911 not common there), but still very useful to allow booting of
7912 ARM disk images in nspawn with the -i option.
7914 * MAC addresses of interfaces created with nspawn's
7915 --network-interface= switch will now be generated from the
7916 machine name, and thus be stable between multiple invocations
7919 * logind will now automatically remove all IPC objects owned
7920 by a user if she or he fully logs out. This makes sure that
7921 users who are logged out cannot continue to consume IPC
7922 resources. This covers SysV memory, semaphores and message
7923 queues as well as POSIX shared memory and message
7924 queues. Traditionally, SysV and POSIX IPC had no lifecycle
7925 limits. With this functionality, that is corrected. This may
7926 be turned off by using the RemoveIPC= switch of logind.conf.
7928 * The systemd-machine-id-setup and tmpfiles tools gained a
7929 --root= switch to operate on a specific root directory,
7932 * journald can now forward logged messages to the TTYs of all
7933 logged in users ("wall"). This is the default for all
7934 emergency messages now.
7936 * A new tool systemd-journal-remote has been added to stream
7937 journal log messages across the network.
7939 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
7940 controller trees are mounted into it. Note that the
7941 directories mounted beneath it are not read-only. This is a
7942 security measure and is particularly useful because glibc
7943 actually includes a search logic to pick any tmpfs it can
7944 find to implement shm_open() if /dev/shm is not available
7945 (which it might very well be in namespaced setups).
7947 * machinectl gained a new "poweroff" command to cleanly power
7948 down a local OS container.
7950 * The PrivateDevices= unit file setting will now also drop the
7951 CAP_MKNOD capability from the capability bound set, and
7952 imply DevicePolicy=closed.
7954 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
7955 comprehensively on all long-running systemd services where
7956 this is appropriate.
7958 * systemd-udevd will now run in a disassociated mount
7959 namespace. To mount directories from udev rules, make sure to
7960 pull in mount units via SYSTEMD_WANTS properties.
7962 * The kdbus support gained support for uploading policy into
7963 the kernel. sd-bus gained support for creating "monitoring"
7964 connections that can eavesdrop into all bus communication
7965 for debugging purposes.
7967 * Timestamps may now be specified in seconds since the UNIX
7968 epoch Jan 1st, 1970 by specifying "@" followed by the value
7971 * Native tcpwrap support in systemd has been removed. tcpwrap
7972 is old code, not really maintained anymore and has serious
7973 shortcomings, and better options such as firewalls
7974 exist. For setups that require tcpwrap usage, please
7975 consider invoking your socket-activated service via tcpd,
7976 like on traditional inetd.
7978 * A new system.conf configuration option
7979 DefaultTimerAccuracySec= has been added that controls the
7980 default AccuracySec= setting of .timer units.
7982 * Timer units gained a new WakeSystem= switch. If enabled,
7983 timers configured this way will cause the system to resume
7984 from system suspend (if the system supports that, which most
7987 * Timer units gained a new Persistent= switch. If enabled,
7988 timers configured this way will save to disk when they have
7989 been last triggered. This information is then used on next
7990 reboot to possible execute overdue timer events, that
7991 could not take place because the system was powered off.
7992 This enables simple anacron-like behaviour for timer units.
7994 * systemctl's "list-timers" will now also list the time a
7995 timer unit was last triggered in addition to the next time
7996 it will be triggered.
7998 * systemd-networkd will now assign predictable IPv4LL
7999 addresses to its local interfaces.
8001 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
8002 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
8003 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
8004 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
8005 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
8006 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
8007 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
8008 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
8011 — Berlin, 2014-03-25
8015 * A new unit file setting RestrictAddressFamilies= has been
8016 added to restrict which socket address families unit
8017 processes gain access to. This takes address family names
8018 like "AF_INET" or "AF_UNIX", and is useful to minimize the
8019 attack surface of services via exotic protocol stacks. This
8020 is built on seccomp system call filters.
8022 * Two new unit file settings RuntimeDirectory= and
8023 RuntimeDirectoryMode= have been added that may be used to
8024 manage a per-daemon runtime directories below /run. This is
8025 an alternative for setting up directory permissions with
8026 tmpfiles snippets, and has the advantage that the runtime
8027 directory's lifetime is bound to the daemon runtime and that
8028 the daemon starts up with an empty directory each time. This
8029 is particularly useful when writing services that drop
8030 privileges using the User= or Group= setting.
8032 * The DeviceAllow= unit setting now supports globbing for
8033 matching against device group names.
8035 * The systemd configuration file system.conf gained new
8036 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
8037 DefaultMemoryAccounting= to globally turn on/off accounting
8038 for specific resources (cgroups) for all units. These
8039 settings may still be overridden individually in each unit
8042 * systemd-gpt-auto-generator is now able to discover /srv and
8043 root partitions in addition to /home and swap partitions. It
8044 also supports LUKS-encrypted partitions now. With this in
8045 place, automatic discovery of partitions to mount following
8046 the Discoverable Partitions Specification
8047 (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
8048 is now a lot more complete. This allows booting without
8049 /etc/fstab and without root= on the kernel command line on
8050 systems prepared appropriately.
8052 * systemd-nspawn gained a new --image= switch which allows
8053 booting up disk images and Linux installations on any block
8054 device that follow the Discoverable Partitions Specification
8055 (see above). This means that installations made with
8056 appropriately updated installers may now be started and
8057 deployed using container managers, completely
8058 unmodified. (We hope that libvirt-lxc will add support for
8059 this feature soon, too.)
8061 * systemd-nspawn gained a new --network-macvlan= setting to
8062 set up a private macvlan interface for the
8063 container. Similarly, systemd-networkd gained a new
8064 Kind=macvlan setting in .netdev files.
8066 * systemd-networkd now supports configuring local addresses
8069 * A new tool systemd-network-wait-online has been added to
8070 synchronously wait for network connectivity using
8073 * The sd-bus.h bus API gained a new sd_bus_track object for
8074 tracking the lifecycle of bus peers. Note that sd-bus.h is
8075 still not a public API though (unless you specify
8076 --enable-kdbus on the configure command line, which however
8077 voids your warranty and you get no API stability guarantee).
8079 * The $XDG_RUNTIME_DIR runtime directories for each user are
8080 now individual tmpfs instances, which has the benefit of
8081 introducing separate pools for each user, with individual
8082 size limits, and thus making sure that unprivileged clients
8083 can no longer negatively impact the system or other users by
8084 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
8085 RuntimeDirectorySize= has been introduced that allows
8086 controlling the default size limit for all users. It
8087 defaults to 10% of the available physical memory. This is no
8088 replacement for quotas on tmpfs though (which the kernel
8089 still does not support), as /dev/shm and /tmp are still
8090 shared resources used by both the system and unprivileged
8093 * logind will now automatically turn off automatic suspending
8094 on laptop lid close when more than one display is
8095 connected. This was previously expected to be implemented
8096 individually in desktop environments (such as GNOME),
8097 however has been added to logind now, in order to fix a
8098 boot-time race where a desktop environment might not have
8099 been started yet and thus not been able to take an inhibitor
8100 lock at the time where logind already suspends the system
8101 due to a closed lid.
8103 * logind will now wait at least 30s after each system
8104 suspend/resume cycle, and 3min after system boot before
8105 suspending the system due to a closed laptop lid. This
8106 should give USB docking stations and similar enough time to
8107 be probed and configured after system resume and boot in
8108 order to then act as suspend blocker.
8110 * systemd-run gained a new --property= setting which allows
8111 initialization of resource control properties (and others)
8112 for the created scope or service unit. Example: "systemd-run
8113 --property=BlockIOWeight=10 updatedb" may be used to run
8114 updatedb at a low block IO scheduling weight.
8116 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
8117 now also work in --scope mode.
8119 * When systemd is compiled with kdbus support, basic support
8120 for enforced policies is now in place. (Note that enabling
8121 kdbus still voids your warranty and no API compatibility
8124 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
8125 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
8126 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
8127 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
8128 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
8129 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
8130 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
8131 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
8132 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
8133 Zbigniew Jędrzejewski-Szmek
8135 — Berlin, 2014-03-12
8139 * systemd will now relabel /dev after loading the SMACK policy
8140 according to SMACK rules.
8142 * A new unit file option AppArmorProfile= has been added to
8143 set the AppArmor profile for the processes of a unit.
8145 * A new condition check ConditionArchitecture= has been added
8146 to conditionalize units based on the system architecture, as
8147 reported by uname()'s "machine" field.
8149 * systemd-networkd now supports matching on the system
8150 virtualization, architecture, kernel command line, hostname
8153 * logind is now a lot more aggressive when suspending the
8154 machine due to a closed laptop lid. Instead of acting only
8155 on the lid close action, it will continuously watch the lid
8156 status and act on it. This is useful for laptops where the
8157 power button is on the outside of the chassis so that it can
8158 be reached without opening the lid (such as the Lenovo
8159 Yoga). On those machines, logind will now immediately
8160 re-suspend the machine if the power button has been
8161 accidentally pressed while the laptop was suspended and in a
8162 backpack or similar.
8164 * logind will now watch SW_DOCK switches and inhibit reaction
8165 to the lid switch if it is pressed. This means that logind
8166 will not suspend the machine anymore if the lid is closed
8167 and the system is docked, if the laptop supports SW_DOCK
8168 notifications via the input layer. Note that ACPI docking
8169 stations do not generate this currently. Also note that this
8170 logic is usually not fully sufficient and Desktop
8171 Environments should take a lid switch inhibitor lock when an
8172 external display is connected, as systemd will not watch
8175 * nspawn will now make use of the devices cgroup controller by
8176 default, and only permit creation of and access to the usual
8177 API device nodes like /dev/null or /dev/random, as well as
8178 access to (but not creation of) the pty devices.
8180 * We will now ship a default .network file for
8181 systemd-networkd that automatically configures DHCP for
8182 network interfaces created by nspawn's --network-veth or
8183 --network-bridge= switches.
8185 * systemd will now understand the usual M, K, G, T suffixes
8186 according to SI conventions (i.e. to the base 1000) when
8187 referring to throughput and hardware metrics. It will stay
8188 with IEC conventions (i.e. to the base 1024) for software
8189 metrics, according to what is customary according to
8190 Wikipedia. We explicitly document which base applies for
8191 each configuration option.
8193 * The DeviceAllow= setting in unit files now supports a syntax to
8194 allow-list an entire group of devices node majors at once, based on
8195 the /proc/devices listing. For example, with the string "char-pts",
8196 it is now possible to allow-list all current and future pseudo-TTYs
8199 * sd-event learned a new "post" event source. Event sources of
8200 this type are triggered by the dispatching of any event
8201 source of a type that is not "post". This is useful for
8202 implementing clean-up and check event sources that are
8203 triggered by other work being done in the program.
8205 * systemd-networkd is no longer statically enabled, but uses
8206 the usual [Install] sections so that it can be
8207 enabled/disabled using systemctl. It still is enabled by
8210 * When creating a veth interface pair with systemd-nspawn, the
8211 host side will now be prefixed with "vb-" if
8212 --network-bridge= is used, and with "ve-" if --network-veth
8213 is used. This way, it is easy to distinguish these cases on
8214 the host, for example to apply different configuration to
8215 them with systemd-networkd.
8217 * The compatibility libraries for libsystemd-journal.so,
8218 libsystem-id128.so, libsystemd-login.so and
8219 libsystemd-daemon.so do not make use of IFUNC
8220 anymore. Instead, we now build libsystemd.so multiple times
8221 under these alternative names. This means that the footprint
8222 is drastically increased, but given that these are
8223 transitional compatibility libraries, this should not matter
8224 much. This change has been made necessary to support the ARM
8225 platform for these compatibility libraries, as the ARM
8226 toolchain is not really at the same level as the toolchain
8227 for other architectures like x86 and does not support
8228 IFUNC. Please make sure to use --enable-compat-libs only
8229 during a transitional period!
8231 * The .include syntax has been deprecated and is not documented
8232 anymore. Drop-in files in .d directories should be used instead.
8234 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
8235 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
8236 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
8237 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
8238 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
8239 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
8240 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
8241 Zbigniew Jędrzejewski-Szmek
8243 — Berlin, 2014-02-24
8247 * A new component "systemd-networkd" has been added that can
8248 be used to configure local network interfaces statically or
8249 via DHCP. It is capable of bringing up bridges, VLANs, and
8250 bonding. Currently, no hook-ups for interactive network
8251 configuration are provided. Use this for your initrd,
8252 container, embedded, or server setup if you need a simple,
8253 yet powerful, network configuration solution. This
8254 configuration subsystem is quite nifty, as it allows wildcard
8255 hotplug matching in interfaces. For example, with a single
8256 configuration snippet, you can configure that all Ethernet
8257 interfaces showing up are automatically added to a bridge,
8258 or similar. It supports link-sensing and more.
8260 * A new tool "systemd-socket-proxyd" has been added which can
8261 act as a bidirectional proxy for TCP sockets. This is
8262 useful for adding socket activation support to services that
8263 do not actually support socket activation, including virtual
8264 machines and the like.
8266 * Add a new tool to save/restore rfkill state on
8269 * Save/restore state of keyboard backlights in addition to
8270 display backlights on shutdown/boot.
8272 * udev learned a new SECLABEL{} construct to label device
8273 nodes with a specific security label when they appear. For
8274 now, only SECLABEL{selinux} is supported, but the syntax is
8275 prepared for additional security frameworks.
8277 * udev gained a new scheme to configure link-level attributes
8278 from files in /etc/systemd/network/*.link. These files can
8279 match against MAC address, device path, driver name and type,
8280 and will apply attributes like the naming policy, link speed,
8281 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
8282 address assignment policy (randomized, ...).
8284 * The configuration of network interface naming rules for
8285 "permanent interface names" has changed: a new NamePolicy=
8286 setting in the [Link] section of .link files determines the
8287 priority of possible naming schemes (onboard, slot, MAC,
8288 path). The default value of this setting is determined by
8289 /usr/lib/net/links/99-default.link. Old
8290 80-net-name-slot.rules udev configuration file has been
8291 removed, so local configuration overriding this file should
8292 be adapted to override 99-default.link instead.
8294 * When the User= switch is used in a unit file, also
8295 initialize $SHELL= based on the user database entry.
8297 * systemd no longer depends on libdbus. All communication is
8298 now done with sd-bus, systemd's low-level bus library
8301 * kdbus support has been added to PID 1 itself. When kdbus is
8302 enabled, this causes PID 1 to set up the system bus and
8303 enable support for a new ".busname" unit type that
8304 encapsulates bus name activation on kdbus. It works a little
8305 bit like ".socket" units, except for bus names. A new
8306 generator has been added that converts classic dbus1 service
8307 activation files automatically into native systemd .busname
8310 * sd-bus: add a light-weight vtable implementation that allows
8311 defining objects on the bus with a simple static const
8312 vtable array of its methods, signals and properties.
8314 * systemd will not generate or install static dbus
8315 introspection data anymore to /usr/share/dbus-1/interfaces,
8316 as the precise format of these files is unclear, and
8317 nothing makes use of it.
8319 * A proxy daemon is now provided to proxy clients connecting
8320 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
8321 compatibility with classic D-Bus.
8323 * A bus driver implementation has been added that supports the
8324 classic D-Bus bus driver calls on kdbus, also for
8325 compatibility purposes.
8327 * A new API "sd-event.h" has been added that implements a
8328 minimal event loop API built around epoll. It provides a
8329 couple of features that direct epoll usage is lacking:
8330 prioritization of events, scales to large numbers of timer
8331 events, per-event timer slack (accuracy), system-wide
8332 coalescing of timer events, exit handlers, watchdog
8333 supervision support using systemd's sd_notify() API, child
8336 * A new API "sd-rntl.h" has been added that provides an API
8337 around the route netlink interface of the kernel, similar in
8338 style to "sd-bus.h".
8340 * A new API "sd-dhcp-client.h" has been added that provides a
8341 small DHCPv4 client-side implementation. This is used by
8344 * There is a new kernel command line option
8345 "systemd.restore_state=0|1". When set to "0", none of the
8346 systemd tools will restore saved runtime state to hardware
8347 devices. More specifically, the rfkill and backlight states
8350 * The FsckPassNo= compatibility option in mount/service units
8351 has been removed. The fstab generator will now add the
8352 necessary dependencies automatically, and does not require
8353 PID1's support for that anymore.
8355 * journalctl gained a new switch, --list-boots, that lists
8356 recent boots with their times and boot IDs.
8358 * The various tools like systemctl, loginctl, timedatectl,
8359 busctl, systemd-run, ... have gained a new switch "-M" to
8360 connect to a specific, local OS container (as direct
8361 connection, without requiring SSH). This works on any
8362 container that is registered with machined, such as those
8363 created by libvirt-lxc or nspawn.
8365 * systemd-run and systemd-analyze also gained support for "-H"
8366 to connect to remote hosts via SSH. This is particularly
8367 useful for systemd-run because it enables queuing of jobs
8368 onto remote systems.
8370 * machinectl gained a new command "login" to open a getty
8371 login in any local container. This works with any container
8372 that is registered with machined (such as those created by
8373 libvirt-lxc or nspawn), and which runs systemd inside.
8375 * machinectl gained a new "reboot" command that may be used to
8376 trigger a reboot on a specific container that is registered
8377 with machined. This works on any container that runs an init
8378 system of some kind.
8380 * systemctl gained a new "list-timers" command to print a nice
8381 listing of installed timer units with the times they elapse
8384 * Alternative reboot() parameters may now be specified on the
8385 "systemctl reboot" command line and are passed to the
8386 reboot() system call.
8388 * systemctl gained a new --job-mode= switch to configure the
8389 mode to queue a job with. This is a more generic version of
8390 --fail, --irreversible, and --ignore-dependencies, which are
8391 still available but not advertised anymore.
8393 * /etc/systemd/system.conf gained new settings to configure
8394 various default timeouts of units, as well as the default
8395 start limit interval and burst. These may still be overridden
8398 * PID1 will now export on the bus profile data of the security
8399 policy upload process (such as the SELinux policy upload to
8402 * journald: when forwarding logs to the console, include
8403 timestamps (following the setting in
8404 /sys/module/printk/parameters/time).
8406 * OnCalendar= in timer units now understands the special
8407 strings "yearly" and "annually". (Both are equivalent)
8409 * The accuracy of timer units is now configurable with the new
8410 AccuracySec= setting. It defaults to 1min.
8412 * A new dependency type JoinsNamespaceOf= has been added that
8413 allows running two services within the same /tmp and network
8414 namespace, if PrivateNetwork= or PrivateTmp= are used.
8416 * A new command "cat" has been added to systemctl. It outputs
8417 the original unit file of a unit, and concatenates the
8418 contents of additional "drop-in" unit file snippets, so that
8419 the full configuration is shown.
8421 * systemctl now supports globbing on the various "list-xyz"
8422 commands, like "list-units" or "list-sockets", as well as on
8423 those commands which take multiple unit names.
8425 * journalctl's --unit= switch gained support for globbing.
8427 * All systemd daemons now make use of the watchdog logic so
8428 that systemd automatically notices when they hang.
8430 * If the $container_ttys environment variable is set,
8431 getty-generator will automatically spawn a getty for each
8432 listed tty. This is useful for container managers to request
8433 login gettys to be spawned on as many ttys as needed.
8435 * %h, %s, %U specifier support is not available anymore when
8436 used in unit files for PID 1. This is because NSS calls are
8437 not safe from PID 1. They stay available for --user
8438 instances of systemd, and as special case for the root user.
8440 * loginctl gained a new "--no-legend" switch to turn off output
8443 * The "sd-login.h" API gained three new calls:
8444 sd_session_is_remote(), sd_session_get_remote_user(),
8445 sd_session_get_remote_host() to query information about
8448 * The udev hardware database now also carries vendor/product
8449 information of SDIO devices.
8451 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
8452 determine whether watchdog notifications are requested by
8455 * Socket-activated per-connection services now include a
8456 short description of the connection parameters in the
8459 * tmpfiles gained a new "--boot" option. When this is not used,
8460 only lines where the command character is not suffixed with
8461 "!" are executed. When this option is specified, those
8462 options are executed too. This partitions tmpfiles
8463 directives into those that can be safely executed at any
8464 time, and those which should be run only at boot (for
8465 example, a line that creates /run/nologin).
8467 * A new API "sd-resolve.h" has been added which provides a simple
8468 asynchronous wrapper around glibc NSS hostname resolution
8469 calls, such as getaddrinfo(). In contrast to glibc's
8470 getaddrinfo_a(), it does not use signals. In contrast to most
8471 other asynchronous name resolution libraries, this one does
8472 not reimplement DNS, but reuses NSS, so that alternate
8473 hostname resolution systems continue to work, such as mDNS,
8474 LDAP, etc. This API is based on libasyncns, but it has been
8475 cleaned up for inclusion in systemd.
8477 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
8478 "sd-daemon.h" are no longer found in individual libraries
8479 libsystemd-journal.so, libsystemd-login.so,
8480 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
8481 merged them into a single library, libsystemd.so, which
8482 provides all symbols. The reason for this is cyclic
8483 dependencies, as these libraries tend to use each other's
8484 symbols. So far, we have managed to workaround that by linking
8485 a copy of a good part of our code into each of these
8486 libraries again and again, which, however, makes certain
8487 things hard to do, like sharing static variables. Also, it
8488 substantially increases footprint. With this change, there
8489 is only one library for the basic APIs systemd
8490 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
8491 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
8492 library as well, however are subject to the --enable-kdbus
8493 switch (see below). Note that "sd-dhcp-client.h" is not part
8494 of this library (this is because it only consumes, never
8495 provides, services of/to other APIs). To make the transition
8496 easy from the separate libraries to the unified one, we
8497 provide the --enable-compat-libs compile-time switch which
8498 will generate stub libraries that are compatible with the
8499 old ones but redirect all calls to the new one.
8501 * All of the kdbus logic and the new APIs "sd-bus.h",
8502 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
8503 and "sd-utf8.h" are compile-time optional via the
8504 "--enable-kdbus" switch, and they are not compiled in by
8505 default. To make use of kdbus, you have to explicitly enable
8506 the switch. Note however, that neither the kernel nor the
8507 userspace API for all of this is considered stable yet. We
8508 want to maintain the freedom to still change the APIs for
8509 now. By specifying this build-time switch, you acknowledge
8510 that you are aware of the instability of the current
8513 * Also, note that while kdbus is pretty much complete,
8514 it lacks one thing: proper policy support. This means you
8515 can build a fully working system with all features; however,
8516 it will be highly insecure. Policy support will be added in
8517 one of the next releases, at the same time that we will
8518 declare the APIs stable.
8520 * When the kernel command line argument "kdbus" is specified,
8521 systemd will automatically load the kdbus.ko kernel module. At
8522 this stage of development, it is only useful for testing kdbus
8523 and should not be used in production. Note: if "--enable-kdbus"
8524 is specified, and the kdbus.ko kernel module is available, and
8525 "kdbus" is added to the kernel command line, the entire system
8526 runs with kdbus instead of dbus-daemon, with the above mentioned
8527 problem of missing the system policy enforcement. Also a future
8528 version of kdbus.ko or a newer systemd will not be compatible with
8529 each other, and will unlikely be able to boot the machine if only
8530 one of them is updated.
8532 * systemctl gained a new "import-environment" command which
8533 uploads the caller's environment (or parts thereof) into the
8534 service manager so that it is inherited by services started
8535 by the manager. This is useful to upload variables like
8536 $DISPLAY into the user service manager.
8538 * A new PrivateDevices= switch has been added to service units
8539 which allows running a service with a namespaced /dev
8540 directory that does not contain any device nodes for
8541 physical devices. More specifically, it only includes devices
8542 such as /dev/null, /dev/urandom, and /dev/zero which are API
8545 * logind has been extended to support behaviour like VT
8546 switching on seats that do not support a VT. This makes
8547 multi-session available on seats that are not the first seat
8548 (seat0), and on systems where kernel support for VTs has
8549 been disabled at compile-time.
8551 * If a process holds a delay lock for system sleep or shutdown
8552 and fails to release it in time, we will now log its
8553 identity. This makes it easier to identify processes that
8554 cause slow suspends or power-offs.
8556 * When parsing /etc/crypttab, support for a new key-slot=
8557 option as supported by Debian is added. It allows indicating
8558 which LUKS slot to use on disk, speeding up key loading.
8560 * The sd_journal_sendv() API call has been checked and
8561 officially declared to be async-signal-safe so that it may
8562 be invoked from signal handlers for logging purposes.
8564 * Boot-time status output is now enabled automatically after a
8565 short timeout if boot does not progress, in order to give
8566 the user an indication what she or he is waiting for.
8568 * The boot-time output has been improved to show how much time
8569 remains until jobs expire.
8571 * The KillMode= switch in service units gained a new possible
8572 value "mixed". If set, and the unit is shut down, then the
8573 initial SIGTERM signal is sent only to the main daemon
8574 process, while the following SIGKILL signal is sent to
8575 all remaining processes of the service.
8577 * When a scope unit is registered, a new property "Controller"
8578 may be set. If set to a valid bus name, systemd will send a
8579 RequestStop() signal to this name when it would like to shut
8580 down the scope. This may be used to hook manager logic into
8581 the shutdown logic of scope units. Also, scope units may now
8582 be put in a special "abandoned" state, in which case the
8583 manager process which created them takes no further
8584 responsibilities for it.
8586 * When reading unit files, systemd will now verify
8587 the access mode of these files, and warn about certain
8588 suspicious combinations. This has been added to make it
8589 easier to track down packaging bugs where unit files are
8590 marked executable or world-writable.
8592 * systemd-nspawn gained a new "--setenv=" switch to set
8593 container-wide environment variables. The similar option in
8594 systemd-activate was renamed from "--environment=" to
8595 "--setenv=" for consistency.
8597 * systemd-nspawn has been updated to create a new kdbus domain
8598 for each container that is invoked, thus allowing each
8599 container to have its own set of system and user buses,
8600 independent of the host.
8602 * systemd-nspawn gained a new --drop-capability= switch to run
8603 the container with less capabilities than the default. Both
8604 --drop-capability= and --capability= now take the special
8605 string "all" for dropping or keeping all capabilities.
8607 * systemd-nspawn gained new switches for executing containers
8608 with specific SELinux labels set.
8610 * systemd-nspawn gained a new --quiet switch to not generate
8611 any additional output but the container's own console
8614 * systemd-nspawn gained a new --share-system switch to run a
8615 container without PID namespacing enabled.
8617 * systemd-nspawn gained a new --register= switch to control
8618 whether the container is registered with systemd-machined or
8619 not. This is useful for containers that do not run full
8620 OS images, but only specific apps.
8622 * systemd-nspawn gained a new --keep-unit which may be used
8623 when invoked as the only program from a service unit, and
8624 results in registration of the unit service itself in
8625 systemd-machined, instead of a newly opened scope unit.
8627 * systemd-nspawn gained a new --network-interface= switch for
8628 moving arbitrary interfaces to the container. The new
8629 --network-veth switch creates a virtual Ethernet connection
8630 between host and container. The new --network-bridge=
8631 switch then allows assigning the host side of this virtual
8632 Ethernet connection to a bridge device.
8634 * systemd-nspawn gained a new --personality= switch for
8635 setting the kernel personality for the container. This is
8636 useful when running a 32-bit container on a 64-bit host. A
8637 similar option Personality= is now also available for service
8640 * logind will now also track a "Desktop" identifier for each
8641 session which encodes the desktop environment of it. This is
8642 useful for desktop environments that want to identify
8643 multiple running sessions of itself easily.
8645 * A new SELinuxContext= setting for service units has been
8646 added that allows setting a specific SELinux execution
8647 context for a service.
8649 * Most systemd client tools will now honour $SYSTEMD_LESS for
8650 settings of the "less" pager. By default, these tools will
8651 override $LESS to allow certain operations to work, such as
8652 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
8653 influence this logic.
8655 * systemd's "seccomp" hook-up has been changed to make use of
8656 the libseccomp library instead of using its own
8657 implementation. This has benefits for portability among
8660 * For usage together with SystemCallFilter=, a new
8661 SystemCallErrorNumber= setting has been introduced that
8662 allows configuration of a system error number to be returned
8663 on filtered system calls, instead of immediately killing the
8664 process. Also, SystemCallArchitectures= has been added to
8665 limit access to system calls of a particular architecture
8666 (in order to turn off support for unused secondary
8667 architectures). There is also a global
8668 SystemCallArchitectures= setting in system.conf now to turn
8669 off support for non-native system calls system-wide.
8671 * systemd requires a kernel with a working name_to_handle_at(),
8672 please see the kernel config requirements in the README file.
8674 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
8675 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
8676 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
8677 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
8678 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
8679 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
8680 Elia Pinto, Florian Weimer, George McCollister, Goffredo
8681 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
8682 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
8683 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
8684 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
8685 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
8686 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
8687 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
8688 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
8689 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
8690 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
8691 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
8692 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
8693 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
8694 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
8695 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
8696 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
8697 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
8699 — Berlin, 2014-02-20
8703 * logind has gained support for facilitating privileged input
8704 and drm device access for unprivileged clients. This work is
8705 useful to allow Wayland display servers (and similar
8706 programs, such as kmscon) to run under the user's ID and
8707 access input and drm devices which are normally
8708 protected. When this is used (and the kernel is new enough)
8709 logind will "mute" IO on the file descriptors passed to
8710 Wayland as long as it is in the background and "unmute" it
8711 if it returns into the foreground. This allows secure
8712 session switching without allowing background sessions to
8713 eavesdrop on input and display data. This also introduces
8714 session switching support if VT support is turned off in the
8715 kernel, and on seats that are not seat0.
8717 * A new kernel command line option luks.options= is understood
8718 now which allows specifying LUKS options for usage for LUKS
8719 encrypted partitions specified with luks.uuid=.
8721 * tmpfiles.d(5) snippets may now use specifier expansion in
8722 path names. More specifically %m, %b, %H, %v, are now
8723 replaced by the local machine id, boot id, hostname, and
8724 kernel version number.
8726 * A new tmpfiles.d(5) command "m" has been introduced which
8727 may be used to change the owner/group/access mode of a file
8728 or directory if it exists, but do nothing if it does not.
8730 * This release removes high-level support for the
8731 MemorySoftLimit= cgroup setting. The underlying kernel
8732 cgroup attribute memory.soft_limit= is currently badly
8733 designed and likely to be removed from the kernel API in its
8734 current form, hence we should not expose it for now.
8736 * The memory.use_hierarchy cgroup attribute is now enabled for
8737 all cgroups systemd creates in the memory cgroup
8738 hierarchy. This option is likely to be come the built-in
8739 default in the kernel anyway, and the non-hierarchical mode
8740 never made much sense in the intrinsically hierarchical
8743 * A new field _SYSTEMD_SLICE= is logged along with all journal
8744 messages containing the slice a message was generated
8745 from. This is useful to allow easy per-customer filtering of
8746 logs among other things.
8748 * systemd-journald will no longer adjust the group of journal
8749 files it creates to the "systemd-journal" group. Instead we
8750 rely on the journal directory to be owned by the
8751 "systemd-journal" group, and its setgid bit set, so that the
8752 kernel file system layer will automatically enforce that
8753 journal files inherit this group assignment. The reason for
8754 this change is that we cannot allow NSS look-ups from
8755 journald which would be necessary to resolve
8756 "systemd-journal" to a numeric GID, because this might
8757 create deadlocks if NSS involves synchronous queries to
8758 other daemons (such as nscd, or sssd) which in turn are
8759 logging clients of journald and might block on it, which
8760 would then dead lock. A tmpfiles.d(5) snippet included in
8761 systemd will make sure the setgid bit and group are
8762 properly set on the journal directory if it exists on every
8763 boot. However, we recommend adjusting it manually after
8764 upgrades too (or from RPM scriptlets), so that the change is
8765 not delayed until next reboot.
8767 * Backlight and random seed files in /var/lib/ have moved into
8768 the /var/lib/systemd/ directory, in order to centralize all
8769 systemd generated files in one directory.
8771 * Boot time performance measurements (as displayed by
8772 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
8773 performance information if that's available to determine how
8774 much time BIOS and boot loader initialization required. With
8775 a sufficiently new BIOS you hence no longer need to boot
8776 with Gummiboot to get access to such information.
8778 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
8779 Cristian Rodríguez, Dave Reisner, David Herrmann, David
8780 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
8781 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
8782 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
8783 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
8784 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
8786 — Berlin, 2013-10-02
8790 * The Restart= option for services now understands a new
8791 on-watchdog setting, which will restart the service
8792 automatically if the service stops sending out watchdog keep
8793 alive messages (as configured with WatchdogSec=).
8795 * The getty generator (which is responsible for bringing up a
8796 getty on configured serial consoles) will no longer only
8797 start a getty on the primary kernel console but on all
8798 others, too. This makes the order in which console= is
8799 specified on the kernel command line less important.
8801 * libsystemd-logind gained a new sd_session_get_vt() call to
8802 retrieve the VT number of a session.
8804 * If the option "tries=0" is set for an entry of /etc/crypttab
8805 its passphrase is queried indefinitely instead of any
8806 maximum number of tries.
8808 * If a service with a configure PID file terminates its PID
8809 file will now be removed automatically if it still exists
8810 afterwards. This should put an end to stale PID files.
8812 * systemd-run will now also take relative binary path names
8813 for execution and no longer insists on absolute paths.
8815 * InaccessibleDirectories= and ReadOnlyDirectories= now take
8816 paths that are optionally prefixed with "-" to indicate that
8817 it should not be considered a failure if they do not exist.
8819 * journalctl -o (and similar commands) now understands a new
8820 output mode "short-precise", it is similar to "short" but
8821 shows timestamps with usec accuracy.
8823 * The option "discard" (as known from Debian) is now
8824 synonymous to "allow-discards" in /etc/crypttab. In fact,
8825 "discard" is preferred now (since it is easier to remember
8828 * Some licensing clean-ups were made, so that more code is now
8829 LGPL-2.1 licensed than before.
8831 * A minimal tool to save/restore the display backlight
8832 brightness across reboots has been added. It will store the
8833 backlight setting as late as possible at shutdown, and
8834 restore it as early as possible during reboot.
8836 * A logic to automatically discover and enable home and swap
8837 partitions on GPT disks has been added. With this in place
8838 /etc/fstab becomes optional for many setups as systemd can
8839 discover certain partitions located on the root disk
8840 automatically. Home partitions are recognized under their
8841 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
8842 partitions are recognized under their GPT type ID
8843 0657fd6da4ab43c484e50933c84b4f4f.
8845 * systemd will no longer pass any environment from the kernel
8846 or initrd to system services. If you want to set an
8847 environment for all services, do so via the kernel command
8848 line systemd.setenv= assignment.
8850 * The systemd-sysctl tool no longer natively reads the file
8851 /etc/sysctl.conf. If desired, the file should be symlinked
8852 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
8853 legacy support by a symlink rather than built-in code, it
8854 also makes the otherwise hidden order of application of the
8855 different files visible. (Note that this partly reverts to a
8856 pre-198 application order of sysctl knobs!)
8858 * The "systemctl set-log-level" and "systemctl dump" commands
8859 have been moved to systemd-analyze.
8861 * systemd-run learned the new --remain-after-exit switch,
8862 which causes the scope unit not to be cleaned up
8863 automatically after the process terminated.
8865 * tmpfiles learned a new --exclude-prefix= switch to exclude
8866 certain paths from operation.
8868 * journald will now automatically flush all messages to disk
8869 as soon as a message at the log level CRIT, ALERT or EMERG
8872 Contributions from: Andrew Cook, Brandon Philips, Christian
8873 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
8874 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
8875 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
8876 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
8877 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
8878 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
8879 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
8880 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
8881 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
8882 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
8883 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
8884 William Giokas, Zbigniew Jędrzejewski-Szmek
8886 — Berlin, 2013-09-13
8890 * The documentation has been updated to cover the various new
8891 concepts introduced with 205.
8893 * Unit files now understand the new %v specifier which
8894 resolves to the kernel version string as returned by "uname
8897 * systemctl now supports filtering the unit list output by
8898 load state, active state and sub state, using the new
8901 * "systemctl status" will now show the results of the
8902 condition checks (like ConditionPathExists= and similar) of
8903 the last start attempts of the unit. They are also logged to
8906 * "journalctl -b" may now be used to look for boot output of a
8907 specific boot. Try "journalctl -b -1" for the previous boot,
8908 but the syntax is substantially more powerful.
8910 * "journalctl --show-cursor" has been added which prints the
8911 cursor string the last shown log line. This may then be used
8912 with the new "journalctl --after-cursor=" switch to continue
8913 browsing logs from that point on.
8915 * "journalctl --force" may now be used to force regeneration
8918 * Creation of "dead" device nodes has been moved from udev
8919 into kmod and tmpfiles. Previously, udev would read the kmod
8920 databases to pre-generate dead device nodes based on meta
8921 information contained in kernel modules, so that these would
8922 be auto-loaded on access rather then at boot. As this
8923 does not really have much to do with the exposing actual
8924 kernel devices to userspace this has always been slightly
8925 alien in the udev codebase. Following the new scheme kmod
8926 will now generate a runtime snippet for tmpfiles from the
8927 module meta information and it now is tmpfiles' job to the
8928 create the nodes. This also allows overriding access and
8929 other parameters for the nodes using the usual tmpfiles
8930 facilities. As side effect this allows us to remove the
8931 CAP_SYS_MKNOD capability bit from udevd entirely.
8933 * logind's device ACLs may now be applied to these "dead"
8934 devices nodes too, thus finally allowing managed access to
8935 devices such as /dev/snd/sequencer without loading the
8936 backing module right-away.
8938 * A new RPM macro has been added that may be used to apply
8939 tmpfiles configuration during package installation.
8941 * systemd-detect-virt and ConditionVirtualization= now can
8942 detect User-Mode-Linux machines (UML).
8944 * journald will now implicitly log the effective capabilities
8945 set of processes in the message metadata.
8947 * systemd-cryptsetup has gained support for TrueCrypt volumes.
8949 * The initrd interface has been simplified (more specifically,
8950 support for passing performance data via environment
8951 variables and fsck results via files in /run has been
8952 removed). These features were non-essential, and are
8953 nowadays available in a much nicer way by having systemd in
8954 the initrd serialize its state and have the hosts systemd
8955 deserialize it again.
8957 * The udev "keymap" data files and tools to apply keyboard
8958 specific mappings of scan to key codes, and force-release
8959 scan code lists have been entirely replaced by a udev
8960 "keyboard" builtin and a hwdb data file.
8962 * systemd will now honour the kernel's "quiet" command line
8963 argument also during late shutdown, resulting in a
8964 completely silent shutdown when used.
8966 * There's now an option to control the SO_REUSEPORT socket
8967 option in .socket units.
8969 * Instance units will now automatically get a per-template
8970 subslice of system.slice unless something else is explicitly
8971 configured. For example, instances of sshd@.service will now
8972 implicitly be placed in system-sshd.slice rather than
8973 system.slice as before.
8975 * Test coverage support may now be enabled at build time.
8977 Contributions from: Dave Reisner, Frederic Crozat, Harald
8978 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
8979 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
8980 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
8981 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
8982 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
8983 Giokas, Zbigniew Jędrzejewski-Szmek
8985 — Berlin, 2013-07-23
8989 * Two new unit types have been introduced:
8991 Scope units are very similar to service units, however, are
8992 created out of pre-existing processes — instead of PID 1
8993 forking off the processes. By using scope units it is
8994 possible for system services and applications to group their
8995 own child processes (worker processes) in a powerful way
8996 which then maybe used to organize them, or kill them
8997 together, or apply resource limits on them.
8999 Slice units may be used to partition system resources in an
9000 hierarchical fashion and then assign other units to them. By
9001 default there are now three slices: system.slice (for all
9002 system services), user.slice (for all user sessions),
9003 machine.slice (for VMs and containers).
9005 Slices and scopes have been introduced primarily in
9006 context of the work to move cgroup handling to a
9007 single-writer scheme, where only PID 1
9008 creates/removes/manages cgroups.
9010 * There's a new concept of "transient" units. In contrast to
9011 normal units these units are created via an API at runtime,
9012 not from configuration from disk. More specifically this
9013 means it is now possible to run arbitrary programs as
9014 independent services, with all execution parameters passed
9015 in via bus APIs rather than read from disk. Transient units
9016 make systemd substantially more dynamic then it ever was,
9017 and useful as a general batch manager.
9019 * logind has been updated to make use of scope and slice units
9020 for managing user sessions. As a user logs in he will get
9021 his own private slice unit, to which all sessions are added
9022 as scope units. We also added support for automatically
9023 adding an instance of user@.service for the user into the
9024 slice. Effectively logind will no longer create cgroup
9025 hierarchies on its own now, it will defer entirely to PID 1
9026 for this by means of scope, service and slice units. Since
9027 user sessions this way become entities managed by PID 1
9028 the output of "systemctl" is now a lot more comprehensive.
9030 * A new mini-daemon "systemd-machined" has been added which
9031 may be used by virtualization managers to register local
9032 VMs/containers. nspawn has been updated accordingly, and
9033 libvirt will be updated shortly. machined will collect a bit
9034 of meta information about the VMs/containers, and assign
9035 them their own scope unit (see above). The collected
9036 meta-data is then made available via the "machinectl" tool,
9037 and exposed in "ps" and similar tools. machined/machinectl
9038 is compile-time optional.
9040 * As discussed earlier, the low-level cgroup configuration
9041 options ControlGroup=, ControlGroupModify=,
9042 ControlGroupPersistent=, ControlGroupAttribute= have been
9043 removed. Please use high-level attribute settings instead as
9044 well as slice units.
9046 * A new bus call SetUnitProperties() has been added to alter
9047 various runtime parameters of a unit. This is primarily
9048 useful to alter cgroup parameters dynamically in a nice way,
9049 but will be extended later on to make more properties
9050 modifiable at runtime. systemctl gained a new set-properties
9051 command that wraps this call.
9053 * A new tool "systemd-run" has been added which can be used to
9054 run arbitrary command lines as transient services or scopes,
9055 while configuring a number of settings via the command
9056 line. This tool is currently very basic, however already
9057 very useful. We plan to extend this tool to even allow
9058 queuing of execution jobs with time triggers from the
9059 command line, similar in fashion to "at".
9061 * nspawn will now inform the user explicitly that kernels with
9062 audit enabled break containers, and suggest the user to turn
9065 * Support for detecting the IMA and AppArmor security
9066 frameworks with ConditionSecurity= has been added.
9068 * journalctl gained a new "-k" switch for showing only kernel
9069 messages, mimicking dmesg output; in addition to "--user"
9070 and "--system" switches for showing only user's own logs
9073 * systemd-delta can now show information about drop-in
9074 snippets extending unit files.
9076 * libsystemd-bus has been substantially updated but is still
9077 not available as public API.
9079 * systemd will now look for the "debug" argument on the kernel
9080 command line and enable debug logging, similar to what
9081 "systemd.log_level=debug" already did before.
9083 * "systemctl set-default", "systemctl get-default" has been
9084 added to configure the default.target symlink, which
9085 controls what to boot into by default.
9087 * "systemctl set-log-level" has been added as a convenient
9088 way to raise and lower systemd logging threshold.
9090 * "systemd-analyze plot" will now show the time the various
9091 generators needed for execution, as well as information
9092 about the unit file loading.
9094 * libsystemd-journal gained a new sd_journal_open_files() call
9095 for opening specific journal files. journactl also gained a
9096 new switch to expose this new functionality. Previously we
9097 only supported opening all files from a directory, or all
9098 files from the system, as opening individual files only is
9099 racy due to journal file rotation.
9101 * systemd gained the new DefaultEnvironment= setting in
9102 /etc/systemd/system.conf to set environment variables for
9105 * If a privileged process logs a journal message with the
9106 OBJECT_PID= field set, then journald will automatically
9107 augment this with additional OBJECT_UID=, OBJECT_GID=,
9108 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
9109 system services want to log events about specific client
9110 processes. journactl/systemctl has been updated to make use
9111 of this information if all log messages regarding a specific
9114 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
9115 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
9116 Reisner, David Coppa, David King, David Strauss, Eelco
9117 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
9118 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
9119 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
9120 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
9121 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
9122 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
9123 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
9124 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
9125 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
9126 Łukasz Stelmach, 장동준
9130 * The Python bindings gained some minimal support for the APIs
9131 exposed by libsystemd-logind.
9133 * ConditionSecurity= gained support for detecting SMACK. Since
9134 this condition already supports SELinux and AppArmor we only
9135 miss IMA for this. Patches welcome!
9137 Contributions from: Karol Lewandowski, Lennart Poettering,
9138 Zbigniew Jędrzejewski-Szmek
9142 * systemd-nspawn will now create /etc/resolv.conf if
9143 necessary, before bind-mounting the host's file onto it.
9145 * systemd-nspawn will now store meta information about a
9146 container on the container's cgroup as extended attribute
9147 fields, including the root directory.
9149 * The cgroup hierarchy has been reworked in many ways. All
9150 objects any of the components systemd creates in the cgroup
9151 tree are now suffixed. More specifically, user sessions are
9152 now placed in cgroups suffixed with ".session", users in
9153 cgroups suffixed with ".user", and nspawn containers in
9154 cgroups suffixed with ".nspawn". Furthermore, all cgroup
9155 names are now escaped in a simple scheme to avoid collision
9156 of userspace object names with kernel filenames. This work
9157 is preparation for making these objects relocatable in the
9158 cgroup tree, in order to allow easy resource partitioning of
9159 these objects without causing naming conflicts.
9161 * systemctl list-dependencies gained the new switches
9162 --plain, --reverse, --after and --before.
9164 * systemd-inhibit now shows the process name of processes that
9165 have taken an inhibitor lock.
9167 * nss-myhostname will now also resolve "localhost"
9168 implicitly. This makes /etc/hosts an optional file and
9169 nicely handles that on IPv6 ::1 maps to both "localhost" and
9172 * libsystemd-logind.so gained a new call
9173 sd_get_machine_names() to enumerate running containers and
9174 VMs (currently only supported by very new libvirt and
9175 nspawn). sd_login_monitor can now be used to watch
9176 VMs/containers coming and going.
9178 * .include is not allowed recursively anymore, and only in
9179 unit files. Usually it is better to use drop-in snippets in
9180 .d/*.conf anyway, as introduced with systemd 198.
9182 * systemd-analyze gained a new "critical-chain" command that
9183 determines the slowest chain of units run during system
9184 boot-up. It is very useful for tracking down where
9185 optimizing boot time is the most beneficial.
9187 * systemd will no longer allow manipulating service paths in
9188 the name=systemd:/system cgroup tree using ControlGroup= in
9189 units. (But is still fine with it in all other dirs.)
9191 * There's a new systemd-nspawn@.service service file that may
9192 be used to easily run nspawn containers as system
9193 services. With the container's root directory in
9194 /var/lib/container/foobar it is now sufficient to run
9195 "systemctl start systemd-nspawn@foobar.service" to boot it.
9197 * systemd-cgls gained a new parameter "--machine" to list only
9198 the processes within a certain container.
9200 * ConditionSecurity= now can check for "apparmor". We still
9201 are lacking checks for SMACK and IMA for this condition
9202 check though. Patches welcome!
9204 * A new configuration file /etc/systemd/sleep.conf has been
9205 added that may be used to configure which kernel operation
9206 systemd is supposed to execute when "suspend", "hibernate"
9207 or "hybrid-sleep" is requested. This makes the new kernel
9208 "freeze" state accessible to the user.
9210 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
9211 the passed argument if applicable.
9213 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
9214 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
9215 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
9216 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
9217 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
9218 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
9219 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
9224 * The output of 'systemctl list-jobs' got some polishing. The
9225 '--type=' argument may now be passed more than once. A new
9226 command 'systemctl list-sockets' has been added which shows
9227 a list of kernel sockets systemd is listening on with the
9228 socket units they belong to, plus the units these socket
9231 * The experimental libsystemd-bus library got substantial
9232 updates to work in conjunction with the (also experimental)
9233 kdbus kernel project. It works well enough to exchange
9234 messages with some sophistication. Note that kdbus is not
9235 ready yet, and the library is mostly an elaborate test case
9236 for now, and not installable.
9238 * systemd gained a new unit 'systemd-static-nodes.service'
9239 that generates static device nodes earlier during boot, and
9240 can run in conjunction with udev.
9242 * libsystemd-login gained a new call sd_pid_get_user_unit()
9243 to retrieve the user systemd unit a process is running
9244 in. This is useful for systems where systemd is used as
9247 * systemd-nspawn now places all containers in the new /machine
9248 top-level cgroup directory in the name=systemd
9249 hierarchy. libvirt will soon do the same, so that we get a
9250 uniform separation of /system, /user and /machine for system
9251 services, user processes and containers/virtual
9252 machines. This new cgroup hierarchy is also useful to stick
9253 stable names to specific container instances, which can be
9254 recognized later this way (this name may be controlled
9255 via systemd-nspawn's new -M switch). libsystemd-login also
9256 gained a new call sd_pid_get_machine_name() to retrieve the
9257 name of the container/VM a specific process belongs to.
9259 * bootchart can now store its data in the journal.
9261 * libsystemd-journal gained a new call
9262 sd_journal_add_conjunction() for AND expressions to the
9263 matching logic. This can be used to express more complex
9264 logical expressions.
9266 * journactl can now take multiple --unit= and --user-unit=
9269 * The cryptsetup logic now understands the "luks.key=" kernel
9270 command line switch for specifying a file to read the
9271 decryption key from. Also, if a configured key file is not
9272 found the tool will now automatically fall back to prompting
9275 * Python systemd.journal module was updated to wrap recently
9276 added functions from libsystemd-journal. The interface was
9277 changed to bring the low level interface in s.j._Reader
9278 closer to the C API, and the high level interface in
9279 s.j.Reader was updated to wrap and convert all data about
9282 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
9283 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
9284 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
9285 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
9286 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
9287 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
9291 * journalctl --update-catalog now understands a new --root=
9292 option to operate on catalogs found in a different root
9295 * During shutdown after systemd has terminated all running
9296 services a final killing loop kills all remaining left-over
9297 processes. We will now print the name of these processes
9298 when we send SIGKILL to them, since this usually indicates a
9301 * If /etc/crypttab refers to password files stored on
9302 configured mount points automatic dependencies will now be
9303 generated to ensure the specific mount is established first
9304 before the key file is attempted to be read.
9306 * 'systemctl status' will now show information about the
9307 network sockets a socket unit is listening on.
9309 * 'systemctl status' will also shown information about any
9310 drop-in configuration file for units. (Drop-In configuration
9311 files in this context are files such as
9312 /etc/systemd/systemd/foobar.service.d/*.conf)
9314 * systemd-cgtop now optionally shows summed up CPU times of
9315 cgroups. Press '%' while running cgtop to switch between
9316 percentage and absolute mode. This is useful to determine
9317 which cgroups use up the most CPU time over the entire
9318 runtime of the system. systemd-cgtop has also been updated
9319 to be 'pipeable' for processing with further shell tools.
9321 * 'hostnamectl set-hostname' will now allow setting of FQDN
9324 * The formatting and parsing of time span values has been
9325 changed. The parser now understands fractional expressions
9326 such as "5.5h". The formatter will now output fractional
9327 expressions for all time spans under 1min, i.e. "5.123456s"
9328 rather than "5s 123ms 456us". For time spans under 1s
9329 millisecond values are shown, for those under 1ms
9330 microsecond values are shown. This should greatly improve
9331 all time-related output of systemd.
9333 * libsystemd-login and libsystemd-journal gained new
9334 functions for querying the poll() events mask and poll()
9335 timeout value for integration into arbitrary event
9338 * localectl gained the ability to list available X11 keymaps
9339 (models, layouts, variants, options).
9341 * 'systemd-analyze dot' gained the ability to filter for
9342 specific units via shell-style globs, to create smaller,
9343 more useful graphs. I.e. it is now possible to create simple
9344 graphs of all the dependencies between only target units, or
9345 of all units that Avahi has dependencies with.
9347 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
9348 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
9349 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
9350 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
9351 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
9352 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
9353 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
9357 * The boot-time readahead implementation for rotating media
9358 will now read the read-ahead data in multiple passes which
9359 consist of all read requests made in equidistant time
9360 intervals. This means instead of strictly reading read-ahead
9361 data in its physical order on disk we now try to find a
9362 middle ground between physical and access time order.
9364 * /etc/os-release files gained a new BUILD_ID= field for usage
9365 on operating systems that provide continuous builds of OS
9368 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
9369 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
9370 William Douglas, Zbigniew Jędrzejewski-Szmek
9374 * systemd-python gained an API exposing libsystemd-daemon.
9376 * The SMACK setup logic gained support for uploading CIPSO
9379 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
9380 ReadOnlyDirectories= and InaccessibleDirectories= has
9381 changed. The private /tmp and /var/tmp directories are now
9382 shared by all processes of a service (which means
9383 ExecStartPre= may now leave data in /tmp that ExecStart= of
9384 the same service can still access). When a service is
9385 stopped its temporary directories are immediately deleted
9386 (normal clean-up with tmpfiles is still done in addition to
9389 * By default, systemd will now set a couple of sysctl
9390 variables in the kernel: the safe sysrq options are turned
9391 on, IP route verification is turned on, and source routing
9392 disabled. The recently added hardlink and softlink
9393 protection of the kernel is turned on. These settings should
9394 be reasonably safe, and good defaults for all new systems.
9396 * The predictable network naming logic may now be turned off
9397 with a new kernel command line switch: net.ifnames=0.
9399 * A new libsystemd-bus module has been added that implements a
9400 pretty complete D-Bus client library. For details see:
9402 https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
9404 * journald will now explicitly flush the journal files to disk
9405 at the latest 5min after each write. The file will then also
9406 be marked offline until the next write. This should increase
9407 reliability in case of a crash. The synchronization delay
9408 can be configured via SyncIntervalSec= in journald.conf.
9410 * There's a new remote-fs-setup.target unit that can be used
9411 to pull in specific services when at least one remote file
9412 system is to be mounted.
9414 * There are new targets timers.target and paths.target as
9415 canonical targets to pull user timer and path units in
9416 from. This complements sockets.target with a similar
9417 purpose for socket units.
9419 * libudev gained a new call udev_device_set_attribute_value()
9420 to set sysfs attributes of a device.
9422 * The udev daemon now sets the default number of worker
9423 processes executed in parallel based on the number of available
9424 CPUs instead of the amount of available RAM. This is supposed
9425 to provide a more reliable default and limit a too aggressive
9426 parallelism for setups with 1000s of devices connected.
9428 Contributions from: Auke Kok, Colin Walters, Cristian
9429 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
9430 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
9431 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
9432 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
9433 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
9434 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
9435 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
9436 Zbigniew Jędrzejewski-Szmek
9440 * Configuration of unit files may now be extended via drop-in
9441 files without having to edit/override the unit files
9442 themselves. More specifically, if the administrator wants to
9443 change one value for a service file foobar.service he can
9444 now do so by dropping in a configuration snippet into
9445 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
9446 will load all these snippets and apply them on top of the
9447 main unit configuration file, possibly extending or
9448 overriding its settings. Using these drop-in snippets is
9449 generally nicer than the two earlier options for changing
9450 unit files locally: copying the files from
9451 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
9452 them there; or creating a new file in /etc/systemd/system/
9453 that incorporates the original one via ".include". Drop-in
9454 snippets into these .d/ directories can be placed in any
9455 directory systemd looks for units in, and the usual
9456 overriding semantics between /usr/lib, /etc and /run apply
9459 * Most unit file settings which take lists of items can now be
9460 reset by assigning the empty string to them. For example,
9461 normally, settings such as Environment=FOO=BAR append a new
9462 environment variable assignment to the environment block,
9463 each time they are used. By assigning Environment= the empty
9464 string the environment block can be reset to empty. This is
9465 particularly useful with the .d/*.conf drop-in snippets
9466 mentioned above, since this adds the ability to reset list
9467 settings from vendor unit files via these drop-ins.
9469 * systemctl gained a new "list-dependencies" command for
9470 listing the dependencies of a unit recursively.
9472 * Inhibitors are now honored and listed by "systemctl
9473 suspend", "systemctl poweroff" (and similar) too, not only
9474 GNOME. These commands will also list active sessions by
9477 * Resource limits (as exposed by the various control group
9478 controllers) can now be controlled dynamically at runtime
9479 for all units. More specifically, you can now use a command
9480 like "systemctl set-cgroup-attr foobar.service cpu.shares
9481 2000" to alter the CPU shares a specific service gets. These
9482 settings are stored persistently on disk, and thus allow the
9483 administrator to easily adjust the resource usage of
9484 services with a few simple commands. This dynamic resource
9485 management logic is also available to other programs via the
9486 bus. Almost any kernel cgroup attribute and controller is
9489 * systemd-vconsole-setup will now copy all font settings to
9490 all allocated VTs, where it previously applied them only to
9493 * libsystemd-login gained the new sd_session_get_tty() API
9496 * This release drops support for a few legacy or
9497 distribution-specific LSB facility names when parsing init
9498 scripts: $x-display-manager, $mail-transfer-agent,
9499 $mail-transport-agent, $mail-transfer-agent, $smtp,
9500 $null. Also, the mail-transfer-agent.target unit backing
9501 this has been removed. Distributions which want to retain
9502 compatibility with this should carry the burden for
9503 supporting this themselves and patch support for these back
9504 in, if they really need to. Also, the facilities $syslog and
9505 $local_fs are now ignored, since systemd does not support
9506 early-boot LSB init scripts anymore, and these facilities
9507 are implied anyway for normal services. syslog.target has
9510 * There are new bus calls on PID1's Manager object for
9511 cancelling jobs, and removing snapshot units. Previously,
9512 both calls were only available on the Job and Snapshot
9515 * systemd-journal-gatewayd gained SSL support.
9517 * The various "environment" files, such as /etc/locale.conf
9518 now support continuation lines with a backslash ("\") as
9519 last character in the line, similarly in style (but different)
9520 to how this is supported in shells.
9522 * For normal user processes the _SYSTEMD_USER_UNIT= field is
9523 now implicitly appended to every log entry logged. systemctl
9524 has been updated to filter by this field when operating on a
9525 user systemd instance.
9527 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
9528 CAP_AUDIT_CONTROL capabilities to the capabilities set for
9529 the container. This makes it easier to boot unmodified
9530 Fedora systems in a container, which however still requires
9531 audit=0 to be passed on the kernel command line. Auditing in
9532 kernel and userspace is unfortunately still too broken in
9533 context of containers, hence we recommend compiling it out
9534 of the kernel or using audit=0. Hopefully this will be fixed
9535 one day for good in the kernel.
9537 * nspawn gained the new --bind= and --bind-ro= parameters to
9538 bind mount specific directories from the host into the
9541 * nspawn will now mount its own devpts file system instance
9542 into the container, in order not to leak pty devices from
9543 the host into the container.
9545 * systemd will now read the firmware boot time performance
9546 information from the EFI variables, if the used boot loader
9547 supports this, and takes it into account for boot performance
9548 analysis via "systemd-analyze". This is currently supported
9549 only in conjunction with Gummiboot, but could be supported
9550 by other boot loaders too. For details see:
9552 https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
9554 * A new generator has been added that automatically mounts the
9555 EFI System Partition (ESP) to /boot, if that directory
9556 exists, is empty, and no other file system has been
9557 configured to be mounted there.
9559 * logind will now send out PrepareForSleep(false) out
9560 unconditionally, after coming back from suspend. This may be
9561 used by applications as asynchronous notification for
9562 system resume events.
9564 * "systemctl unlock-sessions" has been added, that allows
9565 unlocking the screens of all user sessions at once, similar
9566 to how "systemctl lock-sessions" already locked all users
9567 sessions. This is backed by a new D-Bus call UnlockSessions().
9569 * "loginctl seat-status" will now show the master device of a
9570 seat. (i.e. the device of a seat that needs to be around for
9571 the seat to be considered available, usually the graphics
9574 * tmpfiles gained a new "X" line type, that allows
9575 configuration of files and directories (with wildcards) that
9576 shall be excluded from automatic cleanup ("aging").
9578 * udev default rules set the device node permissions now only
9579 at "add" events, and do not change them any longer with a
9580 later "change" event.
9582 * The log messages for lid events and power/sleep keypresses
9583 now carry a message ID.
9585 * We now have a substantially larger unit test suite, but this
9586 continues to be work in progress.
9588 * udevadm hwdb gained a new --root= parameter to change the
9589 root directory to operate relative to.
9591 * logind will now issue a background sync() request to the kernel
9592 early at shutdown, so that dirty buffers are flushed to disk early
9593 instead of at the last moment, in order to optimize shutdown
9596 * A new bootctl tool has been added that is an interface for
9597 certain boot loader operations. This is currently a preview
9598 and is likely to be extended into a small mechanism daemon
9599 like timedated, localed, hostnamed, and can be used by
9600 graphical UIs to enumerate available boot options, and
9601 request boot into firmware operations.
9603 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
9604 the rest of the package. It also has been updated to work
9605 correctly in initrds.
9607 * polkit previously has been runtime optional, and is now also
9608 compile time optional via a configure switch.
9610 * systemd-analyze has been reimplemented in C. Also "systemctl
9611 dot" has moved into systemd-analyze.
9613 * "systemctl status" with no further parameters will now print
9614 the status of all active or failed units.
9616 * Operations such as "systemctl start" can now be executed
9617 with a new mode "--irreversible" which may be used to queue
9618 operations that cannot accidentally be reversed by a later
9619 job queuing. This is by default used to make shutdown
9620 requests more robust.
9622 * The Python API of systemd now gained a new module for
9623 reading journal files.
9625 * A new tool kernel-install has been added that can install
9626 kernel images according to the Boot Loader Specification:
9628 https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
9630 * Boot time console output has been improved to provide
9631 animated boot time output for hanging jobs.
9633 * A new tool systemd-activate has been added which can be used
9634 to test socket activation with, directly from the command
9635 line. This should make it much easier to test and debug
9636 socket activation in daemons.
9638 * journalctl gained a new "--reverse" (or -r) option to show
9639 journal output in reverse order (i.e. newest line first).
9641 * journalctl gained a new "--pager-end" (or -e) option to jump
9642 to immediately jump to the end of the journal in the
9643 pager. This is only supported in conjunction with "less".
9645 * journalctl gained a new "--user-unit=" option, that works
9646 similarly to "--unit=" but filters for user units rather than
9649 * A number of unit files to ease adoption of systemd in
9650 initrds has been added. This moves some minimal logic from
9651 the various initrd implementations into systemd proper.
9653 * The journal files are now owned by a new group
9654 "systemd-journal", which exists specifically to allow access
9655 to the journal, and nothing else. Previously, we used the
9656 "adm" group for that, which however possibly covers more
9657 than just journal/log file access. This new group is now
9658 already used by systemd-journal-gatewayd to ensure this
9659 daemon gets access to the journal files and as little else
9660 as possible. Note that "make install" will also set FS ACLs
9661 up for /var/log/journal to give "adm" and "wheel" read
9662 access to it, in addition to "systemd-journal" which owns
9663 the journal files. We recommend that packaging scripts also
9664 add read access to "adm" + "wheel" to /var/log/journal, and
9665 all existing/future journal files. To normal users and
9666 administrators little changes, however packagers need to
9667 ensure to create the "systemd-journal" system group at
9668 package installation time.
9670 * The systemd-journal-gatewayd now runs as unprivileged user
9671 systemd-journal-gateway:systemd-journal-gateway. Packaging
9672 scripts need to create these system user/group at
9675 * timedated now exposes a new boolean property CanNTP that
9676 indicates whether a local NTP service is available or not.
9678 * systemd-detect-virt will now also detect xen PVs
9680 * The pstore file system is now mounted by default, if it is
9683 * In addition to the SELinux and IMA policies we will now also
9684 load SMACK policies at early boot.
9686 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
9687 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
9688 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
9689 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
9690 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
9691 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
9692 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
9693 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
9694 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
9695 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
9696 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
9697 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
9698 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
9699 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
9703 * Timer units now support calendar time events in addition to
9704 monotonic time events. That means you can now trigger a unit
9705 based on a calendar time specification such as "Thu,Fri
9706 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
9707 or fifth day of any month of the year 2013, given that it is
9708 a thursday or friday. This brings timer event support
9709 considerably closer to cron's capabilities. For details on
9710 the supported calendar time specification language see
9713 * udev now supports a number of different naming policies for
9714 network interfaces for predictable names, and a combination
9715 of these policies is now the default. Please see this wiki
9716 document for details:
9718 https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
9720 * Auke Kok's bootchart implementation has been added to the
9721 systemd tree. It is an optional component that can graph the
9722 boot in quite some detail. It is one of the best bootchart
9723 implementations around and minimal in its code and
9726 * nss-myhostname has been integrated into the systemd source
9727 tree. nss-myhostname guarantees that the local hostname
9728 always stays resolvable via NSS. It has been a weak
9729 requirement of systemd-hostnamed since a long time, and
9730 since its code is actually trivial we decided to just
9731 include it in systemd's source tree. It can be turned off
9732 with a configure switch.
9734 * The read-ahead logic is now capable of properly detecting
9735 whether a btrfs file system is on SSD or rotating media, in
9736 order to optimize the read-ahead scheme. Previously, it was
9737 only capable of detecting this on traditional file systems
9740 * In udev, additional device properties are now read from the
9741 IAB in addition to the OUI database. Also, Bluetooth company
9742 identities are attached to the devices as well.
9744 * In service files %U may be used as specifier that is
9745 replaced by the configured user name of the service.
9747 * nspawn may now be invoked without a controlling TTY. This
9748 makes it suitable for invocation as its own service. This
9749 may be used to set up a simple containerized server system
9750 using only core OS tools.
9752 * systemd and nspawn can now accept socket file descriptors
9753 when they are started for socket activation. This enables
9754 implementation of socket activated nspawn
9755 containers. i.e. think about autospawning an entire OS image
9756 when the first SSH or HTTP connection is received. We expect
9757 that similar functionality will also be added to libvirt-lxc
9760 * journalctl will now suppress ANSI color codes when
9761 presenting log data.
9763 * systemctl will no longer show control group information for
9764 a unit if the control group is empty anyway.
9766 * logind can now automatically suspend/hibernate/shutdown the
9769 * /etc/machine-info and hostnamed now also expose the chassis
9770 type of the system. This can be used to determine whether
9771 the local system is a laptop, desktop, handset or
9772 tablet. This information may either be configured by the
9773 user/vendor or is automatically determined from ACPI and DMI
9774 information if possible.
9776 * A number of polkit actions are now bound together with "imply"
9777 rules. This should simplify creating UIs because many actions
9778 will now authenticate similar ones as well.
9780 * Unit files learnt a new condition ConditionACPower= which
9781 may be used to conditionalize a unit depending on whether an
9782 AC power source is connected or not, of whether the system
9783 is running on battery power.
9785 * systemctl gained a new "is-failed" verb that may be used in
9786 shell scripts and suchlike to check whether a specific unit
9787 is in the "failed" state.
9789 * The EnvironmentFile= setting in unit files now supports file
9790 globbing, and can hence be used to easily read a number of
9791 environment files at once.
9793 * systemd will no longer detect and recognize specific
9794 distributions. All distribution-specific #ifdeffery has been
9795 removed, systemd is now fully generic and
9796 distribution-agnostic. Effectively, not too much is lost as
9797 a lot of the code is still accessible via explicit configure
9798 switches. However, support for some distribution specific
9799 legacy configuration file formats has been dropped. We
9800 recommend distributions to simply adopt the configuration
9801 files everybody else uses now and convert the old
9802 configuration from packaging scripts. Most distributions
9803 already did that. If that's not possible or desirable,
9804 distributions are welcome to forward port the specific
9805 pieces of code locally from the git history.
9807 * When logging a message about a unit systemd will now always
9808 log the unit name in the message meta data.
9810 * localectl will now also discover system locale data that is
9811 not stored in locale archives, but directly unpacked.
9813 * logind will no longer unconditionally use framebuffer
9814 devices as seat masters, i.e. as devices that are required
9815 to be existing before a seat is considered preset. Instead,
9816 it will now look for all devices that are tagged as
9817 "seat-master" in udev. By default, framebuffer devices will
9818 be marked as such, but depending on local systems, other
9819 devices might be marked as well. This may be used to
9820 integrate graphics cards using closed source drivers (such
9821 as NVidia ones) more nicely into logind. Note however, that
9822 we recommend using the open source NVidia drivers instead,
9823 and no udev rules for the closed-source drivers will be
9824 shipped from us upstream.
9826 Contributions from: Adam Williamson, Alessandro Crismani, Auke
9827 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
9828 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
9829 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
9830 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
9831 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
9832 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
9833 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
9834 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
9835 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
9836 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
9837 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
9842 * udev gained support for loading additional device properties
9843 from an indexed database that is keyed by vendor/product IDs
9844 and similar device identifiers. For the beginning this
9845 "hwdb" is populated with data from the well-known PCI and
9846 USB database, but also includes PNP, ACPI and OID data. In
9847 the longer run this indexed database shall grow into
9848 becoming the one central database for non-essential
9849 userspace device metadata. Previously, data from the PCI/USB
9850 database was only attached to select devices, since the
9851 lookup was a relatively expensive operation due to O(n) time
9852 complexity (with n being the number of entries in the
9853 database). Since this is now O(1), we decided to add in this
9854 data for all devices where this is available, by
9855 default. Note that the indexed database needs to be rebuilt
9856 when new data files are installed. To achieve this you need
9857 to update your packaging scripts to invoke "udevadm hwdb
9858 --update" after installation of hwdb data files. For
9859 RPM-based distributions we introduced the new
9860 %udev_hwdb_update macro for this purpose.
9862 * The Journal gained support for the "Message Catalog", an
9863 indexed database to link up additional information with
9864 journal entries. For further details please check:
9866 https://www.freedesktop.org/wiki/Software/systemd/catalog
9868 The indexed message catalog database also needs to be
9869 rebuilt after installation of message catalog files. Use
9870 "journalctl --update-catalog" for this. For RPM-based
9871 distributions we introduced the %journal_catalog_update
9872 macro for this purpose.
9874 * The Python Journal bindings gained support for the standard
9875 Python logging framework.
9877 * The Journal API gained new functions for checking whether
9878 the underlying file system of a journal file is capable of
9879 properly reporting file change notifications, or whether
9880 applications that want to reflect journal changes "live"
9881 need to recheck journal files continuously in appropriate
9884 * It is now possible to set the "age" field for tmpfiles
9885 entries to 0, indicating that files matching this entry
9886 shall always be removed when the directories are cleaned up.
9888 * coredumpctl gained a new "gdb" verb which invokes gdb
9889 right-away on the selected coredump.
9891 * There's now support for "hybrid sleep" on kernels that
9892 support this, in addition to "suspend" and "hibernate". Use
9893 "systemctl hybrid-sleep" to make use of this.
9895 * logind's HandleSuspendKey= setting (and related settings)
9896 now gained support for a new "lock" setting to simply
9897 request the screen lock on all local sessions, instead of
9898 actually executing a suspend or hibernation.
9900 * systemd will now mount the EFI variables file system by
9903 * Socket units now gained support for configuration of the
9904 SMACK security label.
9906 * timedatectl will now output the time of the last and next
9907 daylight saving change.
9909 * We dropped support for various legacy and distro-specific
9910 concepts, such as insserv, early-boot SysV services
9911 (i.e. those for non-standard runlevels such as 'b' or 'S')
9912 or ArchLinux /etc/rc.conf support. We recommend the
9913 distributions who still need support this to either continue
9914 to maintain the necessary patches downstream, or find a
9915 different solution. (Talk to us if you have questions!)
9917 * Various systemd components will now bypass polkit checks for
9918 root and otherwise handle properly if polkit is not found to
9919 be around. This should fix most issues for polkit-less
9920 systems. Quite frankly this should have been this way since
9921 day one. It is absolutely our intention to make systemd work
9922 fine on polkit-less systems, and we consider it a bug if
9923 something does not work as it should if polkit is not around.
9925 * For embedded systems it is now possible to build udev and
9926 systemd without blkid and/or kmod support.
9928 * "systemctl switch-root" is now capable of switching root
9929 more than once. I.e. in addition to transitions from the
9930 initrd to the host OS it is now possible to transition to
9931 further OS images from the host. This is useful to implement
9932 offline updating tools.
9934 * Various other additions have been made to the RPM macros
9935 shipped with systemd. Use %udev_rules_update() after
9936 installing new udev rules files. %_udevhwdbdir,
9937 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
9938 %_sysctldir are now available which resolve to the right
9939 directories for packages to place various data files in.
9941 * journalctl gained the new --full switch (in addition to
9942 --all, to disable ellipsation for long messages.
9944 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
9945 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
9946 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
9947 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
9948 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
9949 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
9950 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
9951 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
9952 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
9956 * journalctl gained new --since= and --until= switches to
9957 filter by time. It also now supports nice filtering for
9958 units via --unit=/-u.
9960 * Type=oneshot services may use ExecReload= and do the
9963 * The journal daemon now supports time-based rotation and
9964 vacuuming, in addition to the usual disk-space based
9967 * The journal will now index the available field values for
9968 each field name. This enables clients to show pretty drop
9969 downs of available match values when filtering. The bash
9970 completion of journalctl has been updated
9971 accordingly. journalctl gained a new switch -F to list all
9972 values a certain field takes in the journal database.
9974 * More service events are now written as structured messages
9975 to the journal, and made recognizable via message IDs.
9977 * The timedated, localed and hostnamed mini-services which
9978 previously only provided support for changing time, locale
9979 and hostname settings from graphical DEs such as GNOME now
9980 also have a minimal (but very useful) text-based client
9981 utility each. This is probably the nicest way to changing
9982 these settings from the command line now, especially since
9983 it lists available options and is fully integrated with bash
9986 * There's now a new tool "systemd-coredumpctl" to list and
9987 extract coredumps from the journal.
9989 * We now install a README each in /var/log/ and
9990 /etc/rc.d/init.d explaining where the system logs and init
9991 scripts went. This hopefully should help folks who go to
9992 that dirs and look into the otherwise now empty void and
9993 scratch their heads.
9995 * When user-services are invoked (by systemd --user) the
9996 $MANAGERPID env var is set to the PID of systemd.
9998 * SIGRTMIN+24 when sent to a --user instance will now result
9999 in immediate termination of systemd.
10001 * gatewayd received numerous feature additions such as a
10002 "follow" mode, for live syncing and filtering.
10004 * browse.html now allows filtering and showing detailed
10005 information on specific entries. Keyboard navigation and
10006 mouse screen support has been added.
10008 * gatewayd/journalctl now supports HTML5/JSON
10009 Server-Sent-Events as output.
10011 * The SysV init script compatibility logic will now
10012 heuristically determine whether a script supports the
10013 "reload" verb, and only then make this available as
10014 "systemctl reload".
10016 * "systemctl status --follow" has been removed, use "journalctl
10019 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
10020 have been removed since they are hardly useful to be
10023 * And I'd like to take the opportunity to specifically mention
10024 Zbigniew for his great contributions. Zbigniew, you rock!
10026 Contributions from: Andrew Eikum, Christian Hesse, Colin
10027 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
10028 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
10029 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
10030 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
10031 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
10032 Jędrzejewski-Szmek, Сковорода Никита Андреевич
10036 * If /etc/vconsole.conf is non-existent or empty we will no
10037 longer load any console font or key map at boot by
10038 default. Instead the kernel defaults will be left
10039 intact. This is definitely the right thing to do, as no
10040 configuration should mean no configuration, and hard-coding
10041 font names that are different on all archs is probably a bad
10042 idea. Also, the kernel default key map and font should be
10043 good enough for most cases anyway, and mostly identical to
10044 the userspace fonts/key maps we previously overloaded them
10045 with. If distributions want to continue to default to a
10046 non-kernel font or key map they should ship a default
10047 /etc/vconsole.conf with the appropriate contents.
10049 Contributions from: Colin Walters, Daniel J Walsh, Dave
10050 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
10051 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
10055 * journalctl gained a new --cursor= switch to show entries
10056 starting from the specified location in the journal.
10058 * We now enforce a size limit on journal entry fields exported
10059 with "-o json" in journalctl. Fields larger than 4K will be
10060 assigned null. This can be turned off with --all.
10062 * An (optional) journal gateway daemon is now available as
10063 "systemd-journal-gatewayd.service". This service provides
10064 access to the journal via HTTP and JSON. This functionality
10065 will be used to implement live log synchronization in both
10066 pull and push modes, but has various other users too, such
10067 as easy log access for debugging of embedded devices. Right
10068 now it is already useful to retrieve the journal via HTTP:
10070 # systemctl start systemd-journal-gatewayd.service
10071 # wget http://localhost:19531/entries
10073 This will download the journal contents in a
10074 /var/log/messages compatible format. The same as JSON:
10076 # curl -H"Accept: application/json" http://localhost:19531/entries
10078 This service is also accessible via a web browser where a
10079 single static HTML5 app is served that uses the JSON logic
10080 to enable the user to do some basic browsing of the
10081 journal. This will be extended later on. Here's an example
10082 screenshot of this app in its current state:
10084 http://0pointer.de/public/journal-gatewayd
10086 Contributions from: Kay Sievers, Lennart Poettering, Robert
10087 Milasan, Tom Gundersen
10091 * The bash completion logic is now available for journalctl
10094 * We do not mount the "cpuset" controller anymore together with
10095 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
10096 started if no parameters are assigned to it. "cpuset" hence
10097 broke code that assumed it could create "cpu" groups and
10100 * journalctl -f will now subscribe to terminal size changes,
10101 and line break accordingly.
10103 Contributions from: Dave Reisner, Kay Sievers, Lennart
10104 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
10108 * nspawn will now create a symlink /etc/localtime in the
10109 container environment, copying the host's timezone
10110 setting. Previously this has been done via a bind mount, but
10111 since symlinks cannot be bind mounted this has now been
10112 changed to create/update the appropriate symlink.
10114 * journalctl -n's line number argument is now optional, and
10115 will default to 10 if omitted.
10117 * journald will now log the maximum size the journal files may
10118 take up on disk. This is particularly useful if the default
10119 built-in logic of determining this parameter from the file
10120 system size is used. Use "systemctl status
10121 systemd-journald.service" to see this information.
10123 * The multi-seat X wrapper tool has been stripped down. As X
10124 is now capable of enumerating graphics devices via udev in a
10125 seat-aware way the wrapper is not strictly necessary
10126 anymore. A stripped down temporary stop-gap is still shipped
10127 until the upstream display managers have been updated to
10128 fully support the new X logic. Expect this wrapper to be
10129 removed entirely in one of the next releases.
10131 * HandleSleepKey= in logind.conf has been split up into
10132 HandleSuspendKey= and HandleHibernateKey=. The old setting
10133 is not available anymore. X11 and the kernel are
10134 distinguishing between these keys and we should too. This
10135 also means the inhibition lock for these keys has been split
10138 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
10139 Poettering, Lukas Nykryn, Václav Pavlín
10143 * Whenever a unit changes state we will now log this to the
10144 journal and show along the unit's own log output in
10145 "systemctl status".
10147 * ConditionPathIsMountPoint= can now properly detect bind
10148 mount points too. (Previously, a bind mount of one file
10149 system to another place in the same file system could not be
10150 detected as mount, since they shared struct stat's st_dev
10153 * We will now mount the cgroup controllers cpu, cpuacct,
10154 cpuset and the controllers net_cls, net_prio together by
10157 * nspawn containers will now have a virtualized boot
10158 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
10159 over with a randomized ID at container initialization). This
10160 has the effect of making "journalctl -b" do the right thing
10163 * The JSON output journal serialization has been updated not
10164 to generate "endless" list objects anymore, but rather one
10165 JSON object per line. This is more in line how most JSON
10166 parsers expect JSON objects. The new output mode
10167 "json-pretty" has been added to provide similar output, but
10168 neatly aligned for readability by humans.
10170 * We dropped all explicit sync() invocations in the shutdown
10171 code. The kernel does this implicitly anyway in the kernel
10172 reboot() syscall. halt(8)'s -n option is now a compatibility
10175 * We now support virtualized reboot() in containers, as
10176 supported by newer kernels. We will fall back to exit() if
10177 CAP_SYS_REBOOT is not available to the container. Also,
10178 nspawn makes use of this now and will actually reboot the
10179 container if the containerized OS asks for that.
10181 * journalctl will only show local log output by default
10182 now. Use --merge (-m) to show remote log output, too.
10184 * libsystemd-journal gained the new sd_journal_get_usage()
10185 call to determine the current disk usage of all journal
10186 files. This is exposed in the new "journalctl --disk-usage"
10189 * journald gained a new configuration setting SplitMode= in
10190 journald.conf which may be used to control how user journals
10191 are split off. See journald.conf(5) for details.
10193 * A new condition type ConditionFileNotEmpty= has been added.
10195 * tmpfiles' "w" lines now support file globbing, to write
10196 multiple files at once.
10198 * We added Python bindings for the journal submission
10199 APIs. More Python APIs for a number of selected APIs will
10200 likely follow. Note that we intend to add native bindings
10201 only for the Python language, as we consider it common
10202 enough to deserve bindings shipped within systemd. There are
10203 various projects outside of systemd that provide bindings
10204 for languages such as PHP or Lua.
10206 * Many conditions will now resolve specifiers such as %i. In
10207 addition, PathChanged= and related directives of .path units
10208 now support specifiers as well.
10210 * There's now a new RPM macro definition for the system preset
10213 * journald will now warn if it ca not forward a message to the
10214 syslog daemon because its socket is full.
10216 * timedated will no longer write or process /etc/timezone,
10217 except on Debian. As we do not support late mounted /usr
10218 anymore /etc/localtime always being a symlink is now safe,
10219 and hence the information in /etc/timezone is not necessary
10222 * logind will now always reserve one VT for a text getty (VT6
10223 by default). Previously if more than 6 X sessions where
10224 started they took up all the VTs with auto-spawned gettys,
10225 so that no text gettys were available anymore.
10227 * udev will now automatically inform the btrfs kernel logic
10228 about btrfs RAID components showing up. This should make
10229 simple hotplug based btrfs RAID assembly work.
10231 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
10232 (but not for its children which will stay at the kernel
10233 default). This should allow setups with a lot more listening
10236 * systemd will now always pass the configured timezone to the
10237 kernel at boot. timedated will do the same when the timezone
10240 * logind's inhibition logic has been updated. By default,
10241 logind will now handle the lid switch, the power and sleep
10242 keys all the time, even in graphical sessions. If DEs want
10243 to handle these events on their own they should take the new
10244 handle-power-key, handle-sleep-key and handle-lid-switch
10245 inhibitors during their runtime. A simple way to achieve
10246 that is to invoke the DE wrapped in an invocation of:
10248 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
10250 * Access to unit operations is now checked via SELinux taking
10251 the unit file label and client process label into account.
10253 * systemd will now notify the administrator in the journal
10254 when he over-mounts a non-empty directory.
10256 * There are new specifiers that are resolved in unit files,
10257 for the hostname (%H), the machine ID (%m) and the boot ID
10260 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
10261 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
10262 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
10263 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
10264 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
10265 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
10266 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
10270 * Support for reading structured kernel messages from
10271 /dev/kmsg has now been added and is enabled by default.
10273 * Support for reading kernel messages from /proc/kmsg has now
10274 been removed. If you want kernel messages in the journal
10275 make sure to run a recent kernel (>= 3.5) that supports
10276 reading structured messages from /dev/kmsg (see
10277 above). /proc/kmsg is now exclusive property of classic
10278 syslog daemons again.
10280 * The libudev API gained the new
10281 udev_device_new_from_device_id() call.
10283 * The logic for file system namespace (ReadOnlyDirectory=,
10284 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
10285 require pivot_root() anymore. This means fewer temporary
10286 directories are created below /tmp for this feature.
10288 * nspawn containers will now see and receive all submounts
10289 made on the host OS below the root file system of the
10292 * Forward Secure Sealing is now supported for Journal files,
10293 which provide cryptographical sealing of journal files so
10294 that attackers cannot alter log history anymore without this
10295 being detectable. Lennart will soon post a blog story about
10296 this explaining it in more detail.
10298 * There are two new service settings RestartPreventExitStatus=
10299 and SuccessExitStatus= which allow configuration of exit
10300 status (exit code or signal) which will be excepted from the
10301 restart logic, resp. consider successful.
10303 * journalctl gained the new --verify switch that can be used
10304 to check the integrity of the structure of journal files and
10305 (if Forward Secure Sealing is enabled) the contents of
10308 * nspawn containers will now be run with /dev/stdin, /dev/fd/
10309 and similar symlinks pre-created. This makes running shells
10310 as container init process a lot more fun.
10312 * The fstab support can now handle PARTUUID= and PARTLABEL=
10315 * A new ConditionHost= condition has been added to match
10316 against the hostname (with globs) and machine ID. This is
10317 useful for clusters where a single OS image is used to
10318 provision a large number of hosts which shall run slightly
10319 different sets of services.
10321 * Services which hit the restart limit will now be placed in a
10324 Contributions from: Bertram Poettering, Dave Reisner, Huang
10325 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
10326 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
10330 * When running in --user mode systemd will now become a
10331 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
10332 tree a lot more organized.
10334 * A new PartOf= unit dependency type has been introduced that
10335 may be used to group services in a natural way.
10337 * "systemctl enable" may now be used to enable instances of
10340 * journalctl now prints error log levels in red, and
10341 warning/notice log levels in bright white. It also supports
10342 filtering by log level now.
10344 * cgtop gained a new -n switch (similar to top), to configure
10345 the maximum number of iterations to run for. It also gained
10346 -b, to run in batch mode (accepting no input).
10348 * The suffix ".service" may now be omitted on most systemctl
10349 command lines involving service unit names.
10351 * There's a new bus call in logind to lock all sessions, as
10352 well as a loginctl verb for it "lock-sessions".
10354 * libsystemd-logind.so gained a new call sd_journal_perror()
10355 that works similar to libc perror() but logs to the journal
10356 and encodes structured information about the error number.
10358 * /etc/crypttab entries now understand the new keyfile-size=
10361 * shutdown(8) now can send a (configurable) wall message when
10362 a shutdown is cancelled.
10364 * The mount propagation mode for the root file system will now
10365 default to "shared", which is useful to make containers work
10366 nicely out-of-the-box so that they receive new mounts from
10367 the host. This can be undone locally by running "mount
10368 --make-rprivate /" if needed.
10370 * The prefdm.service file has been removed. Distributions
10371 should maintain this unit downstream if they intend to keep
10372 it around. However, we recommend writing normal unit files
10373 for display managers instead.
10375 * Since systemd is a crucial part of the OS we will now
10376 default to a number of compiler switches that improve
10377 security (hardening) such as read-only relocations, stack
10378 protection, and suchlike.
10380 * The TimeoutSec= setting for services is now split into
10381 TimeoutStartSec= and TimeoutStopSec= to allow configuration
10382 of individual time outs for the start and the stop phase of
10385 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
10386 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
10387 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
10388 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
10389 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
10390 Gundersen, Zbigniew Jędrzejewski-Szmek
10394 * The journal and id128 C APIs are now fully documented as man
10397 * Extra safety checks have been added when transitioning from
10398 the initial RAM disk to the main system to avoid accidental
10401 * /etc/crypttab entries now understand the new keyfile-offset=
10404 * systemctl -t can now be used to filter by unit load state.
10406 * The journal C API gained the new sd_journal_wait() call to
10407 make writing synchronous journal clients easier.
10409 * journalctl gained the new -D switch to show journals from a
10410 specific directory.
10412 * journalctl now displays a special marker between log
10413 messages of two different boots.
10415 * The journal is now explicitly flushed to /var via a service
10416 systemd-journal-flush.service, rather than implicitly simply
10417 by seeing /var/log/journal to be writable.
10419 * journalctl (and the journal C APIs) can now match for much
10420 more complex expressions, with alternatives and
10423 * When transitioning from the initial RAM disk to the main
10424 system we will now kill all processes in a killing spree to
10425 ensure no processes stay around by accident.
10427 * Three new specifiers may be used in unit files: %u, %h, %s
10428 resolve to the user name, user home directory resp. user
10429 shell. This is useful for running systemd user instances.
10431 * We now automatically rotate journal files if their data
10432 object hash table gets a fill level > 75%. We also size the
10433 hash table based on the configured maximum file size. This
10434 together should lower hash collisions drastically and thus
10435 speed things up a bit.
10437 * journalctl gained the new "--header" switch to introspect
10438 header data of journal files.
10440 * A new setting SystemCallFilters= has been added to services which may
10441 be used to apply deny lists or allow lists to system calls. This is
10442 based on SECCOMP Mode 2 of Linux 3.5.
10444 * nspawn gained a new --link-journal= switch (and quicker: -j)
10445 to link the container journal with the host. This makes it
10446 very easy to centralize log viewing on the host for all
10447 guests while still keeping the journal files separated.
10449 * Many bugfixes and optimizations
10451 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
10452 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
10453 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
10458 * Several tools now understand kernel command line arguments,
10459 which are only read when run in an initial RAM disk. They
10460 usually follow closely their normal counterparts, but are
10463 * There's a new tool to analyze the readahead files that are
10464 automatically generated at boot. Use:
10466 /usr/lib/systemd/systemd-readahead analyze /.readahead
10468 * We now provide an early debug shell on tty9 if this enabled. Use:
10470 systemctl enable debug-shell.service
10472 * All plymouth related units have been moved into the Plymouth
10473 package. Please make sure to upgrade your Plymouth version
10476 * systemd-tmpfiles now supports getting passed the basename of
10477 a configuration file only, in which case it will look for it
10478 in all appropriate directories automatically.
10480 * udevadm info now takes a /dev or /sys path as argument, and
10481 does the right thing. Example:
10483 udevadm info /dev/sda
10484 udevadm info /sys/class/block/sda
10486 * systemctl now prints a warning if a unit is stopped but a
10487 unit that might trigger it continues to run. Example: a
10488 service is stopped but the socket that activates it is left
10491 * "systemctl status" will now mention if the log output was
10492 shortened due to rotation since a service has been started.
10494 * The journal API now exposes functions to determine the
10495 "cutoff" times due to rotation.
10497 * journald now understands SIGUSR1 and SIGUSR2 for triggering
10498 immediately flushing of runtime logs to /var if possible,
10499 resp. for triggering immediate rotation of the journal
10502 * It is now considered an error if a service is attempted to
10503 be stopped that is not loaded.
10505 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
10507 * systemd-analyze now supports Python 3
10509 * tmpfiles now supports cleaning up directories via aging
10510 where the first level dirs are always kept around but
10511 directories beneath it automatically aged. This is enabled
10512 by prefixing the age field with '~'.
10514 * Seat objects now expose CanGraphical, CanTTY properties
10515 which is required to deal with very fast bootups where the
10516 display manager might be running before the graphics drivers
10517 completed initialization.
10519 * Seat objects now expose a State property.
10521 * We now include RPM macros for service enabling/disabling
10522 based on the preset logic. We recommend RPM based
10523 distributions to make use of these macros if possible. This
10524 makes it simpler to reuse RPM spec files across
10527 * We now make sure that the collected systemd unit name is
10528 always valid when services log to the journal via
10531 * There's a new man page kernel-command-line(7) detailing all
10532 command line options we understand.
10534 * The fstab generator may now be disabled at boot by passing
10535 fstab=0 on the kernel command line.
10537 * A new kernel command line option modules-load= is now understood
10538 to load a specific kernel module statically, early at boot.
10540 * Unit names specified on the systemctl command line are now
10541 automatically escaped as needed. Also, if file system or
10542 device paths are specified they are automatically turned
10543 into the appropriate mount or device unit names. Example:
10545 systemctl status /home
10546 systemctl status /dev/sda
10548 * The SysVConsole= configuration option has been removed from
10549 system.conf parsing.
10551 * The SysV search path is no longer exported on the D-Bus
10554 * The Names= option has been removed from unit file parsing.
10556 * There's a new man page bootup(7) detailing the boot process.
10558 * Every unit and every generator we ship with systemd now
10559 comes with full documentation. The self-explanatory boot is
10562 * A couple of services gained "systemd-" prefixes in their
10563 name if they wrap systemd code, rather than only external
10564 code. Among them fsck@.service which is now
10565 systemd-fsck@.service.
10567 * The HaveWatchdog property has been removed from the D-Bus
10570 * systemd.confirm_spawn= on the kernel command line should now
10573 * There's a new man page crypttab(5) which details all options
10574 we actually understand.
10576 * systemd-nspawn gained a new --capability= switch to pass
10577 additional capabilities to the container.
10579 * timedated will now read known NTP implementation unit names
10580 from /usr/lib/systemd/ntp-units.d/*.list,
10581 systemd-timedated-ntp.target has been removed.
10583 * journalctl gained a new switch "-b" that lists log data of
10584 the current boot only.
10586 * The notify socket is in the abstract namespace again, in
10587 order to support daemons which chroot() at start-up.
10589 * There is a new Storage= configuration option for journald
10590 which allows configuration of where log data should go. This
10591 also provides a way to disable journal logging entirely, so
10592 that data collected is only forwarded to the console, the
10593 kernel log buffer or another syslog implementation.
10595 * Many bugfixes and optimizations
10597 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
10598 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
10599 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
10600 Shawn Landden, Tom Gundersen
10604 * "systemctl help <unit>" now shows the man page if one is
10607 * Several new man pages have been added.
10609 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
10610 MaxLevelConsole= can now be specified in
10611 journald.conf. These options allow reducing the amount of
10612 data stored on disk or forwarded by the log level.
10614 * TimerSlackNSec= can now be specified in system.conf for
10615 PID1. This allows system-wide power savings.
10617 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
10618 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
10623 * logind is now capable of (optionally) handling power and
10624 sleep keys as well as the lid switch.
10626 * journalctl now understands the syntax "journalctl
10627 /usr/bin/avahi-daemon" to get all log output of a specific
10630 * CapabilityBoundingSet= in system.conf now also influences
10631 the capability bound set of usermode helpers of the kernel.
10633 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
10634 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
10635 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
10636 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
10640 * Note that we skipped 139 releases here in order to set the
10641 new version to something that is greater than both udev's
10642 and systemd's most recent version number.
10644 * udev: all udev sources are merged into the systemd source tree now.
10645 All future udev development will happen in the systemd tree. It
10646 is still fully supported to use the udev daemon and tools without
10647 systemd running, like in initramfs or other init systems. Building
10648 udev though, will require the *build* of the systemd tree, but
10649 udev can be properly *run* without systemd.
10651 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
10652 should be used to create dead device nodes as workarounds for broken
10655 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
10656 no longer supported. udev_monitor_new_from_netlink() needs to be
10657 used to subscribe to events.
10659 * udev: when udevd is started by systemd, processes which are left
10660 behind by forking them off of udev rules, are unconditionally cleaned
10661 up and killed now after the event handling has finished. Services or
10662 daemons must be started as systemd services. Services can be
10663 pulled-in by udev to get started, but they can no longer be directly
10664 forked by udev rules.
10666 * udev: the daemon binary is called systemd-udevd now and installed
10667 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
10668 to adapt to that, create symlink, or rename the binary after building
10671 * libudev no longer provides these symbols:
10672 udev_monitor_from_socket()
10673 udev_queue_get_failed_list_entry()
10674 udev_get_{dev,sys,run}_path()
10675 The versions number was bumped and symbol versioning introduced.
10677 * systemd-loginctl and systemd-journalctl have been renamed
10678 to loginctl and journalctl to match systemctl.
10680 * The config files: /etc/systemd/systemd-logind.conf and
10681 /etc/systemd/systemd-journald.conf have been renamed to
10682 logind.conf and journald.conf. Package updates should rename
10683 the files to the new names on upgrade.
10685 * For almost all files the license is now LGPL2.1+, changed
10686 from the previous GPL2.0+. Exceptions are some minor stuff
10687 of udev (which will be changed to LGPL2.1 eventually, too),
10688 and the MIT licensed sd-daemon.[ch] library that is suitable
10689 to be used as drop-in files.
10691 * systemd and logind now handle system sleep states, in
10692 particular suspending and hibernating.
10694 * logind now implements a sleep/shutdown/idle inhibiting logic
10695 suitable for a variety of uses. Soonishly Lennart will blog
10696 about this in more detail.
10698 * var-run.mount and var-lock.mount are no longer provided
10699 (which previously bind mounted these directories to their new
10700 places). Distributions which have not converted these
10701 directories to symlinks should consider stealing these files
10702 from git history and add them downstream.
10704 * We introduced the Documentation= field for units and added
10705 this to all our shipped units. This is useful to make it
10706 easier to explore the boot and the purpose of the various
10709 * All smaller setup units (such as
10710 systemd-vconsole-setup.service) now detect properly if they
10711 are run in a container and are skipped when
10712 appropriate. This guarantees an entirely noise-free boot in
10713 Linux container environments such as systemd-nspawn.
10715 * A framework for implementing offline system updates is now
10716 integrated, for details see:
10717 https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates
10719 * A new service type Type=idle is available now which helps us
10720 avoiding ugly interleaving of getty output and boot status
10723 * There's now a system-wide CapabilityBoundingSet= option to
10724 globally reduce the set of capabilities for the
10725 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
10726 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
10727 even CAP_NET_ADMIN system-wide for secure systems.
10729 * There are now system-wide DefaultLimitXXX= options to
10730 globally change the defaults of the various resource limits
10731 for all units started by PID 1.
10733 * Harald Hoyer's systemd test suite has been integrated into
10734 systemd which allows easy testing of systemd builds in qemu
10735 and nspawn. (This is really awesome! Ask us for details!)
10737 * The fstab parser is now implemented as generator, not inside
10740 * systemctl will now warn you if .mount units generated from
10741 /etc/fstab are out of date due to changes in fstab that
10742 have not been read by systemd yet.
10744 * systemd is now suitable for usage in initrds. Dracut has
10745 already been updated to make use of this. With this in place
10746 initrds get a slight bit faster but primarily are much
10747 easier to introspect and debug since "systemctl status" in
10748 the host system can be used to introspect initrd services,
10749 and the journal from the initrd is kept around too.
10751 * systemd-delta has been added, a tool to explore differences
10752 between user/admin configuration and vendor defaults.
10754 * PrivateTmp= now affects both /tmp and /var/tmp.
10756 * Boot time status messages are now much prettier and feature
10757 proper english language. Booting up systemd has never been
10760 * Read-ahead pack files now include the inode number of all
10761 files to pre-cache. When the inode changes the pre-caching
10762 is not attempted. This should be nicer to deal with updated
10763 packages which might result in changes of read-ahead
10766 * We now temporaritly lower the kernel's read_ahead_kb variable
10767 when collecting read-ahead data to ensure the kernel's
10768 built-in read-ahead does not add noise to our measurements
10769 of necessary blocks to pre-cache.
10771 * There's now RequiresMountsFor= to add automatic dependencies
10772 for all mounts necessary for a specific file system path.
10774 * MountAuto= and SwapAuto= have been removed from
10775 system.conf. Mounting file systems at boot has to take place
10778 * nspawn now learned a new switch --uuid= to set the machine
10779 ID on the command line.
10781 * nspawn now learned the -b switch to automatically search
10782 for an init system.
10784 * vt102 is now the default TERM for serial TTYs, upgraded from
10787 * systemd-logind now works on VT-less systems.
10789 * The build tree has been reorganized. The individual
10790 components now have directories of their own.
10792 * A new condition type ConditionPathIsReadWrite= is now available.
10794 * nspawn learned the new -C switch to create cgroups for the
10795 container in other hierarchies.
10797 * We now have support for hardware watchdogs, configurable in
10800 * The scheduled shutdown logic now has a public API.
10802 * We now mount /tmp as tmpfs by default, but this can be
10803 masked and /etc/fstab can override it.
10805 * Since udisks does not make use of /media anymore we are not
10806 mounting a tmpfs on it anymore.
10808 * journalctl gained a new --local switch to only interleave
10809 locally generated journal files.
10811 * We can now load the IMA policy at boot automatically.
10813 * The GTK tools have been split off into a systemd-ui.
10815 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
10816 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
10817 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
10818 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
10819 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
10820 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
10821 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
10822 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
10823 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
10828 * This is mostly a bugfix release
10830 * Support optional initialization of the machine ID from the
10831 KVM or container configured UUID.
10833 * Support immediate reboots with "systemctl reboot -ff"
10835 * Show /etc/os-release data in systemd-analyze output
10837 * Many bugfixes for the journal, including endianness fixes and
10838 ensuring that disk space enforcement works
10840 * sd-login.h is C++ compatible again
10842 * Extend the /etc/os-release format on request of the Debian
10845 * We now refuse non-UTF8 strings used in various configuration
10846 and unit files. This is done to ensure we do not pass invalid
10847 data over D-Bus or expose it elsewhere.
10849 * Register Mimo USB Screens as suitable for automatic seat
10852 * Read SELinux client context from journal clients in a race
10855 * Reorder configuration file lookup order. /etc now always
10856 overrides /run in order to allow the administrator to always
10857 and unconditionally override vendor-supplied or
10858 automatically generated data.
10860 * The various user visible bits of the journal now have man
10861 pages. We still lack man pages for the journal API calls
10864 * We now ship all man pages in HTML format again in the
10867 Contributions from: Dave Reisner, Dirk Eibach, Frederic
10868 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
10869 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
10874 * This is mostly a bugfix release
10876 * systems lacking /etc/os-release are no longer supported.
10878 * Various functionality updates to libsystemd-login.so
10880 * Track class of PAM logins to distinguish greeters from
10881 normal user logins.
10883 Contributions from: Kay Sievers, Lennart Poettering, Michael
10888 * This is an important bugfix release for v41.
10890 * Building man pages is now optional which should be useful
10891 for those building systemd from git but unwilling to install
10894 * Watchdog support for supervising services is now usable. In
10895 a future release support for hardware watchdogs
10896 (i.e. /dev/watchdog) will be added building on this.
10898 * Service start rate limiting is now configurable and can be
10899 turned off per service. When a start rate limit is hit a
10900 reboot can automatically be triggered.
10902 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
10904 Contributions from: Benjamin Franzke, Bill Nottingham,
10905 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
10906 Schmidt, Michał Górny, Piotr Drąg
10910 * The systemd binary is installed /usr/lib/systemd/systemd now;
10911 An existing /sbin/init symlink needs to be adapted with the
10914 * The code that loads kernel modules has been ported to invoke
10915 libkmod directly, instead of modprobe. This means we do not
10916 support systems with module-init-tools anymore.
10918 * Watchdog support is now already useful, but still not
10921 * A new kernel command line option systemd.setenv= is
10922 understood to set system wide environment variables
10923 dynamically at boot.
10925 * We now limit the set of capabilities of systemd-journald.
10927 * We now set SIGPIPE to ignore by default, since it only is
10928 useful in shell pipelines, and has little use in general
10929 code. This can be disabled with IgnoreSIPIPE=no in unit
10932 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
10933 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
10938 * This is mostly a bugfix release
10940 * We now expose the reason why a service failed in the
10941 "Result" D-Bus property.
10943 * Rudimentary service watchdog support (will be completed over
10944 the next few releases.)
10946 * When systemd forks off in order execute some service we will
10947 now immediately changes its argv[0] to reflect which process
10948 it will execute. This is useful to minimize the time window
10949 with a generic argv[0], which makes bootcharts more useful
10951 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
10952 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
10953 Mike Kazantsev, Ray Strode
10957 * This is mostly a test release, but incorporates many
10960 * New systemd-cgtop tool to show control groups by their
10963 * Linking against libacl for ACLs is optional again. If
10964 disabled, support tracking device access for active logins
10965 goes becomes unavailable, and so does access to the user
10966 journals by the respective users.
10968 * If a group "adm" exists, journal files are automatically
10969 owned by them, thus allow members of this group full access
10970 to the system journal as well as all user journals.
10972 * The journal now stores the SELinux context of the logging
10973 client for all entries.
10975 * Add C++ inclusion guards to all public headers
10977 * New output mode "cat" in the journal to print only text
10978 messages, without any meta data like date or time.
10980 * Include tiny X server wrapper as a temporary stop-gap to
10981 teach XOrg udev display enumeration. This is used by display
10982 managers such as gdm, and will go away as soon as XOrg
10983 learned native udev hotplugging for display devices.
10985 * Add new systemd-cat tool for executing arbitrary programs
10986 with STDERR/STDOUT connected to the journal. Can also act as
10987 BSD logger replacement, and does so by default.
10989 * Optionally store all locally generated coredumps in the
10990 journal along with meta data.
10992 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
10993 writing short strings to files (for usage for /sys), and for
10994 creating symlinks, character and block device nodes.
10996 * New unit file option ControlGroupPersistent= to make cgroups
10997 persistent, following the mechanisms outlined in
10998 https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
11000 * Support multiple local RTCs in a sane way
11002 * No longer monopolize IO when replaying readahead data on
11003 rotating disks, since we might starve non-file-system IO to
11004 death, since fanotify() will not see accesses done by blkid,
11007 * Do not show kernel threads in systemd-cgls anymore, unless
11008 requested with new -k switch.
11010 Contributions from: Dan Horák, Kay Sievers, Lennart
11011 Poettering, Michal Schmidt
11015 * This is mostly a test release, but incorporates many
11018 * The git repository moved to:
11019 git://anongit.freedesktop.org/systemd/systemd
11020 ssh://git.freedesktop.org/git/systemd/systemd
11022 * First release with the journal
11023 http://0pointer.de/blog/projects/the-journal.html
11025 * The journal replaces both systemd-kmsg-syslogd and
11026 systemd-stdout-bridge.
11028 * New sd_pid_get_unit() API call in libsystemd-logind
11030 * Many systemadm clean-ups
11032 * Introduce remote-fs-pre.target which is ordered before all
11033 remote mounts and may be used to start services before all
11036 * Added Mageia support
11038 * Add bash completion for systemd-loginctl
11040 * Actively monitor PID file creation for daemons which exit in
11041 the parent process before having finished writing the PID
11042 file in the daemon process. Daemons which do this need to be
11043 fixed (i.e. PID file creation must have finished before the
11044 parent exits), but we now react a bit more gracefully to them.
11046 * Add colourful boot output, mimicking the well-known output
11047 of existing distributions.
11049 * New option PassCredentials= for socket units, for
11050 compatibility with a recent kernel ABI breakage.
11052 * /etc/rc.local is now hooked in via a generator binary, and
11053 thus will no longer act as synchronization point during
11056 * systemctl list-unit-files now supports --root=.
11058 * systemd-tmpfiles now understands two new commands: z, Z for
11059 relabelling files according to the SELinux database. This is
11060 useful to apply SELinux labels to specific files in /sys,
11061 among other things.
11063 * Output of SysV services is now forwarded to both the console
11064 and the journal by default, not only just the console.
11066 * New man pages for all APIs from libsystemd-login.
11068 * The build tree got reorganized and the build system is a
11069 lot more modular allowing embedded setups to specifically
11070 select the components of systemd they are interested in.
11072 * Support for Linux systems lacking the kernel VT subsystem is
11075 * configure's --with-rootdir= got renamed to
11076 --with-rootprefix= to follow the naming used by udev and
11079 * Unless specified otherwise we will now install to /usr instead
11080 of /usr/local by default.
11082 * Processes with '@' in argv[0][0] are now excluded from the
11083 final shut-down killing spree, following the logic explained
11085 https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
11087 * All processes remaining in a service cgroup when we enter
11088 the START or START_PRE states are now killed with
11089 SIGKILL. That means it is no longer possible to spawn
11090 background processes from ExecStart= lines (which was never
11091 supported anyway, and bad style).
11093 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
11094 reloading of units together.
11096 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
11097 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
11098 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
11099 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
11100 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek