1 systemd System and Service Manager
5 * journalctl gained the new options -t/--identifier= to match
6 on the syslog identifier (aka "tag"), as well as --utc to
7 show log timestamps in the UTC timezone. journalctl now also
8 accepts -n/--lines=all to disable line capping in a pager.
10 * Services can notify the manager before they start a reload
11 (by sending RELOADING=1) or shutdown (by sending
12 STOPPING=1). This allows the manager to track and show the
13 internal state of daemons and closes a race condition when
14 the process is still running but has closed its D-Bus
17 * Services with Type=oneshot do not have to have any ExecStart
20 * User units are now loaded also from
21 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
22 /run/systemd/user directory that was already previously
23 supported, but is under the control of the user.
25 * Job timeouts (i.e. time-outs on the time a job that is
26 queued stays in the run queue) can now optionally result in
27 immediate reboot or power-off actions (JobTimeoutAction= and
28 JobTimeoutRebootArgument=). This is useful on ".target"
29 units, to limit the maximum time a target remains
30 undispatched in the run queue, and to trigger an emergency
31 operation in such a case. This is now used by default to
32 turn off the system if boot-up (as defined by everything in
33 basic.target) hangs and does not complete for at least
34 15min. Also, if power-off or reboot hang for at least 30min
35 an immediate power-off/reboot operation is triggered. This
36 functionality is particularly useful to increase reliability
37 on embedded devices, but also on laptops which might
38 accidentally get powered on when carried in a backpack and
39 whose boot stays stuck in a hard disk encryption passphrase
42 * systemd-logind can be configured to also handle lid switch
43 events even when the machine is docked or multiple displays
44 are attached (HandleLidSwitchDocked= option).
46 * A helper binary and a service have been added which can be
47 used to resume from hibernation in the initramfs. A
48 generator will parse the resume= option on the kernel
49 command-line to trigger resume.
51 * A user console daemon systemd-consoled has been
52 added. Currently, it is a preview, and will so far open a
53 single terminal on each session of the user marked as
54 Desktop=systemd-console.
56 * Route metrics can be specified for DHCP routes added by
59 * The SELinux context of socket-actived services can be set
60 from the information provided by the networking stack
61 (SELinuxContextFromNet= option).
63 * Userspace firmware loading support has been removed and
64 the minimum supported kernel version is thus bumped to 3.7.
66 * Timeout for udev workers has been increased from 1 to 3
67 minutes, but a warning will be printed after 1 minute to
68 help diagnose kernel modules that take a long time to load.
70 * Udev rules can now remove tags on devices with TAG-="foobar".
72 * systemd's readahead implementation has been removed. In many
73 circumstatances it didn't give expected benefits even for
74 rotational disk drives and was becoming less relevant in the
75 age of SSDs. As none of the developers has been using
76 rotating media anymore, and nobody stepped up to actively
77 maintain this component of systemd it has now been removed.
79 * Swap units can use Discard= to specify discard options.
80 Discard options specified for swaps in /etc/fstab are now
83 * Docker containers are now detected as a separate type of
86 * The Password Agent protocol gained support for queries where
87 the user input is shown, useful e.g. for usernames.
88 systemd-ask-password gained a new --echo option to turn that
91 * The default sysctl.d/ snippets will now set:
93 net.core.default_qdisc = fq_codel
95 This selects Fair Queueing Controlled Delay as the default
96 queueing discipline for network interfaces. fq_codel helps
97 fight the network bufferbloat problem. It is believed to be
98 a good default with no tuning required for most workloads.
99 Downstream distributions may override this choice. On 10Gbit
100 servers that do not do forwarding, "fq" may perform better.
101 Systems without a good clocksource should use "pfifo_fast".
103 * If kdbus is enabled during build a new option BusPolicy= is
104 available for service units, that allows locking all service
105 processes into a stricter bus policy, in order to limit
106 access to various bus services, or even hide most of them
107 from the service's view entirely.
109 * networkctl will now show the .network and .link file
110 networkd has applied to a specific interface.
112 * sd-login gained a new API call sd_session_get_desktop() to
113 query which desktop environment has been selected for a
116 * UNIX utmp support is now compile-time optional to support
119 * systemctl gained two new commands "add-wants" and
120 "add-requires" for pulling in units from specific targets
123 * If the word "rescue" is specified on the kernel command line
124 the system will now boot into rescue mode (aka
125 rescue.target), which was previously available only by
126 specifying "1" or "systemd.unit=rescue.target" on the kernel
127 command line. This new kernel command line option nicely
128 mirrors the already existing "emergency" kernel command line
131 * New kernel command line options mount.usr=, mount.usrflags=,
132 usrfstype= have been added that match root=, rootflags=,
133 rootfstype= but allow mounting a specific file system to
136 * The $NOTIFY_SOCKET is now also passed to control processesof
137 services, not only the main process.
139 * This version reenables support for fsck's -l switch. This
140 means at least version v2.25 of util-linux is required for
141 operation, otherwise dead-locks on device nodes may
142 occur. Again: you need to update util-linux to at least
143 v2.25 when updating systemd to v217.
145 * The "multi-seat-x" tool has been removed from systemd, as
146 its functionality has been integrated into X servers 1.16,
147 and the tool is hence redundant. It is recommended to update
148 display managers invoking this tool to simply invoke X
149 directly from now on, again.
151 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
152 message flag has been added for all of systemd's PolicyKit
153 authenticated method calls has been added. In particular
154 this now allows optional interactive authorization via
155 PolicyKit for many of PID1's priviliged operations such as
156 unit file enabling and disabling.
158 * "udevadm hwdb --update" learnt a new switch "--usr" for
159 placing the rebuilt hardware database in /usr instead of
160 /etc. When used only hardware database entries stored in
161 /usr will be used, and any user database entries in /etc are
162 ignored. This functionality is useful for vendors to ship a
163 pre-built database on systems where local configuration is
164 unnecessary or unlikely.
166 * Calendar time specifications in .timer units now also
167 understand the strings "semi-annually", "quarterly" and
168 "minutely" as shortcuts (in addition to the pre-existing
169 "anually", "hourly", ...).
171 Contributions from: Andreas Henriksson, Andrei Borzenkov,
172 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
173 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
174 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
175 Herrmann, David Sommerseth, David Strauss, Emil Renner
176 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
177 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
178 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
179 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
180 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
181 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
182 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
183 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
184 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
185 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
186 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
187 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
188 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
189 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
190 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
191 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
194 -- Berlin, 2014-10-28
198 * timedated no longer reads NTP implementation unit names from
199 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
200 implementations should add a
202 Conflicts=systemd-timesyncd.service
204 to their unit files to take over and replace systemd's NTP
205 default functionality.
207 * systemd-sysusers gained a new line type "r" for configuring
208 which UID/GID ranges to allocate system users/groups
209 from. Lines of type "u" may now add an additional column
210 that specifies the home directory for the system user to be
211 created. Also, systemd-sysusers may now optionally read user
212 information from STDIN instead of a file. This is useful for
213 invoking it from RPM preinst scriptlets that need to create
214 users before the first RPM file is installed since these
215 files might need to be owned by them. A new
216 %sysusers_create_inline RPM macro has been introduced to do
217 just that. systemd-sysusers now updates the shadow files as
218 well as the user/group databases, which should enhance
219 compatibility with certain tools like grpck.
221 * A number of bus APIs of PID 1 now optionally consult
222 PolicyKit to permit access for otherwise unprivileged
223 clients under certain conditions. Note that this currently
224 doesn't support interactive authentication yet, but this is
225 expected to be added eventually, too.
227 * /etc/machine-info now has new fields for configuring the
228 deployment environment of the machine, as well as the
229 location of the machine. hostnamectl has been updated with
230 new command to update these fields.
232 * systemd-timesyncd has been updated to automatically acquire
233 NTP server information from systemd-networkd, which might
234 have been discovered via DHCP.
236 * systemd-resolved now includes a caching DNS stub resolver
237 and a complete LLMNR name resolution implementation. A new
238 NSS module "nss-resolve" has been added which can be used
239 instead of glibc's own "nss-dns" to resolve hostnames via
240 systemd-resolved. Hostnames, addresses and arbitrary RRs may
241 be resolved via systemd-resolved D-Bus APIs. In contrast to
242 the glibc internal resolver systemd-resolved is aware of
243 multi-homed system, and keeps DNS server and caches separate
244 and per-interface. Queries are sent simultaneously on all
245 interfaces that have DNS servers configured, in order to
246 properly handle VPNs and local LANs which might resolve
247 separate sets of domain names. systemd-resolved may acquire
248 DNS server information from systemd-networkd automatically,
249 which in turn might have discovered them via DHCP. A tool
250 "systemd-resolve-host" has been added that may be used to
251 query the DNS logic in resolved. systemd-resolved implements
252 IDNA and automatically uses IDNA or UTF-8 encoding depending
253 on whether classic DNS or LLMNR is used as transport. In the
254 next releases we intend to add a DNSSEC and mDNS/DNS-SD
255 implementation to systemd-resolved.
257 * A new NSS module nss-mymachines has been added, that
258 automatically resolves the names of all local registered
259 containers to their respective IP addresses.
261 * A new client tool "networkctl" for systemd-networkd has been
262 added. It currently is entirely passive and will query
263 networking configuration from udev, rtnetlink and networkd,
264 and present it to the user in a very friendly
265 way. Eventually, we hope to extend it to become a full
266 control utility for networkd.
268 * .socket units gained a new DeferAcceptSec= setting that
269 controls the kernels' TCP_DEFER_ACCEPT sockopt for
270 TCP. Similar, support for controlling TCP keep-alive
271 settings has been added (KeepAliveTimeSec=,
272 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
273 turning off Nagle's algorithm on TCP has been added
276 * logind learned a new session type "web", for use in projects
277 like Cockpit which register web clients as PAM sessions.
279 * timer units with at least one OnCalendar= setting will now
280 be started only after timer-sync.target has been
281 reached. This way they will not elapse before the system
282 clock has been corrected by a local NTP client or
283 similar. This is particular useful on RTC-less embedded
284 machines, that come up with an invalid system clock.
286 * systemd-nspawn's --network-veth= switch should now result in
287 stable MAC addresses for both the outer and the inner side
290 * systemd-nspawn gained a new --volatile= switch for running
291 container instances with /etc or /var unpopulated.
293 * The kdbus client code has been updated to use the new Linux
294 3.17 memfd subsystem instead of the old kdbus-specific one.
296 * systemd-networkd's DHCP client and server now support
297 FORCERENEW. There are also new configuration options to
298 configure the vendor client identifier and broadcast mode
301 * systemd will no longer inform the kernel about the current
302 timezone, as this is necessarily incorrect and racy as the
303 kernel has no understanding of DST and similar
304 concepts. This hence means FAT timestamps will be always
305 considered UTC, similar to what Android is already
306 doing. Also, when the RTC is configured to the local time
307 (rather than UTC) systemd will never synchronize back to it,
308 as this might confuse Windows at a later boot.
310 * systemd-analyze gained a new command "verify" for offline
311 validation of unit files.
313 * systemd-networkd gained support for a couple of additional
314 settings for bonding networking setups. Also, the metric for
315 statically configured routes may now be configured. For
316 network interfaces where this is appropriate the peer IP
317 address may now be configured.
319 * systemd-networkd's DHCP client will no longer request
320 broadcasting by default, as this tripped up some networks.
321 For hardware where broadcast is required the feature should
322 be switched back on using RequestBroadcast=yes.
324 * systemd-networkd will now set up IPv4LL addresses (when
325 enabled) even if DHCP is configured successfully.
327 * udev will now default to respect network device names given
328 by the kernel when the kernel indicates that these are
329 predictable. This behavior can be tweaked by changing
330 NamePolicy= in the relevant .link file.
332 * A new library systemd-terminal has been added that
333 implements full TTY stream parsing and rendering. This
334 library is supposed to be used later on for implementing a
335 full userspace VT subsystem, replacing the current kernel
338 * A new tool systemd-journal-upload has been added to push
339 journal data to a remote system running
340 systemd-journal-remote.
342 * journald will no longer forward all local data to another
343 running syslog daemon. This change has been made because
344 rsyslog (which appears to be the most commonly used syslog
345 implementation these days) no longer makes use of this, and
346 instead pulls the data out of the journal on its own. Since
347 forwarding the messages to a non-existent syslog server is
348 more expensive than we assumed we have now turned this
349 off. If you run a syslog server that is not a recent rsyslog
350 version, you have to turn this option on again
351 (ForwardToSyslog= in journald.conf).
353 * journald now optionally supports the LZ4 compressor for
354 larger journal fields. This compressor should perform much
355 better than XZ which was the previous default.
357 * machinectl now shows the IP addresses of local containers,
358 if it knows them, plus the interface name of the container.
360 * A new tool "systemd-escape" has been added that makes it
361 easy to escape strings to build unit names and similar.
363 * sd_notify() messages may now include a new ERRNO= field
364 which is parsed and collected by systemd and shown among the
365 "systemctl status" output for a service.
367 * A new component "systemd-firstboot" has been added that
368 queries the most basic systemd information (timezone,
369 hostname, root password) interactively on first
370 boot. Alternatively it may also be used to provision these
371 things offline on OS images installed into directories.
373 * The default sysctl.d/ snippets will now set
375 net.ipv4.conf.default.promote_secondaries=1
377 This has the benefit of no flushing secondary IP addresses
378 when primary addresses are removed.
380 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
381 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
382 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
383 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
384 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
385 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
386 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
387 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
388 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
389 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
390 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
391 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
392 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
393 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
394 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
396 -- Berlin, 2014-08-19
400 * A new tool systemd-sysusers has been added. This tool
401 creates system users and groups in /etc/passwd and
402 /etc/group, based on static declarative system user/group
403 definitions in /usr/lib/sysusers.d/. This is useful to
404 enable factory resets and volatile systems that boot up with
405 an empty /etc directory, and thus need system users and
406 groups created during early boot. systemd now also ships
407 with two default sysusers.d/ files for the most basic
408 users and groups systemd and the core operating system
411 * A new tmpfiles snippet has been added that rebuilds the
412 essential files in /etc on boot, should they be missing.
414 * A directive for ensuring automatic clean-up of
415 /var/cache/man/ has been removed from the default
416 configuration. This line should now be shipped by the man
417 implementation. The necessary change has been made to the
418 man-db implementation. Note that you need to update your man
419 implementation to one that ships this line, otherwise no
420 automatic clean-up of /var/cache/man will take place.
422 * A new condition ConditionNeedsUpdate= has been added that
423 may conditionalize services to only run when /etc or /var
424 are "older" than the vendor operating system resources in
425 /usr. This is useful for reconstructing or updating /etc
426 after an offline update of /usr or a factory reset, on the
427 next reboot. Services that want to run once after such an
428 update or reset should use this condition and order
429 themselves before the new systemd-update-done.service, which
430 will mark the two directories as fully updated. A number of
431 service files have been added making use of this, to rebuild
432 the udev hardware database, the journald message catalog and
433 dynamic loader cache (ldconfig). The systemd-sysusers tool
434 described above also makes use of this now. With this in
435 place it is now possible to start up a minimal operating
436 system with /etc empty cleanly. For more information on the
437 concepts involved see this recent blog story:
439 http://0pointer.de/blog/projects/stateless.html
441 * A new system group "input" has been introduced, and all
442 input device nodes get this group assigned. This is useful
443 for system-level software to get access to input devices. It
444 complements what is already done for "audio" and "video".
446 * systemd-networkd learnt minimal DHCPv4 server support in
447 addition to the existing DHCPv4 client support. It also
448 learnt DHCPv6 client and IPv6 Router Solicitation client
449 support. The DHCPv4 client gained support for static routes
450 passed in from the server. Note that the [DHCPv4] section
451 known in older systemd-networkd versions has been renamed to
452 [DHCP] and is now also used by the DHCPv6 client. Existing
453 .network files using settings of this section should be
454 updated, though compatibility is maintained. Optionally, the
455 client hostname may now be sent to the DHCP server.
457 * networkd gained support for vxlan virtual networks as well
458 as tun/tap and dummy devices.
460 * networkd gained support for automatic allocation of address
461 ranges for interfaces from a system-wide pool of
462 addresses. This is useful for dynamically managing a large
463 number of interfaces with a single network configuration
464 file. In particular this is useful to easily assign
465 appropriate IP addresses to the veth links of a large number
468 * RPM macros for processing sysusers, sysctl and binfmt
469 drop-in snippets at package installation time have been
472 * The /etc/os-release file should now be placed in
473 /usr/lib/os-release. The old location is automatically
474 created as symlink. /usr/lib is the more appropriate
475 location of this file, since it shall actually describe the
476 vendor operating system shipped in /usr, and not the
477 configuration stored in /etc.
479 * .mount units gained a new boolean SloppyOptions= setting
480 that maps to mount(8)'s -s option which enables permissive
481 parsing of unknown mount options.
483 * tmpfiles learnt a new "L+" directive which creates a symlink
484 but (unlike "L") deletes a pre-existing file first, should
485 it already exist and not already be the correct
486 symlink. Similar, "b+", "c+" and "p+" directives have been
487 added as well, which create block and character devices, as
488 well as fifos in the filesystem, possibly removing any
489 pre-existing files of different types.
491 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
492 'argument' field (which so far specified the source to
493 symlink/copy the files from) is now optional. If omitted the
494 same file os copied from /usr/share/factory/ suffixed by the
495 full destination path. This is useful for populating /etc
496 with essential files, by copying them from vendor defaults
497 shipped in /usr/share/factory/etc.
499 * A new command "systemctl preset-all" has been added that
500 applies the service preset settings to all installed unit
501 files. A new switch --preset-mode= has been added that
502 controls whether only enable or only disable operations
505 * A new command "systemctl is-system-running" has been added
506 that allows checking the overall state of the system, for
507 example whether it is fully up and running.
509 * When the system boots up with an empty /etc, the equivalent
510 to "systemctl preset-all" is executed during early boot, to
511 make sure all default services are enabled after a factory
514 * systemd now contains a minimal preset file that enables the
515 most basic services systemd ships by default.
517 * Unit files' [Install] section gained a new DefaultInstance=
518 field for defining the default instance to create if a
519 template unit is enabled with no instance specified.
521 * A new passive target cryptsetup-pre.target has been added
522 that may be used by services that need to make they run and
523 finish before the first LUKS cryptographic device is set up.
525 * The /dev/loop-control and /dev/btrfs-control device nodes
526 are now owned by the "disk" group by default, opening up
527 access to this group.
529 * systemd-coredump will now automatically generate a
530 stack trace of all core dumps taking place on the system,
531 based on elfutils' libdw library. This stack trace is logged
534 * systemd-coredump may now optionally store coredumps directly
535 on disk (in /var/lib/systemd/coredump, possibly compressed),
536 instead of storing them unconditionally in the journal. This
537 mode is the new default. A new configuration file
538 /etc/systemd/coredump.conf has been added to configure this
539 and other parameters of systemd-coredump.
541 * coredumpctl gained a new "info" verb to show details about a
542 specific coredump. A new switch "-1" has also been added
543 that makes sure to only show information about the most
544 recent entry instead of all entries. Also, as the tool is
545 generally useful now the "systemd-" prefix of the binary
546 name has been removed. Distributions that want to maintain
547 compatibility with the old name should add a symlink from
548 the old name to the new name.
550 * journald's SplitMode= now defaults to "uid". This makes sure
551 that unprivileged users can access their own coredumps with
552 coredumpctl without restrictions.
554 * New kernel command line options "systemd.wants=" (for
555 pulling an additional unit during boot), "systemd.mask="
556 (for masking a specific unit for the boot), and
557 "systemd.debug-shell" (for enabling the debug shell on tty9)
558 have been added. This is implemented in the new generator
559 "systemd-debug-generator".
561 * systemd-nspawn will now by default filter a couple of
562 syscalls for containers, among them those required for
563 kernel module loading, direct x86 IO port access, swap
564 management, and kexec. Most importantly though
565 open_by_handle_at() is now prohibited for containers,
566 closing a hole similar to a recently discussed vulnerability
567 in docker regarding access to files on file hierarchies the
568 container should normally not have access to. Note that for
569 nspawn we generally make no security claims anyway (and
570 this is explicitly documented in the man page), so this is
571 just a fix for one of the most obvious problems.
573 * A new man page file-hierarchy(7) has been added that
574 contains a minimized, modernized version of the file system
575 layout systemd expects, similar in style to the FHS
576 specification or hier(5). A new tool systemd-path(1) has
577 been added to query many of these paths for the local
580 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
581 longer done. Since the directory now has a per-user size
582 limit, and is cleaned on logout this appears unnecessary,
583 in particular since this now brings the lifecycle of this
584 directory closer in line with how IPC objects are handled.
586 * systemd.pc now exports a number of additional directories,
587 including $libdir (which is useful to identify the library
588 path for the primary architecture of the system), and a
589 couple of drop-in directories.
591 * udev's predictable network interface names now use the dev_port
592 sysfs attribute, introduced in linux 3.15 instead of dev_id to
593 distinguish between ports of the same PCI function. dev_id should
594 only be used for ports using the same HW address, hence the need
597 * machined has been updated to export the OS version of a
598 container (read from /etc/os-release and
599 /usr/lib/os-release) on the bus. This is now shown in
600 "machinectl status" for a machine.
602 * A new service setting RestartForceExitStatus= has been
603 added. If configured to a set of exit signals or process
604 return values, the service will be restarted when the main
605 daemon process exits with any of them, regardless of the
608 * systemctl's -H switch for connecting to remote systemd
609 machines has been extended so that it may be used to
610 directly connect to a specific container on the
611 host. "systemctl -H root@foobar:waldi" will now connect as
612 user "root" to host "foobar", and then proceed directly to
613 the container named "waldi". Note that currently you have to
614 authenticate as user "root" for this to work, as entering
615 containers is a privileged operation.
617 Contributions from: Andreas Henriksson, Benjamin Steinwender,
618 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
619 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
620 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
621 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
622 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
623 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
624 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
625 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
626 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
627 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
628 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
630 -- Berlin, 2014-07-03
634 * As an experimental feature, udev now tries to lock the
635 disk device node (flock(LOCK_SH|LOCK_NB)) while it
636 executes events for the disk or any of its partitions.
637 Applications like partitioning programs can lock the
638 disk device node (flock(LOCK_EX)) and claim temporary
639 device ownership that way; udev will entirely skip all event
640 handling for this disk and its partitions. If the disk
641 was opened for writing, the close will trigger a partition
642 table rescan in udev's "watch" facility, and if needed
643 synthesize "change" events for the disk and all its partitions.
644 This is now unconditionally enabled, and if it turns out to
645 cause major problems, we might turn it on only for specific
646 devices, or might need to disable it entirely. Device Mapper
647 devices are excluded from this logic.
649 * We temporarily dropped the "-l" switch for fsck invocations,
650 since they collide with the flock() logic above. util-linux
651 upstream has been changed already to avoid this conflict,
652 and we will readd "-l" as soon as util-linux with this
653 change has been released.
655 * The dependency on libattr has been removed. Since a long
656 time, the extended attribute calls have moved to glibc, and
657 libattr is thus unnecessary.
659 * Virtualization detection works without priviliges now. This
660 means the systemd-detect-virt binary no longer requires
661 CAP_SYS_PTRACE file capabilities, and our daemons can run
662 with fewer privileges.
664 * systemd-networkd now runs under its own "systemd-network"
665 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
666 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
667 loses the ability to write to files owned by root this way.
669 * Similar, systemd-resolved now runs under its own
670 "systemd-resolve" user with no capabilities remaining.
672 * Similar, systemd-bus-proxyd now runs under its own
673 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
675 * systemd-networkd gained support for setting up "veth"
676 virtual ethernet devices for container connectivity, as well
677 as GRE and VTI tunnels.
679 * systemd-networkd will no longer automatically attempt to
680 manually load kernel modules necessary for certain tunnel
681 transports. Instead, it is assumed the kernel loads them
682 automatically when required. This only works correctly on
683 very new kernels. On older kernels, please consider adding
684 the kernel modules to /etc/modules-load.d/ as a work-around.
686 * The resolv.conf file systemd-resolved generates has been
687 moved to /run/systemd/resolve/. If you have a symlink from
688 /etc/resolv.conf, it might be necessary to correct it.
690 * Two new service settings, ProtectHome= and ProtectSystem=,
691 have been added. When enabled, they will make the user data
692 (such as /home) inaccessible or read-only and the system
693 (such as /usr) read-only, for specific services. This allows
694 very light-weight per-service sandboxing to avoid
695 modifications of user data or system files from
696 services. These two new switches have been enabled for all
697 of systemd's long-running services, where appropriate.
699 * Socket units gained new SocketUser= and SocketGroup=
700 settings to set the owner user and group of AF_UNIX sockets
701 and FIFOs in the file system.
703 * Socket units gained a new RemoveOnStop= setting. If enabled,
704 all FIFOS and sockets in the file system will be removed
705 when the specific socket unit is stopped.
707 * Socket units gained a new Symlinks= setting. It takes a list
708 of symlinks to create to file system sockets or FIFOs
709 created by the specific Unix sockets. This is useful to
710 manage symlinks to socket nodes with the same life-cycle as
713 * The /dev/log socket and /dev/initctl FIFO have been moved to
714 /run, and have been replaced by symlinks. This allows
715 connecting to these facilities even if PrivateDevices=yes is
716 used for a service (which makes /dev/log itself unavailable,
717 but /run is left). This also has the benefit of ensuring
718 that /dev only contains device nodes, directories and
719 symlinks, and nothing else.
721 * sd-daemon gained two new calls sd_pid_notify() and
722 sd_pid_notifyf(). They are similar to sd_notify() and
723 sd_notifyf(), but allow overriding of the source PID of
724 notification messages if permissions permit this. This is
725 useful to send notify messages on behalf of a different
726 process (for example, the parent process). The
727 systemd-notify tool has been updated to make use of this
728 when sending messages (so that notification messages now
729 originate from the shell script invoking systemd-notify and
730 not the systemd-notify process itself. This should minimize
731 a race where systemd fails to associate notification
732 messages to services when the originating process already
735 * A new "on-abnormal" setting for Restart= has been added. If
736 set, it will result in automatic restarts on all "abnormal"
737 reasons for a process to exit, which includes unclean
738 signals, core dumps, timeouts and watchdog timeouts, but
739 does not include clean and unclean exit codes or clean
740 signals. Restart=on-abnormal is an alternative for
741 Restart=on-failure for services that shall be able to
742 terminate and avoid restarts on certain errors, by
743 indicating so with an unclean exit code. Restart=on-failure
744 or Restart=on-abnormal is now the recommended setting for
745 all long-running services.
747 * If the InaccessibleDirectories= service setting points to a
748 mount point (or if there are any submounts contained within
749 it), it is now attempted to completely unmount it, to make
750 the file systems truly unavailable for the respective
753 * The ReadOnlyDirectories= service setting and
754 systemd-nspawn's --read-only parameter are now recursively
755 applied to all submounts, too.
757 * Mount units may now be created transiently via the bus APIs.
759 * The support for SysV and LSB init scripts has been removed
760 from the systemd daemon itself. Instead, it is now
761 implemented as a generator that creates native systemd units
762 from these scripts when needed. This enables us to remove a
763 substantial amount of legacy code from PID 1, following the
764 fact that many distributions only ship a very small number
765 of LSB/SysV init scripts nowadays.
767 * Priviliged Xen (dom0) domains are not considered
768 virtualization anymore by the virtualization detection
769 logic. After all, they generally have unrestricted access to
770 the hardware and usually are used to manage the unprivileged
773 * systemd-tmpfiles gained a new "C" line type, for copying
774 files or entire directories.
776 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
777 lines. So far, they have been non-globbing versions of the
778 latter, and have thus been redundant. In future, it is
779 recommended to only use "z". "m" has hence been removed
780 from the documentation, even though it stays supported.
782 * A tmpfiles snippet to recreate the most basic structure in
783 /var has been added. This is enough to create the /var/run →
784 /run symlink and create a couple of structural
785 directories. This allows systems to boot up with an empty or
786 volatile /var. Of course, while with this change, the core OS
787 now is capable with dealing with a volatile /var, not all
788 user services are ready for it. However, we hope that sooner
789 or later, many service daemons will be changed upstream so
790 that they are able to automatically create their necessary
791 directories in /var at boot, should they be missing. This is
792 the first step to allow state-less systems that only require
793 the vendor image for /usr to boot.
795 * systemd-nspawn has gained a new --tmpfs= switch to mount an
796 empty tmpfs instance to a specific directory. This is
797 particularly useful for making use of the automatic
798 reconstruction of /var (see above), by passing --tmpfs=/var.
800 * Access modes specified in tmpfiles snippets may now be
801 prefixed with "~", which indicates that they shall be masked
802 by whether the existing file or directory is currently
803 writable, readable or executable at all. Also, if specified,
804 the sgid/suid/sticky bits will be masked for all
807 * A new passive target unit "network-pre.target" has been
808 added which is useful for services that shall run before any
809 network is configured, for example firewall scripts.
811 * The "floppy" group that previously owned the /dev/fd*
812 devices is no longer used. The "disk" group is now used
813 instead. Distributions should probably deprecate usage of
816 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
817 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
818 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
819 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
820 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
821 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
824 -- Berlin, 2014-06-11
828 * A new "systemd-timesyncd" daemon has been added for
829 synchronizing the system clock across the network. It
830 implements an SNTP client. In contrast to NTP
831 implementations such as chrony or the NTP reference server,
832 this only implements a client side, and does not bother with
833 the full NTP complexity, focusing only on querying time from
834 one remote server and synchronizing the local clock to
835 it. Unless you intend to serve NTP to networked clients or
836 want to connect to local hardware clocks, this simple NTP
837 client should be more than appropriate for most
838 installations. The daemon runs with minimal privileges, and
839 has been hooked up with networkd to only operate when
840 network connectivity is available. The daemon saves the
841 current clock to disk every time a new NTP sync has been
842 acquired, and uses this to possibly correct the system clock
843 early at bootup, in order to accommodate for systems that
844 lack an RTC such as the Raspberry Pi and embedded devices,
845 and to make sure that time monotonically progresses on these
846 systems, even if it is not always correct. To make use of
847 this daemon, a new system user and group "systemd-timesync"
848 needs to be created on installation of systemd.
850 * The queue "seqnum" interface of libudev has been disabled, as
851 it was generally incompatible with device namespacing as
852 sequence numbers of devices go "missing" if the devices are
853 part of a different namespace.
855 * "systemctl list-timers" and "systemctl list-sockets" gained
856 a --recursive switch for showing units of these types also
857 for all local containers, similar in style to the already
858 supported --recursive switch for "systemctl list-units".
860 * A new RebootArgument= setting has been added for service
861 units, which may be used to specify a kernel reboot argument
862 to use when triggering reboots with StartLimitAction=.
864 * A new FailureAction= setting has been added for service
865 units which may be used to specify an operation to trigger
866 when a service fails. This works similarly to
867 StartLimitAction=, but unlike it, controls what is done
868 immediately rather than only after several attempts to
869 restart the service in question.
871 * hostnamed got updated to also expose the kernel name,
872 release, and version on the bus. This is useful for
873 executing commands like hostnamectl with the -H switch.
874 systemd-analyze makes use of this to properly display
875 details when running non-locally.
877 * The bootchart tool can now show cgroup information in the
880 * The CFS CPU quota cgroup attribute is now exposed for
881 services. The new CPUQuota= switch has been added for this
882 which takes a percentage value. Setting this will have the
883 result that a service may never get more CPU time than the
884 specified percentage, even if the machine is otherwise idle.
886 * systemd-networkd learned IPIP and SIT tunnel support.
888 * LSB init scripts exposing a dependency on $network will now
889 get a dependency on network-online.target rather than simply
890 network.target. This should bring LSB handling closer to
891 what it was on SysV systems.
893 * A new fsck.repair= kernel option has been added to control
894 how fsck shall deal with unclean file systems at boot.
896 * The (.ini) configuration file parser will now silently
897 ignore sections whose name begins with "X-". This may be
898 used to maintain application-specific extension sections in unit
901 * machined gained a new API to query the IP addresses of
902 registered containers. "machinectl status" has been updated
903 to show these addresses in its output.
905 * A new call sd_uid_get_display() has been added to the
906 sd-login APIs for querying the "primary" session of a
907 user. The "primary" session of the user is elected from the
908 user's sessions and generally a graphical session is
909 preferred over a text one.
911 * A minimal systemd-resolved daemon has been added. It
912 currently simply acts as a companion to systemd-networkd and
913 manages resolv.conf based on per-interface DNS
914 configuration, possibly supplied via DHCP. In the long run
915 we hope to extend this into a local DNSSEC enabled DNS and
918 * The systemd-networkd-wait-online tool is now enabled by
919 default. It will delay network-online.target until a network
920 connection has been configured. The tool primarily integrates
921 with networkd, but will also make a best effort to make sense
922 of network configuration performed in some other way.
924 * Two new service options StartupCPUShares= and
925 StartupBlockIOWeight= have been added that work similarly to
926 CPUShares= and BlockIOWeight= however only apply during
927 system startup. This is useful to prioritize certain services
928 differently during bootup than during normal runtime.
930 * hostnamed has been changed to prefer the statically
931 configured hostname in /etc/hostname (unless set to
932 'localhost' or empty) over any dynamic one supplied by
933 dhcp. With this change, the rules for picking the hostname
934 match more closely the rules of other configuration settings
935 where the local administrator's configuration in /etc always
936 overrides any other settings.
938 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
939 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
940 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
941 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
942 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
943 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
944 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
945 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
946 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
947 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
948 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
949 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
950 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
951 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
952 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
953 Lindskog, WaLyong Cho, Will Woods, Zbigniew
956 -- Beijing, 2014-05-28
960 * When restoring the screen brightness at boot, stay away from
961 the darkest setting or from the lowest 5% of the available
962 range, depending on which is the larger value of both. This
963 should effectively protect the user from rebooting into a
964 black screen, should the brightness have been set to minimum
967 * sd-login gained a new sd_machine_get_class() call to
968 determine the class ("vm" or "container") of a machine
969 registered with machined.
971 * sd-login gained new calls
972 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
973 to query the identity of the peer of a local AF_UNIX
974 connection. They operate similarly to their sd_pid_get_xyz()
977 * PID 1 will now maintain a system-wide system state engine
978 with the states "starting", "running", "degraded",
979 "maintenance", "stopping". These states are bound to system
980 startup, normal runtime, runtime with at least one failed
981 service, rescue/emergency mode and system shutdown. This
982 state is shown in the "systemctl status" output when no unit
983 name is passed. It is useful to determine system state, in
984 particularly when doing so for many systems or containers at
987 * A new command "list-machines" has been added to "systemctl"
988 that lists all local OS containers and shows their system
989 state (see above), if systemd runs inside of them.
991 * systemctl gained a new "-r" switch to recursively enumerate
992 units on all local containers, when used with the
993 "list-unit" command (which is the default one that is
994 executed when no parameters are specified).
996 * The GPT automatic partition discovery logic will now honour
997 two GPT partition flags: one may be set on a partition to
998 cause it to be mounted read-only, and the other may be set
999 on a partition to ignore it during automatic discovery.
1001 * Two new GPT type UUIDs have been added for automatic root
1002 partition discovery, for 32-bit and 64-bit ARM. This is not
1003 particularly useful for discovering the root directory on
1004 these architectures during bare-metal boots (since UEFI is
1005 not common there), but still very useful to allow booting of
1006 ARM disk images in nspawn with the -i option.
1008 * MAC addresses of interfaces created with nspawn's
1009 --network-interface= switch will now be generated from the
1010 machine name, and thus be stable between multiple invocations
1013 * logind will now automatically remove all IPC objects owned
1014 by a user if she or he fully logs out. This makes sure that
1015 users who are logged out cannot continue to consume IPC
1016 resources. This covers SysV memory, semaphores and message
1017 queues as well as POSIX shared memory and message
1018 queues. Traditionally, SysV and POSIX IPC had no life-cycle
1019 limits. With this functionality, that is corrected. This may
1020 be turned off by using the RemoveIPC= switch of logind.conf.
1022 * The systemd-machine-id-setup and tmpfiles tools gained a
1023 --root= switch to operate on a specific root directory,
1026 * journald can now forward logged messages to the TTYs of all
1027 logged in users ("wall"). This is the default for all
1028 emergency messages now.
1030 * A new tool systemd-journal-remote has been added to stream
1031 journal log messages across the network.
1033 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
1034 controller trees are mounted into it. Note that the
1035 directories mounted beneath it are not read-only. This is a
1036 security measure and is particularly useful because glibc
1037 actually includes a search logic to pick any tmpfs it can
1038 find to implement shm_open() if /dev/shm is not available
1039 (which it might very well be in namespaced setups).
1041 * machinectl gained a new "poweroff" command to cleanly power
1042 down a local OS container.
1044 * The PrivateDevices= unit file setting will now also drop the
1045 CAP_MKNOD capability from the capability bound set, and
1046 imply DevicePolicy=closed.
1048 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
1049 comprehensively on all long-running systemd services where
1050 this is appropriate.
1052 * systemd-udevd will now run in a disassociated mount
1053 namespace. To mount directories from udev rules, make sure to
1054 pull in mount units via SYSTEMD_WANTS properties.
1056 * The kdbus support gained support for uploading policy into
1057 the kernel. sd-bus gained support for creating "monitoring"
1058 connections that can eavesdrop into all bus communication
1059 for debugging purposes.
1061 * Timestamps may now be specified in seconds since the UNIX
1062 epoch Jan 1st, 1970 by specifying "@" followed by the value
1065 * Native tcpwrap support in systemd has been removed. tcpwrap
1066 is old code, not really maintained anymore and has serious
1067 shortcomings, and better options such as firewalls
1068 exist. For setups that require tcpwrap usage, please
1069 consider invoking your socket-activated service via tcpd,
1070 like on traditional inetd.
1072 * A new system.conf configuration option
1073 DefaultTimerAccuracySec= has been added that controls the
1074 default AccuracySec= setting of .timer units.
1076 * Timer units gained a new WakeSystem= switch. If enabled,
1077 timers configured this way will cause the system to resume
1078 from system suspend (if the system supports that, which most
1081 * Timer units gained a new Persistent= switch. If enabled,
1082 timers configured this way will save to disk when they have
1083 been last triggered. This information is then used on next
1084 reboot to possible execute overdue timer events, that
1085 could not take place because the system was powered off.
1086 This enables simple anacron-like behaviour for timer units.
1088 * systemctl's "list-timers" will now also list the time a
1089 timer unit was last triggered in addition to the next time
1090 it will be triggered.
1092 * systemd-networkd will now assign predictable IPv4LL
1093 addresses to its local interfaces.
1095 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
1096 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
1097 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
1098 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
1099 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
1100 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
1101 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
1102 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
1105 -- Berlin, 2014-03-25
1109 * A new unit file setting RestrictAddressFamilies= has been
1110 added to restrict which socket address families unit
1111 processes gain access to. This takes address family names
1112 like "AF_INET" or "AF_UNIX", and is useful to minimize the
1113 attack surface of services via exotic protocol stacks. This
1114 is built on seccomp system call filters.
1116 * Two new unit file settings RuntimeDirectory= and
1117 RuntimeDirectoryMode= have been added that may be used to
1118 manage a per-daemon runtime directories below /run. This is
1119 an alternative for setting up directory permissions with
1120 tmpfiles snippets, and has the advantage that the runtime
1121 directory's lifetime is bound to the daemon runtime and that
1122 the daemon starts up with an empty directory each time. This
1123 is particularly useful when writing services that drop
1124 privileges using the User= or Group= setting.
1126 * The DeviceAllow= unit setting now supports globbing for
1127 matching against device group names.
1129 * The systemd configuration file system.conf gained new
1130 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
1131 DefaultMemoryAccounting= to globally turn on/off accounting
1132 for specific resources (cgroups) for all units. These
1133 settings may still be overridden individually in each unit
1136 * systemd-gpt-auto-generator is now able to discover /srv and
1137 root partitions in addition to /home and swap partitions. It
1138 also supports LUKS-encrypted partitions now. With this in
1139 place, automatic discovery of partitions to mount following
1140 the Discoverable Partitions Specification
1141 (http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
1142 is now a lot more complete. This allows booting without
1143 /etc/fstab and without root= on the kernel command line on
1144 systems prepared appropriately.
1146 * systemd-nspawn gained a new --image= switch which allows
1147 booting up disk images and Linux installations on any block
1148 device that follow the Discoverable Partitions Specification
1149 (see above). This means that installations made with
1150 appropriately updated installers may now be started and
1151 deployed using container managers, completely
1152 unmodified. (We hope that libvirt-lxc will add support for
1153 this feature soon, too.)
1155 * systemd-nspawn gained a new --network-macvlan= setting to
1156 set up a private macvlan interface for the
1157 container. Similarly, systemd-networkd gained a new
1158 Kind=macvlan setting in .netdev files.
1160 * systemd-networkd now supports configuring local addresses
1163 * A new tool systemd-network-wait-online has been added to
1164 synchronously wait for network connectivity using
1167 * The sd-bus.h bus API gained a new sd_bus_track object for
1168 tracking the life-cycle of bus peers. Note that sd-bus.h is
1169 still not a public API though (unless you specify
1170 --enable-kdbus on the configure command line, which however
1171 voids your warranty and you get no API stability guarantee).
1173 * The $XDG_RUNTIME_DIR runtime directories for each user are
1174 now individual tmpfs instances, which has the benefit of
1175 introducing separate pools for each user, with individual
1176 size limits, and thus making sure that unprivileged clients
1177 can no longer negatively impact the system or other users by
1178 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
1179 RuntimeDirectorySize= has been introduced that allows
1180 controlling the default size limit for all users. It
1181 defaults to 10% of the available physical memory. This is no
1182 replacement for quotas on tmpfs though (which the kernel
1183 still does not support), as /dev/shm and /tmp are still
1184 shared resources used by both the system and unprivileged
1187 * logind will now automatically turn off automatic suspending
1188 on laptop lid close when more than one display is
1189 connected. This was previously expected to be implemented
1190 individually in desktop environments (such as GNOME),
1191 however has been added to logind now, in order to fix a
1192 boot-time race where a desktop environment might not have
1193 been started yet and thus not been able to take an inhibitor
1194 lock at the time where logind already suspends the system
1195 due to a closed lid.
1197 * logind will now wait at least 30s after each system
1198 suspend/resume cycle, and 3min after system boot before
1199 suspending the system due to a closed laptop lid. This
1200 should give USB docking stations and similar enough time to
1201 be probed and configured after system resume and boot in
1202 order to then act as suspend blocker.
1204 * systemd-run gained a new --property= setting which allows
1205 initialization of resource control properties (and others)
1206 for the created scope or service unit. Example: "systemd-run
1207 --property=BlockIOWeight=10 updatedb" may be used to run
1208 updatedb at a low block IO scheduling weight.
1210 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
1211 now also work in --scope mode.
1213 * When systemd is compiled with kdbus support, basic support
1214 for enforced policies is now in place. (Note that enabling
1215 kdbus still voids your warranty and no API compatibility
1218 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
1219 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
1220 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
1221 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
1222 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
1223 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
1224 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
1225 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
1226 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
1227 Zbigniew Jędrzejewski-Szmek
1229 -- Berlin, 2014-03-12
1233 * systemd will now relabel /dev after loading the SMACK policy
1234 according to SMACK rules.
1236 * A new unit file option AppArmorProfile= has been added to
1237 set the AppArmor profile for the processes of a unit.
1239 * A new condition check ConditionArchitecture= has been added
1240 to conditionalize units based on the system architecture, as
1241 reported by uname()'s "machine" field.
1243 * systemd-networkd now supports matching on the system
1244 virtualization, architecture, kernel command line, host name
1247 * logind is now a lot more aggressive when suspending the
1248 machine due to a closed laptop lid. Instead of acting only
1249 on the lid close action, it will continuously watch the lid
1250 status and act on it. This is useful for laptops where the
1251 power button is on the outside of the chassis so that it can
1252 be reached without opening the lid (such as the Lenovo
1253 Yoga). On those machines, logind will now immediately
1254 re-suspend the machine if the power button has been
1255 accidentally pressed while the laptop was suspended and in a
1256 backpack or similar.
1258 * logind will now watch SW_DOCK switches and inhibit reaction
1259 to the lid switch if it is pressed. This means that logind
1260 will not suspend the machine anymore if the lid is closed
1261 and the system is docked, if the laptop supports SW_DOCK
1262 notifications via the input layer. Note that ACPI docking
1263 stations do not generate this currently. Also note that this
1264 logic is usually not fully sufficient and Desktop
1265 Environments should take a lid switch inhibitor lock when an
1266 external display is connected, as systemd will not watch
1269 * nspawn will now make use of the devices cgroup controller by
1270 default, and only permit creation of and access to the usual
1271 API device nodes like /dev/null or /dev/random, as well as
1272 access to (but not creation of) the pty devices.
1274 * We will now ship a default .network file for
1275 systemd-networkd that automatically configures DHCP for
1276 network interfaces created by nspawn's --network-veth or
1277 --network-bridge= switches.
1279 * systemd will now understand the usual M, K, G, T suffixes
1280 according to SI conventions (i.e. to the base 1000) when
1281 referring to throughput and hardware metrics. It will stay
1282 with IEC conventions (i.e. to the base 1024) for software
1283 metrics, according to what is customary according to
1284 Wikipedia. We explicitly document which base applies for
1285 each configuration option.
1287 * The DeviceAllow= setting in unit files now supports a syntax
1288 to whitelist an entire group of devices node majors at once,
1289 based on the /proc/devices listing. For example, with the
1290 string "char-pts", it is now possible to whitelist all
1291 current and future pseudo-TTYs at once.
1293 * sd-event learned a new "post" event source. Event sources of
1294 this type are triggered by the dispatching of any event
1295 source of a type that is not "post". This is useful for
1296 implementing clean-up and check event sources that are
1297 triggered by other work being done in the program.
1299 * systemd-networkd is no longer statically enabled, but uses
1300 the usual [Install] sections so that it can be
1301 enabled/disabled using systemctl. It still is enabled by
1304 * When creating a veth interface pair with systemd-nspawn, the
1305 host side will now be prefixed with "vb-" if
1306 --network-bridge= is used, and with "ve-" if --network-veth
1307 is used. This way, it is easy to distinguish these cases on
1308 the host, for example to apply different configuration to
1309 them with systemd-networkd.
1311 * The compatibility libraries for libsystemd-journal.so,
1312 libsystem-id128.so, libsystemd-login.so and
1313 libsystemd-daemon.so do not make use of IFUNC
1314 anymore. Instead, we now build libsystemd.so multiple times
1315 under these alternative names. This means that the footprint
1316 is drastically increased, but given that these are
1317 transitional compatibility libraries, this should not matter
1318 much. This change has been made necessary to support the ARM
1319 platform for these compatibility libraries, as the ARM
1320 toolchain is not really at the same level as the toolchain
1321 for other architectures like x86 and does not support
1322 IFUNC. Please make sure to use --enable-compat-libs only
1323 during a transitional period!
1325 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
1326 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
1327 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
1328 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
1329 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
1330 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
1331 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
1332 Zbigniew Jędrzejewski-Szmek
1334 -- Berlin, 2014-02-24
1338 * A new component "systemd-networkd" has been added that can
1339 be used to configure local network interfaces statically or
1340 via DHCP. It is capable of bringing up bridges, VLANs, and
1341 bonding. Currently, no hook-ups for interactive network
1342 configuration are provided. Use this for your initrd,
1343 container, embedded, or server setup if you need a simple,
1344 yet powerful, network configuration solution. This
1345 configuration subsystem is quite nifty, as it allows wildcard
1346 hotplug matching in interfaces. For example, with a single
1347 configuration snippet, you can configure that all Ethernet
1348 interfaces showing up are automatically added to a bridge,
1349 or similar. It supports link-sensing and more.
1351 * A new tool "systemd-socket-proxyd" has been added which can
1352 act as a bidirectional proxy for TCP sockets. This is
1353 useful for adding socket activation support to services that
1354 do not actually support socket activation, including virtual
1355 machines and the like.
1357 * Add a new tool to save/restore rfkill state on
1360 * Save/restore state of keyboard backlights in addition to
1361 display backlights on shutdown/boot.
1363 * udev learned a new SECLABEL{} construct to label device
1364 nodes with a specific security label when they appear. For
1365 now, only SECLABEL{selinux} is supported, but the syntax is
1366 prepared for additional security frameworks.
1368 * udev gained a new scheme to configure link-level attributes
1369 from files in /etc/systemd/network/*.link. These files can
1370 match against MAC address, device path, driver name and type,
1371 and will apply attributes like the naming policy, link speed,
1372 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
1373 address assignment policy (randomized, ...).
1375 * The configuration of network interface naming rules for
1376 "permanent interface names" has changed: a new NamePolicy=
1377 setting in the [Link] section of .link files determines the
1378 priority of possible naming schemes (onboard, slot, mac,
1379 path). The default value of this setting is determined by
1380 /usr/lib/net/links/99-default.link. Old
1381 80-net-name-slot.rules udev configuration file has been
1382 removed, so local configuration overriding this file should
1383 be adapated to override 99-default.link instead.
1385 * When the User= switch is used in a unit file, also
1386 initialize $SHELL= based on the user database entry.
1388 * systemd no longer depends on libdbus. All communication is
1389 now done with sd-bus, systemd's low-level bus library
1392 * kdbus support has been added to PID 1 itself. When kdbus is
1393 enabled, this causes PID 1 to set up the system bus and
1394 enable support for a new ".busname" unit type that
1395 encapsulates bus name activation on kdbus. It works a little
1396 bit like ".socket" units, except for bus names. A new
1397 generator has been added that converts classic dbus1 service
1398 activation files automatically into native systemd .busname
1401 * sd-bus: add a light-weight vtable implementation that allows
1402 defining objects on the bus with a simple static const
1403 vtable array of its methods, signals and properties.
1405 * systemd will not generate or install static dbus
1406 introspection data anymore to /usr/share/dbus-1/interfaces,
1407 as the precise format of these files is unclear, and
1408 nothing makes use of it.
1410 * A proxy daemon is now provided to proxy clients connecting
1411 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
1412 compatibility with classic D-Bus.
1414 * A bus driver implementation has been added that supports the
1415 classic D-Bus bus driver calls on kdbus, also for
1416 compatibility purposes.
1418 * A new API "sd-event.h" has been added that implements a
1419 minimal event loop API built around epoll. It provides a
1420 couple of features that direct epoll usage is lacking:
1421 prioritization of events, scales to large numbers of timer
1422 events, per-event timer slack (accuracy), system-wide
1423 coalescing of timer events, exit handlers, watchdog
1424 supervision support using systemd's sd_notify() API, child
1427 * A new API "sd-rntl.h" has been added that provides an API
1428 around the route netlink interface of the kernel, similar in
1429 style to "sd-bus.h".
1431 * A new API "sd-dhcp-client.h" has been added that provides a
1432 small DHCPv4 client-side implementation. This is used by
1435 * There is a new kernel command line option
1436 "systemd.restore_state=0|1". When set to "0", none of the
1437 systemd tools will restore saved runtime state to hardware
1438 devices. More specifically, the rfkill and backlight states
1441 * The FsckPassNo= compatibility option in mount/service units
1442 has been removed. The fstab generator will now add the
1443 necessary dependencies automatically, and does not require
1444 PID1's support for that anymore.
1446 * journalctl gained a new switch, --list-boots, that lists
1447 recent boots with their times and boot IDs.
1449 * The various tools like systemctl, loginctl, timedatectl,
1450 busctl, systemd-run, ... have gained a new switch "-M" to
1451 connect to a specific, local OS container (as direct
1452 connection, without requiring SSH). This works on any
1453 container that is registered with machined, such as those
1454 created by libvirt-lxc or nspawn.
1456 * systemd-run and systemd-analyze also gained support for "-H"
1457 to connect to remote hosts via SSH. This is particularly
1458 useful for systemd-run because it enables queuing of jobs
1459 onto remote systems.
1461 * machinectl gained a new command "login" to open a getty
1462 login in any local container. This works with any container
1463 that is registered with machined (such as those created by
1464 libvirt-lxc or nspawn), and which runs systemd inside.
1466 * machinectl gained a new "reboot" command that may be used to
1467 trigger a reboot on a specific container that is registered
1468 with machined. This works on any container that runs an init
1469 system of some kind.
1471 * systemctl gained a new "list-timers" command to print a nice
1472 listing of installed timer units with the times they elapse
1475 * Alternative reboot() parameters may now be specified on the
1476 "systemctl reboot" command line and are passed to the
1477 reboot() system call.
1479 * systemctl gained a new --job-mode= switch to configure the
1480 mode to queue a job with. This is a more generic version of
1481 --fail, --irreversible, and --ignore-dependencies, which are
1482 still available but not advertised anymore.
1484 * /etc/systemd/system.conf gained new settings to configure
1485 various default timeouts of units, as well as the default
1486 start limit interval and burst. These may still be overridden
1489 * PID1 will now export on the bus profile data of the security
1490 policy upload process (such as the SELinux policy upload to
1493 * journald: when forwarding logs to the console, include
1494 timestamps (following the setting in
1495 /sys/module/printk/parameters/time).
1497 * OnCalendar= in timer units now understands the special
1498 strings "yearly" and "annually". (Both are equivalent)
1500 * The accuracy of timer units is now configurable with the new
1501 AccuracySec= setting. It defaults to 1min.
1503 * A new dependency type JoinsNamespaceOf= has been added that
1504 allows running two services within the same /tmp and network
1505 namespace, if PrivateNetwork= or PrivateTmp= are used.
1507 * A new command "cat" has been added to systemctl. It outputs
1508 the original unit file of a unit, and concatenates the
1509 contents of additional "drop-in" unit file snippets, so that
1510 the full configuration is shown.
1512 * systemctl now supports globbing on the various "list-xyz"
1513 commands, like "list-units" or "list-sockets", as well as on
1514 those commands which take multiple unit names.
1516 * journalctl's --unit= switch gained support for globbing.
1518 * All systemd daemons now make use of the watchdog logic so
1519 that systemd automatically notices when they hang.
1521 * If the $container_ttys environment variable is set,
1522 getty-generator will automatically spawn a getty for each
1523 listed tty. This is useful for container managers to request
1524 login gettys to be spawned on as many ttys as needed.
1526 * %h, %s, %U specifier support is not available anymore when
1527 used in unit files for PID 1. This is because NSS calls are
1528 not safe from PID 1. They stay available for --user
1529 instances of systemd, and as special case for the root user.
1531 * loginctl gained a new "--no-legend" switch to turn off output
1534 * The "sd-login.h" API gained three new calls:
1535 sd_session_is_remote(), sd_session_get_remote_user(),
1536 sd_session_get_remote_host() to query information about
1539 * The udev hardware database now also carries vendor/product
1540 information of SDIO devices.
1542 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
1543 determine whether watchdog notifications are requested by
1546 * Socket-activated per-connection services now include a
1547 short description of the connection parameters in the
1550 * tmpfiles gained a new "--boot" option. When this is not used,
1551 only lines where the command character is not suffixed with
1552 "!" are executed. When this option is specified, those
1553 options are executed too. This partitions tmpfiles
1554 directives into those that can be safely executed at any
1555 time, and those which should be run only at boot (for
1556 example, a line that creates /run/nologin).
1558 * A new API "sd-resolve.h" has been added which provides a simple
1559 asynchronous wrapper around glibc NSS host name resolution
1560 calls, such as getaddrinfo(). In contrast to glibc's
1561 getaddrinfo_a(), it does not use signals. In contrast to most
1562 other asynchronous name resolution libraries, this one does
1563 not reimplement DNS, but reuses NSS, so that alternate
1564 host name resolution systems continue to work, such as mDNS,
1565 LDAP, etc. This API is based on libasyncns, but it has been
1566 cleaned up for inclusion in systemd.
1568 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
1569 "sd-daemon.h" are no longer found in individual libraries
1570 libsystemd-journal.so, libsystemd-login.so,
1571 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
1572 merged them into a single library, libsystemd.so, which
1573 provides all symbols. The reason for this is cyclic
1574 dependencies, as these libraries tend to use each other's
1575 symbols. So far, we have managed to workaround that by linking
1576 a copy of a good part of our code into each of these
1577 libraries again and again, which, however, makes certain
1578 things hard to do, like sharing static variables. Also, it
1579 substantially increases footprint. With this change, there
1580 is only one library for the basic APIs systemd
1581 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
1582 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
1583 library as well, however are subject to the --enable-kdbus
1584 switch (see below). Note that "sd-dhcp-client.h" is not part
1585 of this library (this is because it only consumes, never
1586 provides, services of/to other APIs). To make the transition
1587 easy from the separate libraries to the unified one, we
1588 provide the --enable-compat-libs compile-time switch which
1589 will generate stub libraries that are compatible with the
1590 old ones but redirect all calls to the new one.
1592 * All of the kdbus logic and the new APIs "sd-bus.h",
1593 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
1594 and "sd-utf8.h" are compile-time optional via the
1595 "--enable-kdbus" switch, and they are not compiled in by
1596 default. To make use of kdbus, you have to explicitly enable
1597 the switch. Note however, that neither the kernel nor the
1598 userspace API for all of this is considered stable yet. We
1599 want to maintain the freedom to still change the APIs for
1600 now. By specifying this build-time switch, you acknowledge
1601 that you are aware of the instability of the current
1604 * Also, note that while kdbus is pretty much complete,
1605 it lacks one thing: proper policy support. This means you
1606 can build a fully working system with all features; however,
1607 it will be highly insecure. Policy support will be added in
1608 one of the next releases, at the same time that we will
1609 declare the APIs stable.
1611 * When the kernel command-line argument "kdbus" is specified,
1612 systemd will automatically load the kdbus.ko kernel module. At
1613 this stage of development, it is only useful for testing kdbus
1614 and should not be used in production. Note: if "--enable-kdbus"
1615 is specified, and the kdbus.ko kernel module is available, and
1616 "kdbus" is added to the kernel command line, the entire system
1617 runs with kdbus instead of dbus-daemon, with the above mentioned
1618 problem of missing the system policy enforcement. Also a future
1619 version of kdbus.ko or a newer systemd will not be compatible with
1620 each other, and will unlikely be able to boot the machine if only
1621 one of them is updated.
1623 * systemctl gained a new "import-environment" command which
1624 uploads the caller's environment (or parts thereof) into the
1625 service manager so that it is inherited by services started
1626 by the manager. This is useful to upload variables like
1627 $DISPLAY into the user service manager.
1629 * A new PrivateDevices= switch has been added to service units
1630 which allows running a service with a namespaced /dev
1631 directory that does not contain any device nodes for
1632 physical devices. More specifically, it only includes devices
1633 such as /dev/null, /dev/urandom, and /dev/zero which are API
1636 * logind has been extended to support behaviour like VT
1637 switching on seats that do not support a VT. This makes
1638 multi-session available on seats that are not the first seat
1639 (seat0), and on systems where kernel support for VTs has
1640 been disabled at compile-time.
1642 * If a process holds a delay lock for system sleep or shutdown
1643 and fails to release it in time, we will now log its
1644 identity. This makes it easier to identify processes that
1645 cause slow suspends or power-offs.
1647 * When parsing /etc/crypttab, support for a new key-slot=
1648 option as supported by Debian is added. It allows indicating
1649 which LUKS slot to use on disk, speeding up key loading.
1651 * The sd_journald_sendv() API call has been checked and
1652 officially declared to be async-signal-safe so that it may
1653 be invoked from signal handlers for logging purposes.
1655 * Boot-time status output is now enabled automatically after a
1656 short timeout if boot does not progress, in order to give
1657 the user an indication what she or he is waiting for.
1659 * The boot-time output has been improved to show how much time
1660 remains until jobs expire.
1662 * The KillMode= switch in service units gained a new possible
1663 value "mixed". If set, and the unit is shut down, then the
1664 initial SIGTERM signal is sent only to the main daemon
1665 process, while the following SIGKILL signal is sent to
1666 all remaining processes of the service.
1668 * When a scope unit is registered, a new property "Controller"
1669 may be set. If set to a valid bus name, systemd will send a
1670 RequestStop() signal to this name when it would like to shut
1671 down the scope. This may be used to hook manager logic into
1672 the shutdown logic of scope units. Also, scope units may now
1673 be put in a special "abandoned" state, in which case the
1674 manager process which created them takes no further
1675 responsibilities for it.
1677 * When reading unit files, systemd will now verify
1678 the access mode of these files, and warn about certain
1679 suspicious combinations. This has been added to make it
1680 easier to track down packaging bugs where unit files are
1681 marked executable or world-writable.
1683 * systemd-nspawn gained a new "--setenv=" switch to set
1684 container-wide environment variables. The similar option in
1685 systemd-activate was renamed from "--environment=" to
1686 "--setenv=" for consistency.
1688 * systemd-nspawn has been updated to create a new kdbus domain
1689 for each container that is invoked, thus allowing each
1690 container to have its own set of system and user buses,
1691 independent of the host.
1693 * systemd-nspawn gained a new --drop-capability= switch to run
1694 the container with less capabilities than the default. Both
1695 --drop-capability= and --capability= now take the special
1696 string "all" for dropping or keeping all capabilities.
1698 * systemd-nspawn gained new switches for executing containers
1699 with specific SELinux labels set.
1701 * systemd-nspawn gained a new --quiet switch to not generate
1702 any additional output but the container's own console
1705 * systemd-nspawn gained a new --share-system switch to run a
1706 container without PID namespacing enabled.
1708 * systemd-nspawn gained a new --register= switch to control
1709 whether the container is registered with systemd-machined or
1710 not. This is useful for containers that do not run full
1711 OS images, but only specific apps.
1713 * systemd-nspawn gained a new --keep-unit which may be used
1714 when invoked as the only program from a service unit, and
1715 results in registration of the unit service itself in
1716 systemd-machined, instead of a newly opened scope unit.
1718 * systemd-nspawn gained a new --network-interface= switch for
1719 moving arbitrary interfaces to the container. The new
1720 --network-veth switch creates a virtual Ethernet connection
1721 between host and container. The new --network-bridge=
1722 switch then allows assigning the host side of this virtual
1723 Ethernet connection to a bridge device.
1725 * systemd-nspawn gained a new --personality= switch for
1726 setting the kernel personality for the container. This is
1727 useful when running a 32-bit container on a 64-bit host. A
1728 similar option Personality= is now also available for service
1731 * logind will now also track a "Desktop" identifier for each
1732 session which encodes the desktop environment of it. This is
1733 useful for desktop environments that want to identify
1734 multiple running sessions of itself easily.
1736 * A new SELinuxContext= setting for service units has been
1737 added that allows setting a specific SELinux execution
1738 context for a service.
1740 * Most systemd client tools will now honour $SYSTEMD_LESS for
1741 settings of the "less" pager. By default, these tools will
1742 override $LESS to allow certain operations to work, such as
1743 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
1744 influence this logic.
1746 * systemd's "seccomp" hook-up has been changed to make use of
1747 the libseccomp library instead of using its own
1748 implementation. This has benefits for portability among
1751 * For usage together with SystemCallFilter=, a new
1752 SystemCallErrorNumber= setting has been introduced that
1753 allows configuration of a system error number to be returned
1754 on filtered system calls, instead of immediately killing the
1755 process. Also, SystemCallArchitectures= has been added to
1756 limit access to system calls of a particular architecture
1757 (in order to turn off support for unused secondary
1758 architectures). There is also a global
1759 SystemCallArchitectures= setting in system.conf now to turn
1760 off support for non-native system calls system-wide.
1762 * systemd requires a kernel with a working name_to_handle_at(),
1763 please see the kernel config requirements in the README file.
1765 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
1766 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
1767 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
1768 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
1769 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
1770 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
1771 Elia Pinto, Florian Weimer, George McCollister, Goffredo
1772 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
1773 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
1774 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
1775 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
1776 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
1777 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
1778 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
1779 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
1780 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
1781 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
1782 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
1783 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
1784 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
1785 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
1786 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
1787 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
1788 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
1790 -- Berlin, 2014-02-20
1794 * logind has gained support for facilitating privileged input
1795 and drm device access for unprivileged clients. This work is
1796 useful to allow Wayland display servers (and similar
1797 programs, such as kmscon) to run under the user's ID and
1798 access input and drm devices which are normally
1799 protected. When this is used (and the kernel is new enough)
1800 logind will "mute" IO on the file descriptors passed to
1801 Wayland as long as it is in the background and "unmute" it
1802 if it returns into the foreground. This allows secure
1803 session switching without allowing background sessions to
1804 eavesdrop on input and display data. This also introduces
1805 session switching support if VT support is turned off in the
1806 kernel, and on seats that are not seat0.
1808 * A new kernel command line option luks.options= is understood
1809 now which allows specifying LUKS options for usage for LUKS
1810 encrypted partitions specified with luks.uuid=.
1812 * tmpfiles.d(5) snippets may now use specifier expansion in
1813 path names. More specifically %m, %b, %H, %v, are now
1814 replaced by the local machine id, boot id, hostname, and
1815 kernel version number.
1817 * A new tmpfiles.d(5) command "m" has been introduced which
1818 may be used to change the owner/group/access mode of a file
1819 or directory if it exists, but do nothing if it does not.
1821 * This release removes high-level support for the
1822 MemorySoftLimit= cgroup setting. The underlying kernel
1823 cgroup attribute memory.soft_limit= is currently badly
1824 designed and likely to be removed from the kernel API in its
1825 current form, hence we should not expose it for now.
1827 * The memory.use_hierarchy cgroup attribute is now enabled for
1828 all cgroups systemd creates in the memory cgroup
1829 hierarchy. This option is likely to be come the built-in
1830 default in the kernel anyway, and the non-hierarchial mode
1831 never made much sense in the intrinsically hierarchial
1834 * A new field _SYSTEMD_SLICE= is logged along with all journal
1835 messages containing the slice a message was generated
1836 from. This is useful to allow easy per-customer filtering of
1837 logs among other things.
1839 * systemd-journald will no longer adjust the group of journal
1840 files it creates to the "systemd-journal" group. Instead we
1841 rely on the journal directory to be owned by the
1842 "systemd-journal" group, and its setgid bit set, so that the
1843 kernel file system layer will automatically enforce that
1844 journal files inherit this group assignment. The reason for
1845 this change is that we cannot allow NSS look-ups from
1846 journald which would be necessary to resolve
1847 "systemd-journal" to a numeric GID, because this might
1848 create deadlocks if NSS involves synchronous queries to
1849 other daemons (such as nscd, or sssd) which in turn are
1850 logging clients of journald and might block on it, which
1851 would then dead lock. A tmpfiles.d(5) snippet included in
1852 systemd will make sure the setgid bit and group are
1853 properly set on the journal directory if it exists on every
1854 boot. However, we recommend adjusting it manually after
1855 upgrades too (or from RPM scriptlets), so that the change is
1856 not delayed until next reboot.
1858 * Backlight and random seed files in /var/lib/ have moved into
1859 the /var/lib/systemd/ directory, in order to centralize all
1860 systemd generated files in one directory.
1862 * Boot time performance measurements (as displayed by
1863 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
1864 performance information if that's available to determine how
1865 much time BIOS and boot loader initialization required. With
1866 a sufficiently new BIOS you hence no longer need to boot
1867 with Gummiboot to get access to such information.
1869 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
1870 Cristian Rodríguez, Dave Reisner, David Herrmann, David
1871 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
1872 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
1873 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
1874 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
1875 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1877 -- Berlin, 2013-10-02
1881 * The Restart= option for services now understands a new
1882 on-watchdog setting, which will restart the service
1883 automatically if the service stops sending out watchdog keep
1884 alive messages (as configured with WatchdogSec=).
1886 * The getty generator (which is responsible for bringing up a
1887 getty on configured serial consoles) will no longer only
1888 start a getty on the primary kernel console but on all
1889 others, too. This makes the order in which console= is
1890 specified on the kernel command line less important.
1892 * libsystemd-logind gained a new sd_session_get_vt() call to
1893 retrieve the VT number of a session.
1895 * If the option "tries=0" is set for an entry of /etc/crypttab
1896 its passphrase is queried indefinitely instead of any
1897 maximum number of tries.
1899 * If a service with a configure PID file terminates its PID
1900 file will now be removed automatically if it still exists
1901 afterwards. This should put an end to stale PID files.
1903 * systemd-run will now also take relative binary path names
1904 for execution and no longer insists on absolute paths.
1906 * InaccessibleDirectories= and ReadOnlyDirectories= now take
1907 paths that are optionally prefixed with "-" to indicate that
1908 it should not be considered a failure if they do not exist.
1910 * journalctl -o (and similar commands) now understands a new
1911 output mode "short-precise", it is similar to "short" but
1912 shows timestamps with usec accuracy.
1914 * The option "discard" (as known from Debian) is now
1915 synonymous to "allow-discards" in /etc/crypttab. In fact,
1916 "discard" is preferred now (since it is easier to remember
1919 * Some licensing clean-ups were made, so that more code is now
1920 LGPL-2.1 licensed than before.
1922 * A minimal tool to save/restore the display backlight
1923 brightness across reboots has been added. It will store the
1924 backlight setting as late as possible at shutdown, and
1925 restore it as early as possible during reboot.
1927 * A logic to automatically discover and enable home and swap
1928 partitions on GPT disks has been added. With this in place
1929 /etc/fstab becomes optional for many setups as systemd can
1930 discover certain partitions located on the root disk
1931 automatically. Home partitions are recognized under their
1932 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
1933 partitions are recognized under their GPT type ID
1934 0657fd6da4ab43c484e50933c84b4f4f.
1936 * systemd will no longer pass any environment from the kernel
1937 or initrd to system services. If you want to set an
1938 environment for all services, do so via the kernel command
1939 line systemd.setenv= assignment.
1941 * The systemd-sysctl tool no longer natively reads the file
1942 /etc/sysctl.conf. If desired, the file should be symlinked
1943 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
1944 legacy support by a symlink rather than built-in code, it
1945 also makes the otherwise hidden order of application of the
1946 different files visible. (Note that this partly reverts to a
1947 pre-198 application order of sysctl knobs!)
1949 * The "systemctl set-log-level" and "systemctl dump" commands
1950 have been moved to systemd-analyze.
1952 * systemd-run learned the new --remain-after-exit switch,
1953 which causes the scope unit not to be cleaned up
1954 automatically after the process terminated.
1956 * tmpfiles learned a new --exclude-prefix= switch to exclude
1957 certain paths from operation.
1959 * journald will now automatically flush all messages to disk
1960 as soon as a message of the log priorities CRIT, ALERT or
1963 Contributions from: Andrew Cook, Brandon Philips, Christian
1964 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
1965 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
1966 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
1967 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
1968 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
1969 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
1970 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
1971 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
1972 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
1973 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
1974 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
1975 William Giokas, Zbigniew Jędrzejewski-Szmek
1977 -- Berlin, 2013-09-13
1981 * The documentation has been updated to cover the various new
1982 concepts introduced with 205.
1984 * Unit files now understand the new %v specifier which
1985 resolves to the kernel version string as returned by "uname
1988 * systemctl now supports filtering the unit list output by
1989 load state, active state and sub state, using the new
1992 * "systemctl status" will now show the results of the
1993 condition checks (like ConditionPathExists= and similar) of
1994 the last start attempts of the unit. They are also logged to
1997 * "journalctl -b" may now be used to look for boot output of a
1998 specific boot. Try "journalctl -b -1" for the previous boot,
1999 but the syntax is substantially more powerful.
2001 * "journalctl --show-cursor" has been added which prints the
2002 cursor string the last shown log line. This may then be used
2003 with the new "journalctl --after-cursor=" switch to continue
2004 browsing logs from that point on.
2006 * "journalctl --force" may now be used to force regeneration
2009 * Creation of "dead" device nodes has been moved from udev
2010 into kmod and tmpfiles. Previously, udev would read the kmod
2011 databases to pre-generate dead device nodes based on meta
2012 information contained in kernel modules, so that these would
2013 be auto-loaded on access rather then at boot. As this
2014 does not really have much to do with the exposing actual
2015 kernel devices to userspace this has always been slightly
2016 alien in the udev codebase. Following the new scheme kmod
2017 will now generate a runtime snippet for tmpfiles from the
2018 module meta information and it now is tmpfiles' job to the
2019 create the nodes. This also allows overriding access and
2020 other parameters for the nodes using the usual tmpfiles
2021 facilities. As side effect this allows us to remove the
2022 CAP_SYS_MKNOD capability bit from udevd entirely.
2024 * logind's device ACLs may now be applied to these "dead"
2025 devices nodes too, thus finally allowing managed access to
2026 devices such as /dev/snd/sequencer whithout loading the
2027 backing module right-away.
2029 * A new RPM macro has been added that may be used to apply
2030 tmpfiles configuration during package installation.
2032 * systemd-detect-virt and ConditionVirtualization= now can
2033 detect User-Mode-Linux machines (UML).
2035 * journald will now implicitly log the effective capabilities
2036 set of processes in the message metadata.
2038 * systemd-cryptsetup has gained support for TrueCrypt volumes.
2040 * The initrd interface has been simplified (more specifically,
2041 support for passing performance data via environment
2042 variables and fsck results via files in /run has been
2043 removed). These features were non-essential, and are
2044 nowadays available in a much nicer way by having systemd in
2045 the initrd serialize its state and have the hosts systemd
2046 deserialize it again.
2048 * The udev "keymap" data files and tools to apply keyboard
2049 specific mappings of scan to key codes, and force-release
2050 scan code lists have been entirely replaced by a udev
2051 "keyboard" builtin and a hwdb data file.
2053 * systemd will now honour the kernel's "quiet" command line
2054 argument also during late shutdown, resulting in a
2055 completely silent shutdown when used.
2057 * There's now an option to control the SO_REUSEPORT socket
2058 option in .socket units.
2060 * Instance units will now automatically get a per-template
2061 subslice of system.slice unless something else is explicitly
2062 configured. For example, instances of sshd@.service will now
2063 implicitly be placed in system-sshd.slice rather than
2064 system.slice as before.
2066 * Test coverage support may now be enabled at build time.
2068 Contributions from: Dave Reisner, Frederic Crozat, Harald
2069 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
2070 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
2071 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
2072 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
2073 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
2074 Giokas, Zbigniew Jędrzejewski-Szmek
2076 -- Berlin, 2013-07-23
2080 * Two new unit types have been introduced:
2082 Scope units are very similar to service units, however, are
2083 created out of pre-existing processes -- instead of PID 1
2084 forking off the processes. By using scope units it is
2085 possible for system services and applications to group their
2086 own child processes (worker processes) in a powerful way
2087 which then maybe used to organize them, or kill them
2088 together, or apply resource limits on them.
2090 Slice units may be used to partition system resources in an
2091 hierarchial fashion and then assign other units to them. By
2092 default there are now three slices: system.slice (for all
2093 system services), user.slice (for all user sessions),
2094 machine.slice (for VMs and containers).
2096 Slices and scopes have been introduced primarily in
2097 context of the work to move cgroup handling to a
2098 single-writer scheme, where only PID 1
2099 creates/removes/manages cgroups.
2101 * There's a new concept of "transient" units. In contrast to
2102 normal units these units are created via an API at runtime,
2103 not from configuration from disk. More specifically this
2104 means it is now possible to run arbitrary programs as
2105 independent services, with all execution parameters passed
2106 in via bus APIs rather than read from disk. Transient units
2107 make systemd substantially more dynamic then it ever was,
2108 and useful as a general batch manager.
2110 * logind has been updated to make use of scope and slice units
2111 for managing user sessions. As a user logs in he will get
2112 his own private slice unit, to which all sessions are added
2113 as scope units. We also added support for automatically
2114 adding an instance of user@.service for the user into the
2115 slice. Effectively logind will no longer create cgroup
2116 hierarchies on its own now, it will defer entirely to PID 1
2117 for this by means of scope, service and slice units. Since
2118 user sessions this way become entities managed by PID 1
2119 the output of "systemctl" is now a lot more comprehensive.
2121 * A new mini-daemon "systemd-machined" has been added which
2122 may be used by virtualization managers to register local
2123 VMs/containers. nspawn has been updated accordingly, and
2124 libvirt will be updated shortly. machined will collect a bit
2125 of meta information about the VMs/containers, and assign
2126 them their own scope unit (see above). The collected
2127 meta-data is then made available via the "machinectl" tool,
2128 and exposed in "ps" and similar tools. machined/machinectl
2129 is compile-time optional.
2131 * As discussed earlier, the low-level cgroup configuration
2132 options ControlGroup=, ControlGroupModify=,
2133 ControlGroupPersistent=, ControlGroupAttribute= have been
2134 removed. Please use high-level attribute settings instead as
2135 well as slice units.
2137 * A new bus call SetUnitProperties() has been added to alter
2138 various runtime parameters of a unit. This is primarily
2139 useful to alter cgroup parameters dynamically in a nice way,
2140 but will be extended later on to make more properties
2141 modifiable at runtime. systemctl gained a new set-properties
2142 command that wraps this call.
2144 * A new tool "systemd-run" has been added which can be used to
2145 run arbitrary command lines as transient services or scopes,
2146 while configuring a number of settings via the command
2147 line. This tool is currently very basic, however already
2148 very useful. We plan to extend this tool to even allow
2149 queuing of execution jobs with time triggers from the
2150 command line, similar in fashion to "at".
2152 * nspawn will now inform the user explicitly that kernels with
2153 audit enabled break containers, and suggest the user to turn
2156 * Support for detecting the IMA and AppArmor security
2157 frameworks with ConditionSecurity= has been added.
2159 * journalctl gained a new "-k" switch for showing only kernel
2160 messages, mimicking dmesg output; in addition to "--user"
2161 and "--system" switches for showing only user's own logs
2164 * systemd-delta can now show information about drop-in
2165 snippets extending unit files.
2167 * libsystemd-bus has been substantially updated but is still
2168 not available as public API.
2170 * systemd will now look for the "debug" argument on the kernel
2171 command line and enable debug logging, similar to what
2172 "systemd.log_level=debug" already did before.
2174 * "systemctl set-default", "systemctl get-default" has been
2175 added to configure the default.target symlink, which
2176 controls what to boot into by default.
2178 * "systemctl set-log-level" has been added as a convenient
2179 way to raise and lower systemd logging threshold.
2181 * "systemd-analyze plot" will now show the time the various
2182 generators needed for execution, as well as information
2183 about the unit file loading.
2185 * libsystemd-journal gained a new sd_journal_open_files() call
2186 for opening specific journal files. journactl also gained a
2187 new switch to expose this new functionality. Previously we
2188 only supported opening all files from a directory, or all
2189 files from the system, as opening individual files only is
2190 racy due to journal file rotation.
2192 * systemd gained the new DefaultEnvironment= setting in
2193 /etc/systemd/system.conf to set environment variables for
2196 * If a privileged process logs a journal message with the
2197 OBJECT_PID= field set, then journald will automatically
2198 augment this with additional OBJECT_UID=, OBJECT_GID=,
2199 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
2200 system services want to log events about specific client
2201 processes. journactl/systemctl has been updated to make use
2202 of this information if all log messages regarding a specific
2205 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
2206 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
2207 Reisner, David Coppa, David King, David Strauss, Eelco
2208 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
2209 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
2210 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
2211 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
2212 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
2213 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
2214 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
2215 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
2216 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
2217 Łukasz Stelmach, 장동준
2221 * The Python bindings gained some minimal support for the APIs
2222 exposed by libsystemd-logind.
2224 * ConditionSecurity= gained support for detecting SMACK. Since
2225 this condition already supports SELinux and AppArmor we only
2226 miss IMA for this. Patches welcome!
2228 Contributions from: Karol Lewandowski, Lennart Poettering,
2229 Zbigniew Jędrzejewski-Szmek
2233 * systemd-nspawn will now create /etc/resolv.conf if
2234 necessary, before bind-mounting the host's file onto it.
2236 * systemd-nspawn will now store meta information about a
2237 container on the container's cgroup as extended attribute
2238 fields, including the root directory.
2240 * The cgroup hierarchy has been reworked in many ways. All
2241 objects any of the components systemd creates in the cgroup
2242 tree are now suffixed. More specifically, user sessions are
2243 now placed in cgroups suffixed with ".session", users in
2244 cgroups suffixed with ".user", and nspawn containers in
2245 cgroups suffixed with ".nspawn". Furthermore, all cgroup
2246 names are now escaped in a simple scheme to avoid collision
2247 of userspace object names with kernel filenames. This work
2248 is preparation for making these objects relocatable in the
2249 cgroup tree, in order to allow easy resource partitioning of
2250 these objects without causing naming conflicts.
2252 * systemctl list-dependencies gained the new switches
2253 --plain, --reverse, --after and --before.
2255 * systemd-inhibit now shows the process name of processes that
2256 have taken an inhibitor lock.
2258 * nss-myhostname will now also resolve "localhost"
2259 implicitly. This makes /etc/hosts an optional file and
2260 nicely handles that on IPv6 ::1 maps to both "localhost" and
2263 * libsystemd-logind.so gained a new call
2264 sd_get_machine_names() to enumerate running containers and
2265 VMs (currently only supported by very new libvirt and
2266 nspawn). sd_login_monitor can now be used to watch
2267 VMs/containers coming and going.
2269 * .include is not allowed recursively anymore, and only in
2270 unit files. Usually it is better to use drop-in snippets in
2271 .d/*.conf anyway, as introduced with systemd 198.
2273 * systemd-analyze gained a new "critical-chain" command that
2274 determines the slowest chain of units run during system
2275 boot-up. It is very useful for tracking down where
2276 optimizing boot time is the most beneficial.
2278 * systemd will no longer allow manipulating service paths in
2279 the name=systemd:/system cgroup tree using ControlGroup= in
2280 units. (But is still fine with it in all other dirs.)
2282 * There's a new systemd-nspawn@.service service file that may
2283 be used to easily run nspawn containers as system
2284 services. With the container's root directory in
2285 /var/lib/container/foobar it is now sufficient to run
2286 "systemctl start systemd-nspawn@foobar.service" to boot it.
2288 * systemd-cgls gained a new parameter "--machine" to list only
2289 the processes within a certain container.
2291 * ConditionSecurity= now can check for "apparmor". We still
2292 are lacking checks for SMACK and IMA for this condition
2293 check though. Patches welcome!
2295 * A new configuration file /etc/systemd/sleep.conf has been
2296 added that may be used to configure which kernel operation
2297 systemd is supposed to execute when "suspend", "hibernate"
2298 or "hybrid-sleep" is requested. This makes the new kernel
2299 "freeze" state accessible to the user.
2301 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
2302 the passed argument if applicable.
2304 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
2305 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
2306 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
2307 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
2308 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
2309 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
2310 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
2315 * The output of 'systemctl list-jobs' got some polishing. The
2316 '--type=' argument may now be passed more than once. A new
2317 command 'systemctl list-sockets' has been added which shows
2318 a list of kernel sockets systemd is listening on with the
2319 socket units they belong to, plus the units these socket
2322 * The experimental libsystemd-bus library got substantial
2323 updates to work in conjunction with the (also experimental)
2324 kdbus kernel project. It works well enough to exchange
2325 messages with some sophistication. Note that kdbus is not
2326 ready yet, and the library is mostly an elaborate test case
2327 for now, and not installable.
2329 * systemd gained a new unit 'systemd-static-nodes.service'
2330 that generates static device nodes earlier during boot, and
2331 can run in conjunction with udev.
2333 * libsystemd-login gained a new call sd_pid_get_user_unit()
2334 to retrieve the user systemd unit a process is running
2335 in. This is useful for systems where systemd is used as
2338 * systemd-nspawn now places all containers in the new /machine
2339 top-level cgroup directory in the name=systemd
2340 hierarchy. libvirt will soon do the same, so that we get a
2341 uniform separation of /system, /user and /machine for system
2342 services, user processes and containers/virtual
2343 machines. This new cgroup hierarchy is also useful to stick
2344 stable names to specific container instances, which can be
2345 recognized later this way (this name may be controlled
2346 via systemd-nspawn's new -M switch). libsystemd-login also
2347 gained a new call sd_pid_get_machine_name() to retrieve the
2348 name of the container/VM a specific process belongs to.
2350 * bootchart can now store its data in the journal.
2352 * libsystemd-journal gained a new call
2353 sd_journal_add_conjunction() for AND expressions to the
2354 matching logic. This can be used to express more complex
2355 logical expressions.
2357 * journactl can now take multiple --unit= and --user-unit=
2360 * The cryptsetup logic now understands the "luks.key=" kernel
2361 command line switch for specifying a file to read the
2362 decryption key from. Also, if a configured key file is not
2363 found the tool will now automatically fall back to prompting
2366 * Python systemd.journal module was updated to wrap recently
2367 added functions from libsystemd-journal. The interface was
2368 changed to bring the low level interface in s.j._Reader
2369 closer to the C API, and the high level interface in
2370 s.j.Reader was updated to wrap and convert all data about
2373 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
2374 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
2375 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
2376 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
2377 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
2378 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
2382 * journalctl --update-catalog now understands a new --root=
2383 option to operate on catalogs found in a different root
2386 * During shutdown after systemd has terminated all running
2387 services a final killing loop kills all remaining left-over
2388 processes. We will now print the name of these processes
2389 when we send SIGKILL to them, since this usually indicates a
2392 * If /etc/crypttab refers to password files stored on
2393 configured mount points automatic dependencies will now be
2394 generated to ensure the specific mount is established first
2395 before the key file is attempted to be read.
2397 * 'systemctl status' will now show information about the
2398 network sockets a socket unit is listening on.
2400 * 'systemctl status' will also shown information about any
2401 drop-in configuration file for units. (Drop-In configuration
2402 files in this context are files such as
2403 /etc/systemd/systemd/foobar.service.d/*.conf)
2405 * systemd-cgtop now optionally shows summed up CPU times of
2406 cgroups. Press '%' while running cgtop to switch between
2407 percentage and absolute mode. This is useful to determine
2408 which cgroups use up the most CPU time over the entire
2409 runtime of the system. systemd-cgtop has also been updated
2410 to be 'pipeable' for processing with further shell tools.
2412 * 'hostnamectl set-hostname' will now allow setting of FQDN
2415 * The formatting and parsing of time span values has been
2416 changed. The parser now understands fractional expressions
2417 such as "5.5h". The formatter will now output fractional
2418 expressions for all time spans under 1min, i.e. "5.123456s"
2419 rather than "5s 123ms 456us". For time spans under 1s
2420 millisecond values are shown, for those under 1ms
2421 microsecond values are shown. This should greatly improve
2422 all time-related output of systemd.
2424 * libsystemd-login and libsystemd-journal gained new
2425 functions for querying the poll() events mask and poll()
2426 timeout value for integration into arbitrary event
2429 * localectl gained the ability to list available X11 keymaps
2430 (models, layouts, variants, options).
2432 * 'systemd-analyze dot' gained the ability to filter for
2433 specific units via shell-style globs, to create smaller,
2434 more useful graphs. I.e. it is now possible to create simple
2435 graphs of all the dependencies between only target units, or
2436 of all units that Avahi has dependencies with.
2438 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
2439 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
2440 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
2441 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
2442 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
2443 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
2444 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
2448 * The boot-time readahead implementation for rotating media
2449 will now read the read-ahead data in multiple passes which
2450 consist of all read requests made in equidistant time
2451 intervals. This means instead of strictly reading read-ahead
2452 data in its physical order on disk we now try to find a
2453 middle ground between physical and access time order.
2455 * /etc/os-release files gained a new BUILD_ID= field for usage
2456 on operating systems that provide continuous builds of OS
2459 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
2460 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
2461 William Douglas, Zbigniew Jędrzejewski-Szmek
2465 * systemd-python gained an API exposing libsystemd-daemon.
2467 * The SMACK setup logic gained support for uploading CIPSO
2470 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
2471 ReadOnlyDirectories= and InaccessibleDirectories= has
2472 changed. The private /tmp and /var/tmp directories are now
2473 shared by all processes of a service (which means
2474 ExecStartPre= may now leave data in /tmp that ExecStart= of
2475 the same service can still access). When a service is
2476 stopped its temporary directories are immediately deleted
2477 (normal clean-up with tmpfiles is still done in addition to
2480 * By default, systemd will now set a couple of sysctl
2481 variables in the kernel: the safe sysrq options are turned
2482 on, IP route verification is turned on, and source routing
2483 disabled. The recently added hardlink and softlink
2484 protection of the kernel is turned on. These settings should
2485 be reasonably safe, and good defaults for all new systems.
2487 * The predictable network naming logic may now be turned off
2488 with a new kernel command line switch: net.ifnames=0.
2490 * A new libsystemd-bus module has been added that implements a
2491 pretty complete D-Bus client library. For details see:
2493 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
2495 * journald will now explicitly flush the journal files to disk
2496 at the latest 5min after each write. The file will then also
2497 be marked offline until the next write. This should increase
2498 reliability in case of a crash. The synchronization delay
2499 can be configured via SyncIntervalSec= in journald.conf.
2501 * There's a new remote-fs-setup.target unit that can be used
2502 to pull in specific services when at least one remote file
2503 system is to be mounted.
2505 * There are new targets timers.target and paths.target as
2506 canonical targets to pull user timer and path units in
2507 from. This complements sockets.target with a similar
2508 purpose for socket units.
2510 * libudev gained a new call udev_device_set_attribute_value()
2511 to set sysfs attributes of a device.
2513 * The udev daemon now sets the default number of worker
2514 processes executed in parallel based on the number of available
2515 CPUs instead of the amount of available RAM. This is supposed
2516 to provide a more reliable default and limit a too aggressive
2517 paralellism for setups with 1000s of devices connected.
2519 Contributions from: Auke Kok, Colin Walters, Cristian
2520 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
2521 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
2522 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
2523 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
2524 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
2525 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
2526 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
2527 Zbigniew Jędrzejewski-Szmek
2531 * Configuration of unit files may now be extended via drop-in
2532 files without having to edit/override the unit files
2533 themselves. More specifically, if the administrator wants to
2534 change one value for a service file foobar.service he can
2535 now do so by dropping in a configuration snippet into
2536 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
2537 will load all these snippets and apply them on top of the
2538 main unit configuration file, possibly extending or
2539 overriding its settings. Using these drop-in snippets is
2540 generally nicer than the two earlier options for changing
2541 unit files locally: copying the files from
2542 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
2543 them there; or creating a new file in /etc/systemd/system/
2544 that incorporates the original one via ".include". Drop-in
2545 snippets into these .d/ directories can be placed in any
2546 directory systemd looks for units in, and the usual
2547 overriding semantics between /usr/lib, /etc and /run apply
2550 * Most unit file settings which take lists of items can now be
2551 reset by assigning the empty string to them. For example,
2552 normally, settings such as Environment=FOO=BAR append a new
2553 environment variable assignment to the environment block,
2554 each time they are used. By assigning Environment= the empty
2555 string the environment block can be reset to empty. This is
2556 particularly useful with the .d/*.conf drop-in snippets
2557 mentioned above, since this adds the ability to reset list
2558 settings from vendor unit files via these drop-ins.
2560 * systemctl gained a new "list-dependencies" command for
2561 listing the dependencies of a unit recursively.
2563 * Inhibitors are now honored and listed by "systemctl
2564 suspend", "systemctl poweroff" (and similar) too, not only
2565 GNOME. These commands will also list active sessions by
2568 * Resource limits (as exposed by the various control group
2569 controllers) can now be controlled dynamically at runtime
2570 for all units. More specifically, you can now use a command
2571 like "systemctl set-cgroup-attr foobar.service cpu.shares
2572 2000" to alter the CPU shares a specific service gets. These
2573 settings are stored persistently on disk, and thus allow the
2574 administrator to easily adjust the resource usage of
2575 services with a few simple commands. This dynamic resource
2576 management logic is also available to other programs via the
2577 bus. Almost any kernel cgroup attribute and controller is
2580 * systemd-vconsole-setup will now copy all font settings to
2581 all allocated VTs, where it previously applied them only to
2584 * libsystemd-login gained the new sd_session_get_tty() API
2587 * This release drops support for a few legacy or
2588 distribution-specific LSB facility names when parsing init
2589 scripts: $x-display-manager, $mail-transfer-agent,
2590 $mail-transport-agent, $mail-transfer-agent, $smtp,
2591 $null. Also, the mail-transfer-agent.target unit backing
2592 this has been removed. Distributions which want to retain
2593 compatibility with this should carry the burden for
2594 supporting this themselves and patch support for these back
2595 in, if they really need to. Also, the facilities $syslog and
2596 $local_fs are now ignored, since systemd does not support
2597 early-boot LSB init scripts anymore, and these facilities
2598 are implied anyway for normal services. syslog.target has
2601 * There are new bus calls on PID1's Manager object for
2602 cancelling jobs, and removing snapshot units. Previously,
2603 both calls were only available on the Job and Snapshot
2606 * systemd-journal-gatewayd gained SSL support.
2608 * The various "environment" files, such as /etc/locale.conf
2609 now support continuation lines with a backslash ("\") as
2610 last character in the line, similarly in style (but different)
2611 to how this is supported in shells.
2613 * For normal user processes the _SYSTEMD_USER_UNIT= field is
2614 now implicitly appended to every log entry logged. systemctl
2615 has been updated to filter by this field when operating on a
2616 user systemd instance.
2618 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
2619 CAP_AUDIT_CONTROL capabilities to the capabilities set for
2620 the container. This makes it easier to boot unmodified
2621 Fedora systems in a container, which however still requires
2622 audit=0 to be passed on the kernel command line. Auditing in
2623 kernel and userspace is unfortunately still too broken in
2624 context of containers, hence we recommend compiling it out
2625 of the kernel or using audit=0. Hopefully this will be fixed
2626 one day for good in the kernel.
2628 * nspawn gained the new --bind= and --bind-ro= parameters to
2629 bind mount specific directories from the host into the
2632 * nspawn will now mount its own devpts file system instance
2633 into the container, in order not to leak pty devices from
2634 the host into the container.
2636 * systemd will now read the firmware boot time performance
2637 information from the EFI variables, if the used boot loader
2638 supports this, and takes it into account for boot performance
2639 analysis via "systemd-analyze". This is currently supported
2640 only in conjunction with Gummiboot, but could be supported
2641 by other boot loaders too. For details see:
2643 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
2645 * A new generator has been added that automatically mounts the
2646 EFI System Partition (ESP) to /boot, if that directory
2647 exists, is empty, and no other file system has been
2648 configured to be mounted there.
2650 * logind will now send out PrepareForSleep(false) out
2651 unconditionally, after coming back from suspend. This may be
2652 used by applications as asynchronous notification for
2653 system resume events.
2655 * "systemctl unlock-sessions" has been added, that allows
2656 unlocking the screens of all user sessions at once, similar
2657 to how "systemctl lock-sessions" already locked all users
2658 sessions. This is backed by a new D-Bus call UnlockSessions().
2660 * "loginctl seat-status" will now show the master device of a
2661 seat. (i.e. the device of a seat that needs to be around for
2662 the seat to be considered available, usually the graphics
2665 * tmpfiles gained a new "X" line type, that allows
2666 configuration of files and directories (with wildcards) that
2667 shall be excluded from automatic cleanup ("aging").
2669 * udev default rules set the device node permissions now only
2670 at "add" events, and do not change them any longer with a
2671 later "change" event.
2673 * The log messages for lid events and power/sleep keypresses
2674 now carry a message ID.
2676 * We now have a substantially larger unit test suite, but this
2677 continues to be work in progress.
2679 * udevadm hwdb gained a new --root= parameter to change the
2680 root directory to operate relative to.
2682 * logind will now issue a background sync() request to the kernel
2683 early at shutdown, so that dirty buffers are flushed to disk early
2684 instead of at the last moment, in order to optimize shutdown
2687 * A new bootctl tool has been added that is an interface for
2688 certain boot loader operations. This is currently a preview
2689 and is likely to be extended into a small mechanism daemon
2690 like timedated, localed, hostnamed, and can be used by
2691 graphical UIs to enumerate available boot options, and
2692 request boot into firmware operations.
2694 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
2695 the rest of the package. It also has been updated to work
2696 correctly in initrds.
2698 * Policykit previously has been runtime optional, and is now
2699 also compile time optional via a configure switch.
2701 * systemd-analyze has been reimplemented in C. Also "systemctl
2702 dot" has moved into systemd-analyze.
2704 * "systemctl status" with no further parameters will now print
2705 the status of all active or failed units.
2707 * Operations such as "systemctl start" can now be executed
2708 with a new mode "--irreversible" which may be used to queue
2709 operations that cannot accidentally be reversed by a later
2710 job queuing. This is by default used to make shutdown
2711 requests more robust.
2713 * The Python API of systemd now gained a new module for
2714 reading journal files.
2716 * A new tool kernel-install has been added that can install
2717 kernel images according to the Boot Loader Specification:
2719 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
2721 * Boot time console output has been improved to provide
2722 animated boot time output for hanging jobs.
2724 * A new tool systemd-activate has been added which can be used
2725 to test socket activation with, directly from the command
2726 line. This should make it much easier to test and debug
2727 socket activation in daemons.
2729 * journalctl gained a new "--reverse" (or -r) option to show
2730 journal output in reverse order (i.e. newest line first).
2732 * journalctl gained a new "--pager-end" (or -e) option to jump
2733 to immediately jump to the end of the journal in the
2734 pager. This is only supported in conjunction with "less".
2736 * journalctl gained a new "--user-unit=" option, that works
2737 similarly to "--unit=" but filters for user units rather than
2740 * A number of unit files to ease adoption of systemd in
2741 initrds has been added. This moves some minimal logic from
2742 the various initrd implementations into systemd proper.
2744 * The journal files are now owned by a new group
2745 "systemd-journal", which exists specifically to allow access
2746 to the journal, and nothing else. Previously, we used the
2747 "adm" group for that, which however possibly covers more
2748 than just journal/log file access. This new group is now
2749 already used by systemd-journal-gatewayd to ensure this
2750 daemon gets access to the journal files and as little else
2751 as possible. Note that "make install" will also set FS ACLs
2752 up for /var/log/journal to give "adm" and "wheel" read
2753 access to it, in addition to "systemd-journal" which owns
2754 the journal files. We recommend that packaging scripts also
2755 add read access to "adm" + "wheel" to /var/log/journal, and
2756 all existing/future journal files. To normal users and
2757 administrators little changes, however packagers need to
2758 ensure to create the "systemd-journal" system group at
2759 package installation time.
2761 * The systemd-journal-gatewayd now runs as unprivileged user
2762 systemd-journal-gateway:systemd-journal-gateway. Packaging
2763 scripts need to create these system user/group at
2766 * timedated now exposes a new boolean property CanNTP that
2767 indicates whether a local NTP service is available or not.
2769 * systemd-detect-virt will now also detect xen PVs
2771 * The pstore file system is now mounted by default, if it is
2774 * In addition to the SELinux and IMA policies we will now also
2775 load SMACK policies at early boot.
2777 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
2778 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
2779 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
2780 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
2781 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
2782 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
2783 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
2784 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
2785 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
2786 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
2787 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
2788 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
2789 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
2790 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
2794 * Timer units now support calendar time events in addition to
2795 monotonic time events. That means you can now trigger a unit
2796 based on a calendar time specification such as "Thu,Fri
2797 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
2798 or fifth day of any month of the year 2013, given that it is
2799 a thursday or friday. This brings timer event support
2800 considerably closer to cron's capabilities. For details on
2801 the supported calendar time specification language see
2804 * udev now supports a number of different naming policies for
2805 network interfaces for predictable names, and a combination
2806 of these policies is now the default. Please see this wiki
2807 document for details:
2809 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
2811 * Auke Kok's bootchart implementation has been added to the
2812 systemd tree. It is an optional component that can graph the
2813 boot in quite some detail. It is one of the best bootchart
2814 implementations around and minimal in its code and
2817 * nss-myhostname has been integrated into the systemd source
2818 tree. nss-myhostname guarantees that the local hostname
2819 always stays resolvable via NSS. It has been a weak
2820 requirement of systemd-hostnamed since a long time, and
2821 since its code is actually trivial we decided to just
2822 include it in systemd's source tree. It can be turned off
2823 with a configure switch.
2825 * The read-ahead logic is now capable of properly detecting
2826 whether a btrfs file system is on SSD or rotating media, in
2827 order to optimize the read-ahead scheme. Previously, it was
2828 only capable of detecting this on traditional file systems
2831 * In udev, additional device properties are now read from the
2832 IAB in addition to the OUI database. Also, Bluetooth company
2833 identities are attached to the devices as well.
2835 * In service files %U may be used as specifier that is
2836 replaced by the configured user name of the service.
2838 * nspawn may now be invoked without a controlling TTY. This
2839 makes it suitable for invocation as its own service. This
2840 may be used to set up a simple containerized server system
2841 using only core OS tools.
2843 * systemd and nspawn can now accept socket file descriptors
2844 when they are started for socket activation. This enables
2845 implementation of socket activated nspawn
2846 containers. i.e. think about autospawning an entire OS image
2847 when the first SSH or HTTP connection is received. We expect
2848 that similar functionality will also be added to libvirt-lxc
2851 * journalctl will now suppress ANSI color codes when
2852 presenting log data.
2854 * systemctl will no longer show control group information for
2855 a unit if a the control group is empty anyway.
2857 * logind can now automatically suspend/hibernate/shutdown the
2860 * /etc/machine-info and hostnamed now also expose the chassis
2861 type of the system. This can be used to determine whether
2862 the local system is a laptop, desktop, handset or
2863 tablet. This information may either be configured by the
2864 user/vendor or is automatically determined from ACPI and DMI
2865 information if possible.
2867 * A number of PolicyKit actions are now bound together with
2868 "imply" rules. This should simplify creating UIs because
2869 many actions will now authenticate similar ones as well.
2871 * Unit files learnt a new condition ConditionACPower= which
2872 may be used to conditionalize a unit depending on whether an
2873 AC power source is connected or not, of whether the system
2874 is running on battery power.
2876 * systemctl gained a new "is-failed" verb that may be used in
2877 shell scripts and suchlike to check whether a specific unit
2878 is in the "failed" state.
2880 * The EnvironmentFile= setting in unit files now supports file
2881 globbing, and can hence be used to easily read a number of
2882 environment files at once.
2884 * systemd will no longer detect and recognize specific
2885 distributions. All distribution-specific #ifdeffery has been
2886 removed, systemd is now fully generic and
2887 distribution-agnostic. Effectively, not too much is lost as
2888 a lot of the code is still accessible via explicit configure
2889 switches. However, support for some distribution specific
2890 legacy configuration file formats has been dropped. We
2891 recommend distributions to simply adopt the configuration
2892 files everybody else uses now and convert the old
2893 configuration from packaging scripts. Most distributions
2894 already did that. If that's not possible or desirable,
2895 distributions are welcome to forward port the specific
2896 pieces of code locally from the git history.
2898 * When logging a message about a unit systemd will now always
2899 log the unit name in the message meta data.
2901 * localectl will now also discover system locale data that is
2902 not stored in locale archives, but directly unpacked.
2904 * logind will no longer unconditionally use framebuffer
2905 devices as seat masters, i.e. as devices that are required
2906 to be existing before a seat is considered preset. Instead,
2907 it will now look for all devices that are tagged as
2908 "seat-master" in udev. By default framebuffer devices will
2909 be marked as such, but depending on local systems other
2910 devices might be marked as well. This may be used to
2911 integrate graphics cards using closed source drivers (such
2912 as NVidia ones) more nicely into logind. Note however, that
2913 we recommend using the open source NVidia drivers instead,
2914 and no udev rules for the closed-source drivers will be
2915 shipped from us upstream.
2917 Contributions from: Adam Williamson, Alessandro Crismani, Auke
2918 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
2919 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
2920 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
2921 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
2922 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
2923 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
2924 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
2925 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
2926 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
2927 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
2928 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
2933 * udev gained support for loading additional device properties
2934 from an indexed database that is keyed by vendor/product IDs
2935 and similar device identifiers. For the beginning this
2936 "hwdb" is populated with data from the well-known PCI and
2937 USB database, but also includes PNP, ACPI and OID data. In
2938 the longer run this indexed database shall grow into
2939 becoming the one central database for non-essential
2940 userspace device metadata. Previously, data from the PCI/USB
2941 database was only attached to select devices, since the
2942 lookup was a relatively expensive operation due to O(n) time
2943 complexity (with n being the number of entries in the
2944 database). Since this is now O(1), we decided to add in this
2945 data for all devices where this is available, by
2946 default. Note that the indexed database needs to be rebuilt
2947 when new data files are installed. To achieve this you need
2948 to update your packaging scripts to invoke "udevadm hwdb
2949 --update" after installation of hwdb data files. For
2950 RPM-based distributions we introduced the new
2951 %udev_hwdb_update macro for this purpose.
2953 * The Journal gained support for the "Message Catalog", an
2954 indexed database to link up additional information with
2955 journal entries. For further details please check:
2957 http://www.freedesktop.org/wiki/Software/systemd/catalog
2959 The indexed message catalog database also needs to be
2960 rebuilt after installation of message catalog files. Use
2961 "journalctl --update-catalog" for this. For RPM-based
2962 distributions we introduced the %journal_catalog_update
2963 macro for this purpose.
2965 * The Python Journal bindings gained support for the standard
2966 Python logging framework.
2968 * The Journal API gained new functions for checking whether
2969 the underlying file system of a journal file is capable of
2970 properly reporting file change notifications, or whether
2971 applications that want to reflect journal changes "live"
2972 need to recheck journal files continuously in appropriate
2975 * It is now possible to set the "age" field for tmpfiles
2976 entries to 0, indicating that files matching this entry
2977 shall always be removed when the directories are cleaned up.
2979 * coredumpctl gained a new "gdb" verb which invokes gdb
2980 right-away on the selected coredump.
2982 * There's now support for "hybrid sleep" on kernels that
2983 support this, in addition to "suspend" and "hibernate". Use
2984 "systemctl hybrid-sleep" to make use of this.
2986 * logind's HandleSuspendKey= setting (and related settings)
2987 now gained support for a new "lock" setting to simply
2988 request the screen lock on all local sessions, instead of
2989 actually executing a suspend or hibernation.
2991 * systemd will now mount the EFI variables file system by
2994 * Socket units now gained support for configuration of the
2995 SMACK security label.
2997 * timedatectl will now output the time of the last and next
2998 daylight saving change.
3000 * We dropped support for various legacy and distro-specific
3001 concepts, such as insserv, early-boot SysV services
3002 (i.e. those for non-standard runlevels such as 'b' or 'S')
3003 or ArchLinux /etc/rc.conf support. We recommend the
3004 distributions who still need support this to either continue
3005 to maintain the necessary patches downstream, or find a
3006 different solution. (Talk to us if you have questions!)
3008 * Various systemd components will now bypass PolicyKit checks
3009 for root and otherwise handle properly if PolicyKit is not
3010 found to be around. This should fix most issues for
3011 PolicyKit-less systems. Quite frankly this should have been
3012 this way since day one. It is absolutely our intention to
3013 make systemd work fine on PolicyKit-less systems, and we
3014 consider it a bug if something does not work as it should if
3015 PolicyKit is not around.
3017 * For embedded systems it is now possible to build udev and
3018 systemd without blkid and/or kmod support.
3020 * "systemctl switch-root" is now capable of switching root
3021 more than once. I.e. in addition to transitions from the
3022 initrd to the host OS it is now possible to transition to
3023 further OS images from the host. This is useful to implement
3024 offline updating tools.
3026 * Various other additions have been made to the RPM macros
3027 shipped with systemd. Use %udev_rules_update() after
3028 installing new udev rules files. %_udevhwdbdir,
3029 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
3030 %_sysctldir are now available which resolve to the right
3031 directories for packages to place various data files in.
3033 * journalctl gained the new --full switch (in addition to
3034 --all, to disable ellipsation for long messages.
3036 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
3037 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
3038 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
3039 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
3040 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
3041 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
3042 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
3043 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
3044 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
3048 * journalctl gained new --since= and --until= switches to
3049 filter by time. It also now supports nice filtering for
3050 units via --unit=/-u.
3052 * Type=oneshot services may use ExecReload= and do the
3055 * The journal daemon now supports time-based rotation and
3056 vacuuming, in addition to the usual disk-space based
3059 * The journal will now index the available field values for
3060 each field name. This enables clients to show pretty drop
3061 downs of available match values when filtering. The bash
3062 completion of journalctl has been updated
3063 accordingly. journalctl gained a new switch -F to list all
3064 values a certain field takes in the journal database.
3066 * More service events are now written as structured messages
3067 to the journal, and made recognizable via message IDs.
3069 * The timedated, localed and hostnamed mini-services which
3070 previously only provided support for changing time, locale
3071 and hostname settings from graphical DEs such as GNOME now
3072 also have a minimal (but very useful) text-based client
3073 utility each. This is probably the nicest way to changing
3074 these settings from the command line now, especially since
3075 it lists available options and is fully integrated with bash
3078 * There's now a new tool "systemd-coredumpctl" to list and
3079 extract coredumps from the journal.
3081 * We now install a README each in /var/log/ and
3082 /etc/rc.d/init.d explaining where the system logs and init
3083 scripts went. This hopefully should help folks who go to
3084 that dirs and look into the otherwise now empty void and
3085 scratch their heads.
3087 * When user-services are invoked (by systemd --user) the
3088 $MANAGERPID env var is set to the PID of systemd.
3090 * SIGRTMIN+24 when sent to a --user instance will now result
3091 in immediate termination of systemd.
3093 * gatewayd received numerous feature additions such as a
3094 "follow" mode, for live syncing and filtering.
3096 * browse.html now allows filtering and showing detailed
3097 information on specific entries. Keyboard navigation and
3098 mouse screen support has been added.
3100 * gatewayd/journalctl now supports HTML5/JSON
3101 Server-Sent-Events as output.
3103 * The SysV init script compatibility logic will now
3104 heuristically determine whether a script supports the
3105 "reload" verb, and only then make this available as
3108 * "systemctl status --follow" has been removed, use "journalctl
3111 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
3112 have been removed since they are hardly useful to be
3115 * And I'd like to take the opportunity to specifically mention
3116 Zbigniew for his great contributions. Zbigniew, you rock!
3118 Contributions from: Andrew Eikum, Christian Hesse, Colin
3119 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
3120 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
3121 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
3122 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
3123 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
3124 Jędrzejewski-Szmek, Сковорода Никита Андреевич
3128 * If /etc/vconsole.conf is non-existent or empty we will no
3129 longer load any console font or key map at boot by
3130 default. Instead the kernel defaults will be left
3131 intact. This is definitely the right thing to do, as no
3132 configuration should mean no configuration, and hard-coding
3133 font names that are different on all archs is probably a bad
3134 idea. Also, the kernel default key map and font should be
3135 good enough for most cases anyway, and mostly identical to
3136 the userspace fonts/key maps we previously overloaded them
3137 with. If distributions want to continue to default to a
3138 non-kernel font or key map they should ship a default
3139 /etc/vconsole.conf with the appropriate contents.
3141 Contributions from: Colin Walters, Daniel J Walsh, Dave
3142 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
3143 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
3147 * journalctl gained a new --cursor= switch to show entries
3148 starting from the specified location in the journal.
3150 * We now enforce a size limit on journal entry fields exported
3151 with "-o json" in journalctl. Fields larger than 4K will be
3152 assigned null. This can be turned off with --all.
3154 * An (optional) journal gateway daemon is now available as
3155 "systemd-journal-gatewayd.service". This service provides
3156 access to the journal via HTTP and JSON. This functionality
3157 will be used to implement live log synchronization in both
3158 pull and push modes, but has various other users too, such
3159 as easy log access for debugging of embedded devices. Right
3160 now it is already useful to retrieve the journal via HTTP:
3162 # systemctl start systemd-journal-gatewayd.service
3163 # wget http://localhost:19531/entries
3165 This will download the journal contents in a
3166 /var/log/messages compatible format. The same as JSON:
3168 # curl -H"Accept: application/json" http://localhost:19531/entries
3170 This service is also accessible via a web browser where a
3171 single static HTML5 app is served that uses the JSON logic
3172 to enable the user to do some basic browsing of the
3173 journal. This will be extended later on. Here's an example
3174 screenshot of this app in its current state:
3176 http://0pointer.de/public/journal-gatewayd
3178 Contributions from: Kay Sievers, Lennart Poettering, Robert
3179 Milasan, Tom Gundersen
3183 * The bash completion logic is now available for journalctl
3186 * We do not mount the "cpuset" controller anymore together with
3187 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
3188 started if no parameters are assigned to it. "cpuset" hence
3189 broke code that assumed it it could create "cpu" groups and
3192 * journalctl -f will now subscribe to terminal size changes,
3193 and line break accordingly.
3195 Contributions from: Dave Reisner, Kay Sievers, Lennart
3196 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
3200 * nspawn will now create a symlink /etc/localtime in the
3201 container environment, copying the host's timezone
3202 setting. Previously this has been done via a bind mount, but
3203 since symlinks cannot be bind mounted this has now been
3204 changed to create/update the appropriate symlink.
3206 * journalctl -n's line number argument is now optional, and
3207 will default to 10 if omitted.
3209 * journald will now log the maximum size the journal files may
3210 take up on disk. This is particularly useful if the default
3211 built-in logic of determining this parameter from the file
3212 system size is used. Use "systemctl status
3213 systemd-journald.service" to see this information.
3215 * The multi-seat X wrapper tool has been stripped down. As X
3216 is now capable of enumerating graphics devices via udev in a
3217 seat-aware way the wrapper is not strictly necessary
3218 anymore. A stripped down temporary stop-gap is still shipped
3219 until the upstream display managers have been updated to
3220 fully support the new X logic. Expect this wrapper to be
3221 removed entirely in one of the next releases.
3223 * HandleSleepKey= in logind.conf has been split up into
3224 HandleSuspendKey= and HandleHibernateKey=. The old setting
3225 is not available anymore. X11 and the kernel are
3226 distuingishing between these keys and we should too. This
3227 also means the inhibition lock for these keys has been split
3230 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
3231 Poettering, Lukas Nykryn, Václav Pavlín
3235 * Whenever a unit changes state we will now log this to the
3236 journal and show along the unit's own log output in
3239 * ConditionPathIsMountPoint= can now properly detect bind
3240 mount points too. (Previously, a bind mount of one file
3241 system to another place in the same file system could not be
3242 detected as mount, since they shared struct stat's st_dev
3245 * We will now mount the cgroup controllers cpu, cpuacct,
3246 cpuset and the controllers net_cls, net_prio together by
3249 * nspawn containers will now have a virtualized boot
3250 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
3251 over with a randomized ID at container initialization). This
3252 has the effect of making "journalctl -b" do the right thing
3255 * The JSON output journal serialization has been updated not
3256 to generate "endless" list objects anymore, but rather one
3257 JSON object per line. This is more in line how most JSON
3258 parsers expect JSON objects. The new output mode
3259 "json-pretty" has been added to provide similar output, but
3260 neatly aligned for readability by humans.
3262 * We dropped all explicit sync() invocations in the shutdown
3263 code. The kernel does this implicitly anyway in the kernel
3264 reboot() syscall. halt(8)'s -n option is now a compatibility
3267 * We now support virtualized reboot() in containers, as
3268 supported by newer kernels. We will fall back to exit() if
3269 CAP_SYS_REBOOT is not available to the container. Also,
3270 nspawn makes use of this now and will actually reboot the
3271 container if the containerized OS asks for that.
3273 * journalctl will only show local log output by default
3274 now. Use --merge (-m) to show remote log output, too.
3276 * libsystemd-journal gained the new sd_journal_get_usage()
3277 call to determine the current disk usage of all journal
3278 files. This is exposed in the new "journalctl --disk-usage"
3281 * journald gained a new configuration setting SplitMode= in
3282 journald.conf which may be used to control how user journals
3283 are split off. See journald.conf(5) for details.
3285 * A new condition type ConditionFileNotEmpty= has been added.
3287 * tmpfiles' "w" lines now support file globbing, to write
3288 multiple files at once.
3290 * We added Python bindings for the journal submission
3291 APIs. More Python APIs for a number of selected APIs will
3292 likely follow. Note that we intend to add native bindings
3293 only for the Python language, as we consider it common
3294 enough to deserve bindings shipped within systemd. There are
3295 various projects outside of systemd that provide bindings
3296 for languages such as PHP or Lua.
3298 * Many conditions will now resolve specifiers such as %i. In
3299 addition, PathChanged= and related directives of .path units
3300 now support specifiers as well.
3302 * There's now a new RPM macro definition for the system preset
3305 * journald will now warn if it ca not forward a message to the
3306 syslog daemon because its socket is full.
3308 * timedated will no longer write or process /etc/timezone,
3309 except on Debian. As we do not support late mounted /usr
3310 anymore /etc/localtime always being a symlink is now safe,
3311 and hence the information in /etc/timezone is not necessary
3314 * logind will now always reserve one VT for a text getty (VT6
3315 by default). Previously if more than 6 X sessions where
3316 started they took up all the VTs with auto-spawned gettys,
3317 so that no text gettys were available anymore.
3319 * udev will now automatically inform the btrfs kernel logic
3320 about btrfs RAID components showing up. This should make
3321 simple hotplug based btrfs RAID assembly work.
3323 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
3324 (but not for its children which will stay at the kernel
3325 default). This should allow setups with a lot more listening
3328 * systemd will now always pass the configured timezone to the
3329 kernel at boot. timedated will do the same when the timezone
3332 * logind's inhibition logic has been updated. By default,
3333 logind will now handle the lid switch, the power and sleep
3334 keys all the time, even in graphical sessions. If DEs want
3335 to handle these events on their own they should take the new
3336 handle-power-key, handle-sleep-key and handle-lid-switch
3337 inhibitors during their runtime. A simple way to achiveve
3338 that is to invoke the DE wrapped in an invocation of:
3340 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
3342 * Access to unit operations is now checked via SELinux taking
3343 the unit file label and client process label into account.
3345 * systemd will now notify the administrator in the journal
3346 when he over-mounts a non-empty directory.
3348 * There are new specifiers that are resolved in unit files,
3349 for the host name (%H), the machine ID (%m) and the boot ID
3352 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
3353 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
3354 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
3355 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
3356 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
3357 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
3358 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
3362 * Support for reading structured kernel messages from
3363 /dev/kmsg has now been added and is enabled by default.
3365 * Support for reading kernel messages from /proc/kmsg has now
3366 been removed. If you want kernel messages in the journal
3367 make sure to run a recent kernel (>= 3.5) that supports
3368 reading structured messages from /dev/kmsg (see
3369 above). /proc/kmsg is now exclusive property of classic
3370 syslog daemons again.
3372 * The libudev API gained the new
3373 udev_device_new_from_device_id() call.
3375 * The logic for file system namespace (ReadOnlyDirectory=,
3376 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
3377 require pivot_root() anymore. This means fewer temporary
3378 directories are created below /tmp for this feature.
3380 * nspawn containers will now see and receive all submounts
3381 made on the host OS below the root file system of the
3384 * Forward Secure Sealing is now supported for Journal files,
3385 which provide cryptographical sealing of journal files so
3386 that attackers cannot alter log history anymore without this
3387 being detectable. Lennart will soon post a blog story about
3388 this explaining it in more detail.
3390 * There are two new service settings RestartPreventExitStatus=
3391 and SuccessExitStatus= which allow configuration of exit
3392 status (exit code or signal) which will be excepted from the
3393 restart logic, resp. consider successful.
3395 * journalctl gained the new --verify switch that can be used
3396 to check the integrity of the structure of journal files and
3397 (if Forward Secure Sealing is enabled) the contents of
3400 * nspawn containers will now be run with /dev/stdin, /dev/fd/
3401 and similar symlinks pre-created. This makes running shells
3402 as container init process a lot more fun.
3404 * The fstab support can now handle PARTUUID= and PARTLABEL=
3407 * A new ConditionHost= condition has been added to match
3408 against the hostname (with globs) and machine ID. This is
3409 useful for clusters where a single OS image is used to
3410 provision a large number of hosts which shall run slightly
3411 different sets of services.
3413 * Services which hit the restart limit will now be placed in a
3416 Contributions from: Bertram Poettering, Dave Reisner, Huang
3417 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
3418 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
3422 * When running in --user mode systemd will now become a
3423 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
3424 tree a lot more organized.
3426 * A new PartOf= unit dependency type has been introduced that
3427 may be used to group services in a natural way.
3429 * "systemctl enable" may now be used to enable instances of
3432 * journalctl now prints error log levels in red, and
3433 warning/notice log levels in bright white. It also supports
3434 filtering by log level now.
3436 * cgtop gained a new -n switch (similar to top), to configure
3437 the maximum number of iterations to run for. It also gained
3438 -b, to run in batch mode (accepting no input).
3440 * The suffix ".service" may now be omitted on most systemctl
3441 command lines involving service unit names.
3443 * There's a new bus call in logind to lock all sessions, as
3444 well as a loginctl verb for it "lock-sessions".
3446 * libsystemd-logind.so gained a new call sd_journal_perror()
3447 that works similar to libc perror() but logs to the journal
3448 and encodes structured information about the error number.
3450 * /etc/crypttab entries now understand the new keyfile-size=
3453 * shutdown(8) now can send a (configurable) wall message when
3454 a shutdown is cancelled.
3456 * The mount propagation mode for the root file system will now
3457 default to "shared", which is useful to make containers work
3458 nicely out-of-the-box so that they receive new mounts from
3459 the host. This can be undone locally by running "mount
3460 --make-rprivate /" if needed.
3462 * The prefdm.service file has been removed. Distributions
3463 should maintain this unit downstream if they intend to keep
3464 it around. However, we recommend writing normal unit files
3465 for display managers instead.
3467 * Since systemd is a crucial part of the OS we will now
3468 default to a number of compiler switches that improve
3469 security (hardening) such as read-only relocations, stack
3470 protection, and suchlike.
3472 * The TimeoutSec= setting for services is now split into
3473 TimeoutStartSec= and TimeoutStopSec= to allow configuration
3474 of individual time outs for the start and the stop phase of
3477 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
3478 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
3479 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
3480 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
3481 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
3482 Gundersen, Zbigniew Jędrzejewski-Szmek
3486 * The journal and id128 C APIs are now fully documented as man
3489 * Extra safety checks have been added when transitioning from
3490 the initial RAM disk to the main system to avoid accidental
3493 * /etc/crypttab entries now understand the new keyfile-offset=
3496 * systemctl -t can now be used to filter by unit load state.
3498 * The journal C API gained the new sd_journal_wait() call to
3499 make writing synchronous journal clients easier.
3501 * journalctl gained the new -D switch to show journals from a
3504 * journalctl now displays a special marker between log
3505 messages of two different boots.
3507 * The journal is now explicitly flushed to /var via a service
3508 systemd-journal-flush.service, rather than implicitly simply
3509 by seeing /var/log/journal to be writable.
3511 * journalctl (and the journal C APIs) can now match for much
3512 more complex expressions, with alternatives and
3515 * When transitioning from the initial RAM disk to the main
3516 system we will now kill all processes in a killing spree to
3517 ensure no processes stay around by accident.
3519 * Three new specifiers may be used in unit files: %u, %h, %s
3520 resolve to the user name, user home directory resp. user
3521 shell. This is useful for running systemd user instances.
3523 * We now automatically rotate journal files if their data
3524 object hash table gets a fill level > 75%. We also size the
3525 hash table based on the configured maximum file size. This
3526 together should lower hash collisions drastically and thus
3527 speed things up a bit.
3529 * journalctl gained the new "--header" switch to introspect
3530 header data of journal files.
3532 * A new setting SystemCallFilters= has been added to services
3533 which may be used to apply blacklists or whitelists to
3534 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
3536 * nspawn gained a new --link-journal= switch (and quicker: -j)
3537 to link the container journal with the host. This makes it
3538 very easy to centralize log viewing on the host for all
3539 guests while still keeping the journal files separated.
3541 * Many bugfixes and optimizations
3543 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
3544 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
3545 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
3550 * Several tools now understand kernel command line arguments,
3551 which are only read when run in an initial RAM disk. They
3552 usually follow closely their normal counterparts, but are
3555 * There's a new tool to analyze the readahead files that are
3556 automatically generated at boot. Use:
3558 /usr/lib/systemd/systemd-readahead analyze /.readahead
3560 * We now provide an early debug shell on tty9 if this enabled. Use:
3562 systemctl enable debug-shell.service
3564 * All plymouth related units have been moved into the Plymouth
3565 package. Please make sure to upgrade your Plymouth version
3568 * systemd-tmpfiles now supports getting passed the basename of
3569 a configuration file only, in which case it will look for it
3570 in all appropriate directories automatically.
3572 * udevadm info now takes a /dev or /sys path as argument, and
3573 does the right thing. Example:
3575 udevadm info /dev/sda
3576 udevadm info /sys/class/block/sda
3578 * systemctl now prints a warning if a unit is stopped but a
3579 unit that might trigger it continues to run. Example: a
3580 service is stopped but the socket that activates it is left
3583 * "systemctl status" will now mention if the log output was
3584 shortened due to rotation since a service has been started.
3586 * The journal API now exposes functions to determine the
3587 "cutoff" times due to rotation.
3589 * journald now understands SIGUSR1 and SIGUSR2 for triggering
3590 immediately flushing of runtime logs to /var if possible,
3591 resp. for triggering immediate rotation of the journal
3594 * It is now considered an error if a service is attempted to
3595 be stopped that is not loaded.
3597 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
3599 * systemd-analyze now supports Python 3
3601 * tmpfiles now supports cleaning up directories via aging
3602 where the first level dirs are always kept around but
3603 directories beneath it automatically aged. This is enabled
3604 by prefixing the age field with '~'.
3606 * Seat objects now expose CanGraphical, CanTTY properties
3607 which is required to deal with very fast bootups where the
3608 display manager might be running before the graphics drivers
3609 completed initialization.
3611 * Seat objects now expose a State property.
3613 * We now include RPM macros for service enabling/disabling
3614 based on the preset logic. We recommend RPM based
3615 distributions to make use of these macros if possible. This
3616 makes it simpler to reuse RPM spec files across
3619 * We now make sure that the collected systemd unit name is
3620 always valid when services log to the journal via
3623 * There's a new man page kernel-command-line(7) detailing all
3624 command line options we understand.
3626 * The fstab generator may now be disabled at boot by passing
3627 fstab=0 on the kernel command line.
3629 * A new kernel command line option modules-load= is now understood
3630 to load a specific kernel module statically, early at boot.
3632 * Unit names specified on the systemctl command line are now
3633 automatically escaped as needed. Also, if file system or
3634 device paths are specified they are automatically turned
3635 into the appropriate mount or device unit names. Example:
3637 systemctl status /home
3638 systemctl status /dev/sda
3640 * The SysVConsole= configuration option has been removed from
3641 system.conf parsing.
3643 * The SysV search path is no longer exported on the D-Bus
3646 * The Names= option is been removed from unit file parsing.
3648 * There's a new man page bootup(7) detailing the boot process.
3650 * Every unit and every generator we ship with systemd now
3651 comes with full documentation. The self-explanatory boot is
3654 * A couple of services gained "systemd-" prefixes in their
3655 name if they wrap systemd code, rather than only external
3656 code. Among them fsck@.service which is now
3657 systemd-fsck@.service.
3659 * The HaveWatchdog property has been removed from the D-Bus
3662 * systemd.confirm_spawn= on the kernel command line should now
3665 * There's a new man page crypttab(5) which details all options
3666 we actually understand.
3668 * systemd-nspawn gained a new --capability= switch to pass
3669 additional capabilities to the container.
3671 * timedated will now read known NTP implementation unit names
3672 from /usr/lib/systemd/ntp-units.d/*.list,
3673 systemd-timedated-ntp.target has been removed.
3675 * journalctl gained a new switch "-b" that lists log data of
3676 the current boot only.
3678 * The notify socket is in the abstract namespace again, in
3679 order to support daemons which chroot() at start-up.
3681 * There is a new Storage= configuration option for journald
3682 which allows configuration of where log data should go. This
3683 also provides a way to disable journal logging entirely, so
3684 that data collected is only forwarded to the console, the
3685 kernel log buffer or another syslog implementation.
3687 * Many bugfixes and optimizations
3689 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
3690 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
3691 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
3692 Shawn Landden, Tom Gundersen
3696 * "systemctl help <unit>" now shows the man page if one is
3699 * Several new man pages have been added.
3701 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
3702 MaxLevelConsole= can now be specified in
3703 journald.conf. These options allow reducing the amount of
3704 data stored on disk or forwarded by the log level.
3706 * TimerSlackNSec= can now be specified in system.conf for
3707 PID1. This allows system-wide power savings.
3709 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
3710 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
3715 * logind is now capable of (optionally) handling power and
3716 sleep keys as well as the lid switch.
3718 * journalctl now understands the syntax "journalctl
3719 /usr/bin/avahi-daemon" to get all log output of a specific
3722 * CapabilityBoundingSet= in system.conf now also influences
3723 the capability bound set of usermode helpers of the kernel.
3725 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
3726 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
3727 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
3728 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
3732 * Note that we skipped 139 releases here in order to set the
3733 new version to something that is greater than both udev's
3734 and systemd's most recent version number.
3736 * udev: all udev sources are merged into the systemd source tree now.
3737 All future udev development will happen in the systemd tree. It
3738 is still fully supported to use the udev daemon and tools without
3739 systemd running, like in initramfs or other init systems. Building
3740 udev though, will require the *build* of the systemd tree, but
3741 udev can be properly *run* without systemd.
3743 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
3744 should be used to create dead device nodes as workarounds for broken
3747 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
3748 no longer supported. udev_monitor_new_from_netlink() needs to be
3749 used to subscribe to events.
3751 * udev: when udevd is started by systemd, processes which are left
3752 behind by forking them off of udev rules, are unconditionally cleaned
3753 up and killed now after the event handling has finished. Services or
3754 daemons must be started as systemd services. Services can be
3755 pulled-in by udev to get started, but they can no longer be directly
3756 forked by udev rules.
3758 * udev: the daemon binary is called systemd-udevd now and installed
3759 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
3760 to adapt to that, create symlink, or rename the binary after building
3763 * libudev no longer provides these symbols:
3764 udev_monitor_from_socket()
3765 udev_queue_get_failed_list_entry()
3766 udev_get_{dev,sys,run}_path()
3767 The versions number was bumped and symbol versioning introduced.
3769 * systemd-loginctl and systemd-journalctl have been renamed
3770 to loginctl and journalctl to match systemctl.
3772 * The config files: /etc/systemd/systemd-logind.conf and
3773 /etc/systemd/systemd-journald.conf have been renamed to
3774 logind.conf and journald.conf. Package updates should rename
3775 the files to the new names on upgrade.
3777 * For almost all files the license is now LGPL2.1+, changed
3778 from the previous GPL2.0+. Exceptions are some minor stuff
3779 of udev (which will be changed to LGPL2.1 eventually, too),
3780 and the MIT licensed sd-daemon.[ch] library that is suitable
3781 to be used as drop-in files.
3783 * systemd and logind now handle system sleep states, in
3784 particular suspending and hibernating.
3786 * logind now implements a sleep/shutdown/idle inhibiting logic
3787 suitable for a variety of uses. Soonishly Lennart will blog
3788 about this in more detail.
3790 * var-run.mount and var-lock.mount are no longer provided
3791 (which prevously bind mounted these directories to their new
3792 places). Distributions which have not converted these
3793 directories to symlinks should consider stealing these files
3794 from git history and add them downstream.
3796 * We introduced the Documentation= field for units and added
3797 this to all our shipped units. This is useful to make it
3798 easier to explore the boot and the purpose of the various
3801 * All smaller setup units (such as
3802 systemd-vconsole-setup.service) now detect properly if they
3803 are run in a container and are skipped when
3804 appropriate. This guarantees an entirely noise-free boot in
3805 Linux container environments such as systemd-nspawn.
3807 * A framework for implementing offline system updates is now
3808 integrated, for details see:
3809 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
3811 * A new service type Type=idle is available now which helps us
3812 avoiding ugly interleaving of getty output and boot status
3815 * There's now a system-wide CapabilityBoundingSet= option to
3816 globally reduce the set of capabilities for the
3817 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
3818 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
3819 even CAP_NET_ADMIN system-wide for secure systems.
3821 * There are now system-wide DefaultLimitXXX= options to
3822 globally change the defaults of the various resource limits
3823 for all units started by PID 1.
3825 * Harald Hoyer's systemd test suite has been integrated into
3826 systemd which allows easy testing of systemd builds in qemu
3827 and nspawn. (This is really awesome! Ask us for details!)
3829 * The fstab parser is now implemented as generator, not inside
3832 * systemctl will now warn you if .mount units generated from
3833 /etc/fstab are out of date due to changes in fstab that
3834 have not been read by systemd yet.
3836 * systemd is now suitable for usage in initrds. Dracut has
3837 already been updated to make use of this. With this in place
3838 initrds get a slight bit faster but primarily are much
3839 easier to introspect and debug since "systemctl status" in
3840 the host system can be used to introspect initrd services,
3841 and the journal from the initrd is kept around too.
3843 * systemd-delta has been added, a tool to explore differences
3844 between user/admin configuration and vendor defaults.
3846 * PrivateTmp= now affects both /tmp and /var/tmp.
3848 * Boot time status messages are now much prettier and feature
3849 proper english language. Booting up systemd has never been
3852 * Read-ahead pack files now include the inode number of all
3853 files to pre-cache. When the inode changes the pre-caching
3854 is not attempted. This should be nicer to deal with updated
3855 packages which might result in changes of read-ahead
3858 * We now temporaritly lower the kernel's read_ahead_kb variable
3859 when collecting read-ahead data to ensure the kernel's
3860 built-in read-ahead does not add noise to our measurements
3861 of necessary blocks to pre-cache.
3863 * There's now RequiresMountsFor= to add automatic dependencies
3864 for all mounts necessary for a specific file system path.
3866 * MountAuto= and SwapAuto= have been removed from
3867 system.conf. Mounting file systems at boot has to take place
3870 * nspawn now learned a new switch --uuid= to set the machine
3871 ID on the command line.
3873 * nspawn now learned the -b switch to automatically search
3876 * vt102 is now the default TERM for serial TTYs, upgraded from
3879 * systemd-logind now works on VT-less systems.
3881 * The build tree has been reorganized. The individual
3882 components now have directories of their own.
3884 * A new condition type ConditionPathIsReadWrite= is now available.
3886 * nspawn learned the new -C switch to create cgroups for the
3887 container in other hierarchies.
3889 * We now have support for hardware watchdogs, configurable in
3892 * The scheduled shutdown logic now has a public API.
3894 * We now mount /tmp as tmpfs by default, but this can be
3895 masked and /etc/fstab can override it.
3897 * Since udisks does not make use of /media anymore we are not
3898 mounting a tmpfs on it anymore.
3900 * journalctl gained a new --local switch to only interleave
3901 locally generated journal files.
3903 * We can now load the IMA policy at boot automatically.
3905 * The GTK tools have been split off into a systemd-ui.
3907 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
3908 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
3909 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
3910 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
3911 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
3912 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
3913 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
3914 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
3915 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
3920 * This is mostly a bugfix release
3922 * Support optional initialization of the machine ID from the
3923 KVM or container configured UUID.
3925 * Support immediate reboots with "systemctl reboot -ff"
3927 * Show /etc/os-release data in systemd-analyze output
3929 * Many bugfixes for the journal, including endianness fixes and
3930 ensuring that disk space enforcement works
3932 * sd-login.h is C++ comptaible again
3934 * Extend the /etc/os-release format on request of the Debian
3937 * We now refuse non-UTF8 strings used in various configuration
3938 and unit files. This is done to ensure we do not pass invalid
3939 data over D-Bus or expose it elsewhere.
3941 * Register Mimo USB Screens as suitable for automatic seat
3944 * Read SELinux client context from journal clients in a race
3947 * Reorder configuration file lookup order. /etc now always
3948 overrides /run in order to allow the administrator to always
3949 and unconditionally override vendor supplied or
3950 automatically generated data.
3952 * The various user visible bits of the journal now have man
3953 pages. We still lack man pages for the journal API calls
3956 * We now ship all man pages in HTML format again in the
3959 Contributions from: Dave Reisner, Dirk Eibach, Frederic
3960 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
3961 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
3966 * This is mostly a bugfix release
3968 * systems lacking /etc/os-release are no longer supported.
3970 * Various functionality updates to libsystemd-login.so
3972 * Track class of PAM logins to distuingish greeters from
3975 Contributions from: Kay Sievers, Lennart Poettering, Michael
3980 * This is an important bugfix release for v41.
3982 * Building man pages is now optional which should be useful
3983 for those building systemd from git but unwilling to install
3986 * Watchdog support for supervising services is now usable. In
3987 a future release support for hardware watchdogs
3988 (i.e. /dev/watchdog) will be added building on this.
3990 * Service start rate limiting is now configurable and can be
3991 turned off per service. When a start rate limit is hit a
3992 reboot can automatically be triggered.
3994 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
3996 Contributions from: Benjamin Franzke, Bill Nottingham,
3997 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
3998 Schmidt, Michał Górny, Piotr Drąg
4002 * The systemd binary is installed /usr/lib/systemd/systemd now;
4003 An existing /sbin/init symlink needs to be adapted with the
4006 * The code that loads kernel modules has been ported to invoke
4007 libkmod directly, instead of modprobe. This means we do not
4008 support systems with module-init-tools anymore.
4010 * Watchdog support is now already useful, but still not
4013 * A new kernel command line option systemd.setenv= is
4014 understood to set system wide environment variables
4015 dynamically at boot.
4017 * We now limit the set of capabilities of systemd-journald.
4019 * We now set SIGPIPE to ignore by default, since it only is
4020 useful in shell pipelines, and has little use in general
4021 code. This can be disabled with IgnoreSIPIPE=no in unit
4024 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
4025 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
4030 * This is mostly a bugfix release
4032 * We now expose the reason why a service failed in the
4033 "Result" D-Bus property.
4035 * Rudimentary service watchdog support (will be completed over
4036 the next few releases.)
4038 * When systemd forks off in order execute some service we will
4039 now immediately changes its argv[0] to reflect which process
4040 it will execute. This is useful to minimize the time window
4041 with a generic argv[0], which makes bootcharts more useful
4043 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
4044 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
4045 Mike Kazantsev, Ray Strode
4049 * This is mostly a test release, but incorporates many
4052 * New systemd-cgtop tool to show control groups by their
4055 * Linking against libacl for ACLs is optional again. If
4056 disabled, support tracking device access for active logins
4057 goes becomes unavailable, and so does access to the user
4058 journals by the respective users.
4060 * If a group "adm" exists, journal files are automatically
4061 owned by them, thus allow members of this group full access
4062 to the system journal as well as all user journals.
4064 * The journal now stores the SELinux context of the logging
4065 client for all entries.
4067 * Add C++ inclusion guards to all public headers
4069 * New output mode "cat" in the journal to print only text
4070 messages, without any meta data like date or time.
4072 * Include tiny X server wrapper as a temporary stop-gap to
4073 teach XOrg udev display enumeration. This is used by display
4074 managers such as gdm, and will go away as soon as XOrg
4075 learned native udev hotplugging for display devices.
4077 * Add new systemd-cat tool for executing arbitrary programs
4078 with STDERR/STDOUT connected to the journal. Can also act as
4079 BSD logger replacement, and does so by default.
4081 * Optionally store all locally generated coredumps in the
4082 journal along with meta data.
4084 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
4085 writing short strings to files (for usage for /sys), and for
4086 creating symlinks, character and block device nodes.
4088 * New unit file option ControlGroupPersistent= to make cgroups
4089 persistent, following the mechanisms outlined in
4090 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
4092 * Support multiple local RTCs in a sane way
4094 * No longer monopolize IO when replaying readahead data on
4095 rotating disks, since we might starve non-file-system IO to
4096 death, since fanotify() will not see accesses done by blkid,
4099 * Do not show kernel threads in systemd-cgls anymore, unless
4100 requested with new -k switch.
4102 Contributions from: Dan Horák, Kay Sievers, Lennart
4103 Poettering, Michal Schmidt
4107 * This is mostly a test release, but incorporates many
4110 * The git repository moved to:
4111 git://anongit.freedesktop.org/systemd/systemd
4112 ssh://git.freedesktop.org/git/systemd/systemd
4114 * First release with the journal
4115 http://0pointer.de/blog/projects/the-journal.html
4117 * The journal replaces both systemd-kmsg-syslogd and
4118 systemd-stdout-bridge.
4120 * New sd_pid_get_unit() API call in libsystemd-logind
4122 * Many systemadm clean-ups
4124 * Introduce remote-fs-pre.target which is ordered before all
4125 remote mounts and may be used to start services before all
4128 * Added Mageia support
4130 * Add bash completion for systemd-loginctl
4132 * Actively monitor PID file creation for daemons which exit in
4133 the parent process before having finished writing the PID
4134 file in the daemon process. Daemons which do this need to be
4135 fixed (i.e. PID file creation must have finished before the
4136 parent exits), but we now react a bit more gracefully to them.
4138 * Add colourful boot output, mimicking the well-known output
4139 of existing distributions.
4141 * New option PassCredentials= for socket units, for
4142 compatibility with a recent kernel ABI breakage.
4144 * /etc/rc.local is now hooked in via a generator binary, and
4145 thus will no longer act as synchronization point during
4148 * systemctl list-unit-files now supports --root=.
4150 * systemd-tmpfiles now understands two new commands: z, Z for
4151 relabelling files according to the SELinux database. This is
4152 useful to apply SELinux labels to specific files in /sys,
4155 * Output of SysV services is now forwarded to both the console
4156 and the journal by default, not only just the console.
4158 * New man pages for all APIs from libsystemd-login.
4160 * The build tree got reorganized and a the build system is a
4161 lot more modular allowing embedded setups to specifically
4162 select the components of systemd they are interested in.
4164 * Support for Linux systems lacking the kernel VT subsystem is
4167 * configure's --with-rootdir= got renamed to
4168 --with-rootprefix= to follow the naming used by udev and
4171 * Unless specified otherwise we will now install to /usr instead
4172 of /usr/local by default.
4174 * Processes with '@' in argv[0][0] are now excluded from the
4175 final shut-down killing spree, following the logic explained
4177 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
4179 * All processes remaining in a service cgroup when we enter
4180 the START or START_PRE states are now killed with
4181 SIGKILL. That means it is no longer possible to spawn
4182 background processes from ExecStart= lines (which was never
4183 supported anyway, and bad style).
4185 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
4186 reloading of units together.
4188 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
4189 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
4190 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
4191 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
4192 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek