]>
Commit | Line | Data |
---|---|---|
d657c51f | 1 | systemd System and Service Manager |
220a21d3 | 2 | |
a2b53448 | 3 | CHANGES WITH 233: |
d08ee7cb | 4 | |
fb7c4eff MG |
5 | * DBus policy files are now installed into /usr rather than /etc. Make |
6 | sure your system has dbus >= 1.9.18 running before upgrading to this | |
7 | version, or override the install path with --with-dbuspolicydir= . | |
8 | ||
5cfc0a84 LP |
9 | * The shell invoked by debug-shell.service now defaults to /bin/sh in |
10 | all cases. If distributions want to use a different shell for this | |
11 | purpose (for example Fedora's /sbin/sushell) they need to specify | |
12 | this explicitly at configure time using --with-debug-shell=. | |
13 | ||
2bcc3309 FB |
14 | * The confirmation spawn prompt has been reworked to offer the |
15 | following choices: | |
16 | ||
b0eb2944 | 17 | (c)ontinue, proceed without asking anymore |
dd6f9ac0 | 18 | (D)ump, show the state of the unit |
2bcc3309 | 19 | (f)ail, don't execute the command and pretend it failed |
d172b175 | 20 | (h)elp |
eedf223a | 21 | (i)nfo, show a short summary of the unit |
56fde33a | 22 | (j)obs, show jobs that are in progress |
2bcc3309 FB |
23 | (s)kip, don't execute the command and pretend it succeeded |
24 | (y)es, execute the command | |
25 | ||
26 | The 'n' choice for the confirmation spawn prompt has been removed, | |
27 | because its meaning was confusing. | |
28 | ||
d08ee7cb LP |
29 | The prompt may now also be redirected to an alternative console by |
30 | specifying the console as parameter to systemd.confirm_spawn=. | |
31 | ||
8e458bfe JW |
32 | * Services of Type=notify require a READY=1 notification to be sent |
33 | during startup. If no such message is sent, the service now fails, | |
34 | even if the main process exited with a successful exit code. | |
35 | ||
85266f9b LP |
36 | * Services that fail to start up correctly now always have their |
37 | ExecStopPost= commands executed. Previously, they'd enter "failed" | |
38 | state directly, without executing these commands. | |
39 | ||
baf32786 MP |
40 | * The option MulticastDNS= of network configuration files has acquired |
41 | an actual implementation. With MulticastDNS=yes a host can resolve | |
d08ee7cb LP |
42 | names of remote hosts and to reply to mDNS's A and AAAA requests from |
43 | the hosts. | |
fa8b4499 | 44 | |
631b676b LP |
45 | * When units are about to be started an additional check is now done to |
46 | ensure that all dependencies of type BindsTo= (when used in | |
47 | combination with After=) have been started. | |
48 | ||
d08ee7cb LP |
49 | * systemd-analyze gained a new verb "syscall-filter" which shows which |
50 | system call groups are defined for the SystemCallFilter= unit file | |
51 | setting, and which system calls they precisely contain. | |
52 | ||
53 | * A new system call filter group "@filesystem" has been added, | |
54 | consisting of various file system related system calls. A group | |
55 | "@reboot" has been added, covering reboot, kexec and shutdown related | |
56 | calls. Finally, a group "@swap" has been added covering swap | |
57 | configuration related calls. | |
58 | ||
59 | * A new unit file option RestrictNamespaces= has been added that may be | |
60 | used to restrict access to the various process namespace types the | |
61 | Linux kernel provides. Specifically, it may be used to take away the | |
62 | right for specific service units to create additional file system, | |
63 | networking, user, and other namespaces. This sandboxing option is | |
64 | particularly relevant due to the high amount of recently discovered | |
65 | namespacing related vulnerabilities in the kernel. | |
66 | ||
67 | * .link gained support for a new AutoNegotiation= setting for | |
68 | configuring Ethernet auto-negotiation. | |
69 | ||
70 | * systemd-networkd's .network files gained support for a new | |
71 | ListenPort= setting in the [DHCP] section to explicitly configure the | |
72 | UDP client port the DHCP client shall listen on. | |
73 | ||
74 | * New systemd-specific mount options are now understood in /etc/fstab: | |
75 | ||
76 | x-systemd.mount-timeout= may be used to configure the maximum | |
77 | permitted runtime of the mount command. | |
78 | ||
79 | x-systemd.device-bound may be set to bind a mount point to its | |
80 | backing device unit, in order to automatically remove a mount point | |
81 | if its backing device is unplugged. This option may also be | |
82 | configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property | |
83 | on the block device, which is now automatically set for all CDROM | |
84 | drives, so that mounted CDs are automatically unmounted when they are | |
85 | removed from the drive. | |
86 | ||
87 | x-systemd.after= and x-systemd.before= may be use to explicitly order | |
88 | a mount after or before another unit or mount point. | |
89 | ||
90 | * Enqueued start jobs for device units are now automatically garbage | |
91 | collected if there are no jobs waiting for them anymore. | |
92 | ||
baf32786 MP |
93 | * systemctl list-jobs gained two new switches: With --after, every |
94 | queued job shows which other queued job is waiting for it; with | |
95 | --before it shows which other jobs every job is waiting for. | |
d08ee7cb LP |
96 | |
97 | * systemd-nspawn gained support for ephemeral boots from disk images | |
98 | (or in other words: --ephemeral and --image= may now be | |
99 | combined). Moreover, ephemeral boots are now supported for normal | |
100 | directories, even if the backing file system is not btrfs. Of course, | |
baf32786 MP |
101 | if the file system does not support file system snapshots or |
102 | reflinks, the initial copy operation will be relatively expensive, but | |
103 | this should still be suitable for many use cases. | |
d08ee7cb LP |
104 | |
105 | * Calendar time specifications in .timer units now support | |
106 | specifications relative to the end of a month by using "~" instead of | |
107 | "-" as separator between month and day. For example, "*-02~03" means | |
108 | "The third last day in February". In addition a new syntax for | |
109 | repeated events has been added using the "/" character. For example, | |
110 | "9..17/2:00" means "every two hours from 9am to 5pm". | |
111 | ||
112 | * systemd-socket-proxyd gained a new parameter --connections-max= for | |
113 | configuring the maximum number of concurrent connections. | |
114 | ||
115 | * All python scripts shipped with systemd (specifically: the various | |
116 | tests written in Python) now require Python 3. | |
117 | ||
baf32786 MP |
118 | * sd-id128 gained a new API for generating unique IDs for the host |
119 | that does not leak the machine ID. Specifically, | |
d08ee7cb | 120 | sd_id128_get_machine_app_specific() derives an ID based on the |
baf32786 | 121 | machine ID a in well-defined, non-reversible, stable way. This is |
d08ee7cb LP |
122 | useful whenever an identifier for the host is needed but where the |
123 | identifier shall not be useful to identify the system beyond the | |
124 | scope of the application itself. (Internally this uses HMAC-SHA256 as | |
125 | keyed hash function using the machine ID as input.) | |
126 | ||
127 | * NotifyAccess= gained a new supported value "exec". When set | |
128 | notifications are accepted from all processes systemd itself invoked, | |
129 | including all control processes. | |
130 | ||
131 | * .nspawn files gained support for defining overlay mounts using the | |
132 | Overlay= and OverlayReadOnly= options. Previously this functionality | |
133 | was only available on the systemd-nspawn command line. | |
134 | ||
135 | * systemd-nspawn's --bind= and --overlay= options gained support for | |
136 | bind/overlay mounts whose source lies within the container tree by | |
137 | prefixing the source path with "+". | |
138 | ||
139 | * systemd-nspawn's --bind= and --overlay= options gained support for | |
140 | automatically allocating a temporary source directory in /var/tmp | |
141 | that is removed when the container dies. Specifically, if the source | |
142 | directory is specified as empty string this mechanism is selected. An | |
143 | example usage is --overlay=+/var::/var, which creates an overlay | |
144 | mount based on the original /var contained in the image, overlayed | |
145 | with a temporary directory in the host's /var/tmp. This way changes | |
146 | to /var are automatically flushed when the container shuts down. | |
147 | ||
148 | * .network files gained a new Unmanaged= boolean setting for explicitly | |
149 | excluding one or more interfaces from management by systemd-networkd. | |
150 | ||
baf32786 MP |
151 | * systemd-nspawn --image= option does now permit raw file system block |
152 | devices (in addition to images containing partition tables, as | |
153 | before). | |
d08ee7cb LP |
154 | |
155 | * The disk image dissection logic in systemd-nspawn gained support for | |
156 | automatically setting up LUKS encrypted as well as Verity protected | |
157 | partitions. When a container is booted from an encrypted image the | |
158 | passphrase is queried at start-up time. When a container with Verity | |
159 | data is started, the root hash is search in a ".roothash" file | |
160 | accompanying the disk image (alternatively, pass the root hash via | |
161 | the new --root-hash= command line option). | |
162 | ||
163 | * A new tool /usr/lib/systemd/systemd-dissect has been added that may | |
164 | be used to dissect disk images the same way as systemd-nspawn does | |
165 | it, following the Bootable Partition Specification. It may even be | |
166 | used to mount disk images with complex partition setups (including | |
167 | LUKS and Verity partitions) to a local host directory, in order to | |
168 | inspect them. This tool is not considered public API (yet), and is | |
169 | thus not installed into /usr/bin. Please do not rely on its | |
3b31c466 | 170 | existence, since it might go away or be changed in later systemd |
d08ee7cb LP |
171 | versions. |
172 | ||
173 | * A new generator "systemd-verity-generator" has been added, similar in | |
baf32786 | 174 | style to "systemd-cryptsetup-generator", permitting automatic setup of |
d08ee7cb LP |
175 | Verity root partitions when systemd boots up. In order to make use of |
176 | this your partition setup should follow the Discoverable Partitions | |
177 | Specification, and the GPT partition ID of the root file system | |
178 | partition should be identical to the upper 128bit of the Verity root | |
179 | hash. The GPT partition ID of the Verity partition protecting it | |
180 | should be the lower 128bit of the Verity root hash. If the partition | |
181 | image follows this model it is sufficient to specify a single | |
182 | "roothash=" kernel command line argument to both configure which root | |
183 | image and verity partition to use as well as the root hash for | |
184 | it. Note that systemd-nspawn's Verity support follows the same | |
185 | semantics, meaning that disk images with proper Verity data in place | |
186 | may be booted in containers with systemd-nspawn as well as on | |
187 | physical systems via the verity generator. Also note that the "mkosi" | |
188 | tool available at https://github.com/systemd/mkosi has been updated | |
189 | to generate Verity protected disk images following this scheme. In | |
190 | fact, it has been updated to generate disk images that optionally | |
191 | implement a complete UEFI SecureBoot trust chain, involving a signed | |
192 | kernel and initrd image that incorporates such a root hash as well as | |
193 | a Verity-enabled root partition. | |
194 | ||
195 | * Support for the %c, %r, %R specifiers in unit files has been | |
196 | removed. Specifiers are not supposed to be dependent on configuration | |
197 | of unit files themselves (so that they resolve to the same regardless | |
198 | where used in the unit files), but these options were due to the | |
199 | existence of the Slice= option. | |
200 | ||
201 | * The various options in the [Match] section of .network files gained | |
202 | support for negative matching. | |
203 | ||
204 | * The hardware database (hwdb) udev supports has been updated to carry | |
205 | accelerometer quirks. | |
206 | ||
207 | * All system services are now run with a fresh kernel keyring set up | |
208 | for them. The invocation ID is stored by default in it, thus | |
209 | providing a safe, non-overridable way to determine the invocation | |
210 | ID of each service. | |
211 | ||
212 | * Service unit files gained new BindPaths= and BindReadOnlyPaths= | |
213 | options for bind mounting arbitrary paths in a service-specific | |
214 | way. When these options are used, arbitrary host or service files and | |
215 | directories may be mounted to arbitrary locations in the service's | |
216 | view. | |
217 | ||
218 | * Documentation has been added that lists all of systemd's low-level | |
219 | environment variables: | |
220 | ||
221 | https://github.com/systemd/systemd/blob/master/ENVIRONMENT.md | |
222 | ||
223 | * sd-daemon gained a new API sd_is_socket_sockaddr() for determining | |
224 | whether a specific socket file descriptor matches a specified socket | |
225 | address. | |
226 | ||
227 | * systemd-firstboot has been updated to check for the | |
228 | systemd.firstboot= kernel command line option. It accepts a boolean | |
229 | and when set to false the first boot questions are skipped. | |
230 | ||
231 | * The systemd-networkd ProxyARP= option has been renamed to | |
232 | IPV4ProxyARP=. Similar, VXLAN-specific option ARPProxy= has been | |
233 | renamed to ReduceARPProxy=. The old names continue to be available | |
234 | for compatibility. | |
235 | ||
236 | * systemd-networkd's bonding device support gained support for two new | |
237 | configuration options ActiveSlave= and PrimarySlave=. | |
238 | ||
239 | * systemd-fstab-generator has been updated to check for the | |
240 | systemd.volatile= kernel command line option, which either takes a | |
241 | boolean parameter or the special value "state". If used the system | |
242 | may be booted in a "volatile" boot mode. Specifically, | |
243 | systemd.volatile=yes is used, the root directory will be mounted as | |
244 | tmpfs, and only /usr is mounted from the actual root file system. If | |
245 | systemd.volatile=state is used, the root directory will be mounted as | |
246 | usual, but /var is mounted as tmpfs. This concept provides similar | |
247 | functionality as systemd-nspawn's --volatile= option, but provides it | |
248 | on physical boots. Use this option for implementing stateless | |
249 | systems, or testing systems with all state and/or configuration reset | |
250 | to the defaults. (Note though that many distributions are not | |
251 | prepared to boot up without a populated /etc or /var, though) | |
252 | ||
253 | * systemd-gpt-auto-generator gained support for LUKS encrypted root | |
254 | partitions. Previously it only supported LUKS encrypted partitions | |
255 | for all other uses, except for the root partition itself. | |
256 | ||
257 | * Socket units gained support for listening on AF_VSOCK sockets for | |
258 | communication in virtualized QEMU environments. | |
259 | ||
260 | * The "configure" script gained a new option --with-fallback-hostname= | |
261 | for specifying the fallback hostname to use if none is configured in | |
262 | /etc/hostname. For example, by specifying | |
263 | --with-fallback-hostname=fedora it is possible to default to a | |
264 | hostname of "fedora" when the user didn't specify anything | |
265 | explicitly. | |
266 | ||
267 | * systemd-cgls gained support for a new --unit= switch for listing only | |
268 | the control groups of a specific unit. Similar --user-unit= has been | |
269 | added for listing only the control groups of a specific user unit. | |
270 | ||
271 | * systemd-mount gained a new --umount switch for unmounting a mount or | |
272 | automount point (and all mount/automount points below it). | |
273 | ||
274 | * systemd will now refuse full configuration reloads (via systemctl | |
275 | daemon-reload and related calls) unless at least 16MiB of free space | |
276 | are available in /run. This is a safety precaution in order to ensure | |
277 | that generators can safely operate after the reload completed. | |
278 | ||
279 | * A new unit file option RootImage= has been added, which has a similar | |
280 | effect as RootDirectory= but mounts the service's root directory from | |
281 | a disk image instead of plain directory. This logic reuses the same | |
282 | image dissection and mount logic that systemd-nspawn already uses, | |
283 | and hence supports any disk images systemd-nspawn supports, including | |
284 | those following the Discoverable Partition Specification, as well as | |
285 | Verity enabled images. This option enables systemd to run system | |
286 | services directly off disk images acting as resource bundles, | |
287 | possibly even including full integrity data. | |
288 | ||
289 | * A new MountAPIVFS= unit file option has been added, taking a boolean | |
baf32786 | 290 | argument. If enabled /proc, /sys and /dev (collectively called the |
d08ee7cb LP |
291 | "API VFS") will be mounted for the service. This is only relevant if |
292 | RootDirectory= or RootImage= is used for the service, as these mounts | |
293 | are of course in place in the host mount namespace anyway. | |
294 | ||
295 | * systemd-nspawn gained support for a new --pivot-root= switch. If | |
296 | specified the root directory within the container image is pivoted to | |
297 | the specified mount point, while the original root disk is moved to a | |
298 | different place. This option enables booting of ostree images | |
299 | directly with systemd-nspawn. | |
300 | ||
301 | * systemd-networkd gained support for configuring IPv6 Proxy NDP | |
302 | addresses via the new IPv6ProxyNDPAddress= .network file setting. | |
303 | ||
304 | * The systemd build scripts will no longer complain if the NTP server | |
305 | addresses are not changed from the defaults. Google is now supporting | |
306 | these NTP servers officially. We still recommend downstreams to | |
307 | properly register an NTP pool with the NTP pool project though. | |
308 | ||
309 | * coredumpctl gained new new "--reverse" option for printing the list | |
310 | of coredumps in reverse order. | |
311 | ||
312 | * The systemd-coredump logic has been improved so that it may be reused | |
313 | for collecting backtraces in non-compiled languages, for example in | |
314 | scripting languages such as Python. | |
315 | ||
316 | * machinectl will now show the UID shift of local containers, if user | |
317 | namespacing is enabled for them. | |
318 | ||
baf32786 | 319 | * systemd will now optionally run "environment generator" binaries at |
d08ee7cb LP |
320 | configuration load time. They may be used to add environment |
321 | variables to the environment block passed to services invoked. One | |
baf32786 | 322 | user environment generator is shipped by default that sets up |
d08ee7cb LP |
323 | environment variables based on files dropped into |
324 | ~/.config/environment.d/. | |
325 | ||
a2b53448 LP |
326 | * systemd-resolved now includes the new, recently published 2017 DNSSEC |
327 | root key (KSK). | |
328 | ||
329 | * coredumpctl will now show additional information about truncated and | |
330 | inaccessible coredumps, as well as coredumps that are still being | |
331 | processed. It also gained a new --quiet switch for suppressing | |
332 | additional informational message in its output. | |
333 | ||
334 | * coredumpctl gained support for only showing coredumps newer and/or | |
335 | older than specific timestamps, using the new --since= and --until= | |
336 | options, reminiscent of journalctl's options by the same name. | |
337 | ||
338 | * hostnamed has been updated to report a new chassis type of | |
339 | "convertible" to cover "foldable" laptops that can both act as a | |
340 | tablet and as a laptop, such as various Lenovo Yoga devices. | |
341 | ||
342 | * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH, | |
343 | CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the | |
344 | kernel. | |
345 | ||
346 | * The "hybrid" control group mode has been modified to improve | |
347 | compatibility with "legacy" cgroupsv1 setups. Specifically, the | |
348 | "hybrid" setup of /sys/fs/cgroup is now pretty much identical to | |
349 | "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named | |
350 | cgroupsv1 hierarchy), the only externally visible change being that | |
351 | the cgroupsv2 hierarchy is also mounted, to | |
352 | /sys/fs/cgroup/unified. This should provide a large degree of | |
353 | compatibility with "legacy" cgroupsv1, while taking benefit of the | |
354 | better management capabilities of cgroupsv2. | |
355 | ||
356 | * The default control group setup mode may be selected both a boot-time | |
357 | via a set of kernel command line parameters (specifically: | |
358 | systemd.unified_cgroup_hierarchy= and | |
359 | systemd.legacy_systemd_cgroup_controller=), as well as a compile-time | |
360 | default selected on the configure command line | |
361 | (--with-default-hierarchy=). The upstream default is "hybrid" | |
362 | (i.e. the cgroupsv1 + cgroupsv2 mixture discussed above) now, but | |
363 | this will change in a future systemd version to be "unified" (pure | |
364 | cgroupsv2 mode). The third option for the compile time option is | |
365 | "legacy", to enter pure cgroupsv1 mode. We recommend downstream | |
366 | distributions to default to "hybrid" mode for release distributions, | |
367 | starting with v233. We recommend "unified" for development | |
368 | distributions (specifically: distributions such as Fedora's rawhide) | |
369 | as that's where things are headed in the long run. Use "legacy" for | |
370 | greatest stability and compatibility only. | |
371 | ||
372 | * Note one current limitation of "unified" and "hybrid" control group | |
373 | setup modes: the kernel currently does not permit the systemd --user | |
374 | instance (i.e. unprivileged code) to migrate processes between two | |
375 | disconnected cgroup subtrees, even if both are managed and owned by | |
376 | the user. This effectively means "systemd-run --user --scope" doesn't | |
377 | work when invoked from outside of any "systemd --user" service or | |
378 | scope. Specifically, it is not supported from session scopes. We are | |
379 | working on fixing this in a future systemd version. (See #3388 for | |
380 | further details about this.) | |
381 | ||
d08ee7cb LP |
382 | Contributions from: Adrián López, Alexander Galanin, Alexander |
383 | Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch | |
384 | Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric | |
385 | Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri, | |
386 | Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner, | |
387 | David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry | |
388 | Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly, | |
389 | Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn | |
390 | Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter, | |
a2b53448 LP |
391 | Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede, |
392 | Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan | |
393 | Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen, | |
394 | Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart | |
395 | Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de | |
396 | Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry, | |
397 | Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de | |
398 | Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal | |
399 | Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak, | |
400 | Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip | |
401 | Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin | |
402 | Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx, | |
403 | Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi, | |
404 | Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève, | |
405 | Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor | |
406 | Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar | |
407 | Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb, | |
408 | Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun, | |
409 | YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр | |
410 | Тихонов | |
411 | ||
412 | — Berlin, 2017-03-01 | |
d08ee7cb | 413 | |
54b24597 | 414 | CHANGES WITH 232: |
76153ad4 | 415 | |
05f426d2 LP |
416 | * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and |
417 | RestrictAddressFamilies= enabled. These sandboxing options should | |
418 | generally be compatible with the various external udev call-out | |
419 | binaries we are aware of, however there may be exceptions, in | |
420 | particular when exotic languages for these call-outs are used. In | |
421 | this case, consider turning off these settings locally. | |
422 | ||
4ffe2479 ZJS |
423 | * The new RemoveIPC= option can be used to remove IPC objects owned by |
424 | the user or group of a service when that service exits. | |
425 | ||
6fa44114 | 426 | * The new ProtectKernelModules= option can be used to disable explicit |
4c37970d LP |
427 | load and unload operations of kernel modules by a service. In |
428 | addition access to /usr/lib/modules is removed if this option is set. | |
6fa44114 | 429 | |
4a77c53d ZJS |
430 | * ProtectSystem= option gained a new value "strict", which causes the |
431 | whole file system tree with the exception of /dev, /proc, and /sys, | |
432 | to be remounted read-only for a service. | |
433 | ||
e49e2c25 | 434 | * The new ProtectKernelTunables= option can be used to disable |
4a77c53d ZJS |
435 | modification of configuration files in /sys and /proc by a service. |
436 | Various directories and files are remounted read-only, so access is | |
437 | restricted even if the file permissions would allow it. | |
438 | ||
6fa44114 | 439 | * The new ProtectControlGroups= option can be used to disable write |
4a77c53d ZJS |
440 | access by a service to /sys/fs/cgroup. |
441 | ||
442 | * Various systemd services have been hardened with | |
443 | ProtectKernelTunables=yes, ProtectControlGroups=yes, | |
444 | RestrictAddressFamilies=. | |
445 | ||
4ffe2479 ZJS |
446 | * Support for dynamically creating users for the lifetime of a service |
447 | has been added. If DynamicUser=yes is specified, user and group IDs | |
448 | will be allocated from the range 61184..65519 for the lifetime of the | |
449 | service. They can be resolved using the new nss-systemd.so NSS | |
450 | module. The module must be enabled in /etc/nsswitch.conf. Services | |
451 | started in this way have PrivateTmp= and RemoveIPC= enabled, so that | |
452 | any resources allocated by the service will be cleaned up when the | |
4a77c53d ZJS |
453 | service exits. They also have ProtectHome=read-only and |
454 | ProtectSystem=strict enabled, so they are not able to make any | |
455 | permanent modifications to the system. | |
4ffe2479 | 456 | |
171ae2cd | 457 | * The nss-systemd module also always resolves root and nobody, making |
4ffe2479 | 458 | it possible to have no /etc/passwd or /etc/group files in minimal |
171ae2cd | 459 | container or chroot environments. |
4ffe2479 ZJS |
460 | |
461 | * Services may be started with their own user namespace using the new | |
171ae2cd LP |
462 | boolean PrivateUsers= option. Only root, nobody, and the uid/gid |
463 | under which the service is running are mapped. All other users are | |
464 | mapped to nobody. | |
4ffe2479 ZJS |
465 | |
466 | * Support for the cgroup namespace has been added to systemd-nspawn. If | |
467 | supported by kernel, the container system started by systemd-nspawn | |
468 | will have its own view of the cgroup hierarchy. This new behaviour | |
469 | can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable. | |
470 | ||
471 | * The new MemorySwapMax= option can be used to limit the maximum swap | |
472 | usage under the unified cgroup hierarchy. | |
473 | ||
474 | * Support for the CPU controller in the unified cgroup hierarchy has | |
475 | been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting= | |
476 | options. This controller requires out-of-tree patches for the kernel | |
477 | and the support is provisional. | |
478 | ||
171ae2cd LP |
479 | * Mount and automount units may now be created transiently |
480 | (i.e. dynamically at runtime via the bus API, instead of requiring | |
481 | unit files in the file system). | |
482 | ||
483 | * systemd-mount is a new tool which may mount file systems – much like | |
484 | mount(8), optionally pulling in additional dependencies through | |
485 | transient .mount and .automount units. For example, this tool | |
486 | automatically runs fsck on a backing block device before mounting, | |
487 | and allows the automount logic to be used dynamically from the | |
488 | command line for establishing mount points. This tool is particularly | |
489 | useful when dealing with removable media, as it will ensure fsck is | |
490 | run – if necessary – before the first access and that the file system | |
491 | is quickly unmounted after each access by utilizing the automount | |
492 | logic. This maximizes the chance that the file system on the | |
493 | removable media stays in a clean state, and if it isn't in a clean | |
494 | state is fixed automatically. | |
4ffe2479 ZJS |
495 | |
496 | * LazyUnmount=yes option for mount units has been added to expose the | |
497 | umount --lazy option. Similarly, ForceUnmount=yes exposes the --force | |
498 | option. | |
499 | ||
500 | * /efi will be used as the mount point of the EFI boot partition, if | |
501 | the directory is present, and the mount point was not configured | |
502 | through other means (e.g. fstab). If /efi directory does not exist, | |
503 | /boot will be used as before. This makes it easier to automatically | |
504 | mount the EFI partition on systems where /boot is used for something | |
505 | else. | |
506 | ||
171ae2cd LP |
507 | * When operating on GPT disk images for containers, systemd-nspawn will |
508 | now mount the ESP to /boot or /efi according to the same rules as PID | |
509 | 1 running on a host. This allows tools like "bootctl" to operate | |
510 | correctly within such containers, in order to make container images | |
511 | bootable on physical systems. | |
512 | ||
4a77c53d | 513 | * disk/by-id and disk/by-path symlinks are now created for NVMe drives. |
4ffe2479 ZJS |
514 | |
515 | * Two new user session targets have been added to support running | |
516 | graphical sessions under the systemd --user instance: | |
517 | graphical-session.target and graphical-session-pre.target. See | |
518 | systemd.special(7) for a description of how those targets should be | |
519 | used. | |
520 | ||
521 | * The vconsole initialization code has been significantly reworked to | |
d4c08299 | 522 | use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better |
4ffe2479 ZJS |
523 | support unicode keymaps. Font and keymap configuration will now be |
524 | copied to all allocated virtual consoles. | |
525 | ||
05ecf467 | 526 | * FreeBSD's bhyve virtualization is now detected. |
4ffe2479 | 527 | |
d4c08299 | 528 | * Information recorded in the journal for core dumps now includes the |
4ffe2479 ZJS |
529 | contents of /proc/mountinfo and the command line of the process at |
530 | the top of the process hierarchy (which is usually the init process | |
531 | of the container). | |
532 | ||
171ae2cd | 533 | * systemd-journal-gatewayd learned the --directory= option to serve |
4ffe2479 ZJS |
534 | files from the specified location. |
535 | ||
536 | * journalctl --root=… can be used to peruse the journal in the | |
537 | /var/log/ directories inside of a container tree. This is similar to | |
538 | the existing --machine= option, but does not require the container to | |
539 | be active. | |
540 | ||
541 | * The hardware database has been extended to support | |
542 | ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify | |
543 | trackball devices. | |
544 | ||
545 | MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to | |
546 | specify the click rate for mice which include a horizontal wheel with | |
547 | a click rate that is different than the one for the vertical wheel. | |
548 | ||
549 | * systemd-run gained a new --wait option that makes service execution | |
171ae2cd LP |
550 | synchronous. (Specifically, the command will not return until the |
551 | specified service binary exited.) | |
4ffe2479 | 552 | |
171ae2cd | 553 | * systemctl gained a new --wait option that causes the start command to |
4a77c53d ZJS |
554 | wait until the units being started have terminated again. |
555 | ||
171ae2cd | 556 | * A new journal output mode "short-full" has been added which displays |
4ffe2479 | 557 | timestamps with abbreviated English day names and adds a timezone |
171ae2cd LP |
558 | suffix. Those timestamps include more information than the default |
559 | "short" output mode, and can be passed directly to journalctl's | |
560 | --since= and --until= options. | |
4ffe2479 ZJS |
561 | |
562 | * /etc/resolv.conf will be bind-mounted into containers started by | |
563 | systemd-nspawn, if possible, so any changes to resolv.conf contents | |
564 | are automatically propagated to the container. | |
565 | ||
566 | * The number of instances for socket-activated services originating | |
171ae2cd LP |
567 | from a single IP address can be limited with |
568 | MaxConnectionsPerSource=, extending the existing setting of | |
569 | MaxConnections=. | |
4ffe2479 | 570 | |
4a77c53d ZJS |
571 | * systemd-networkd gained support for vcan ("Virtual CAN") interface |
572 | configuration. | |
573 | ||
574 | * .netdev and .network configuration can now be extended through | |
575 | drop-ins. | |
576 | ||
4ffe2479 ZJS |
577 | * UDP Segmentation Offload, TCP Segmentation Offload, Generic |
578 | Segmentation Offload, Generic Receive Offload, Large Receive Offload | |
579 | can be enabled and disabled using the new UDPSegmentationOffload=, | |
580 | TCPSegmentationOffload=, GenericSegmentationOffload=, | |
581 | GenericReceiveOffload=, LargeReceiveOffload= options in the | |
582 | [Link] section of .link files. | |
583 | ||
171ae2cd LP |
584 | * The Spanning Tree Protocol, Priority, Aging Time, and the Default |
585 | Port VLAN ID can be configured for bridge devices using the new STP=, | |
586 | Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge] | |
587 | section of .netdev files. | |
4ffe2479 | 588 | |
171ae2cd | 589 | * The route table to which routes received over DHCP or RA should be |
4a77c53d ZJS |
590 | added can be configured with the new RouteTable= option in the [DHCP] |
591 | and [IPv6AcceptRA] sections of .network files. | |
592 | ||
171ae2cd | 593 | * The Address Resolution Protocol can be disabled on links managed by |
4ffe2479 ZJS |
594 | systemd-networkd using the ARP=no setting in the [Link] section of |
595 | .network files. | |
596 | ||
171ae2cd LP |
597 | * New environment variables $SERVICE_RESULT, $EXIT_CODE and |
598 | $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and | |
599 | encode information about the result and exit codes of the current | |
600 | service runtime cycle. | |
4ffe2479 | 601 | |
4a77c53d | 602 | * systemd-sysctl will now configure kernel parameters in the order |
1f4f4cf7 | 603 | they occur in the configuration files. This matches what sysctl |
4a77c53d ZJS |
604 | has been traditionally doing. |
605 | ||
606 | * kernel-install "plugins" that are executed to perform various | |
607 | tasks after a new kernel is added and before an old one is removed | |
608 | can now return a special value to terminate the procedure and | |
609 | prevent any later plugins from running. | |
610 | ||
76153ad4 | 611 | * Journald's SplitMode=login setting has been deprecated. It has been |
d4c08299 | 612 | removed from documentation, and its use is discouraged. In a future |
76153ad4 ZJS |
613 | release it will be completely removed, and made equivalent to current |
614 | default of SplitMode=uid. | |
615 | ||
4a77c53d ZJS |
616 | * Storage=both option setting in /etc/systemd/coredump.conf has been |
617 | removed. With fast LZ4 compression storing the core dump twice is not | |
618 | useful. | |
619 | ||
4ffe2479 ZJS |
620 | * The --share-system systemd-nspawn option has been replaced with an |
621 | (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of | |
622 | this functionality is discouraged. In addition the variables | |
623 | $SYSTEMD_NSPAWN_SHARE_NS_IPC, $SYSTEMD_NSPAWN_SHARE_NS_PID, | |
624 | $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of | |
625 | individual namespaces. | |
626 | ||
171ae2cd LP |
627 | * "machinectl list" now shows the IP address of running containers in |
628 | the output, as well as OS release information. | |
629 | ||
630 | * "loginctl list" now shows the TTY of each session in the output. | |
631 | ||
632 | * sd-bus gained new API calls sd_bus_track_set_recursive(), | |
633 | sd_bus_track_get_recursive(), sd_bus_track_count_name(), | |
634 | sd_bus_track_count_sender(). They permit usage of sd_bus_track peer | |
635 | tracking objects in a "recursive" mode, where a single client can be | |
636 | counted multiple times, if it takes multiple references. | |
637 | ||
638 | * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and | |
639 | sd_bus_get_exit_on_disconnect(). They may be used to to make a | |
640 | process using sd-bus automatically exit if the bus connection is | |
641 | severed. | |
642 | ||
643 | * Bus clients of the service manager may now "pin" loaded units into | |
644 | memory, by taking an explicit reference on them. This is useful to | |
645 | ensure the client can retrieve runtime data about the service even | |
646 | after the service completed execution. Taking such a reference is | |
647 | available only for privileged clients and should be helpful to watch | |
648 | running services in a race-free manner, and in particular collect | |
649 | information about exit statuses and results. | |
650 | ||
4c37970d LP |
651 | * The nss-resolve module has been changed to strictly return UNAVAIL |
652 | when communication via D-Bus with resolved failed, and NOTFOUND when | |
653 | a lookup completed but was negative. This means it is now possible to | |
654 | neatly configure fallbacks using nsswitch.conf result checking | |
655 | expressions. Taking benefit of this, the new recommended | |
656 | configuration line for the "hosts" entry in /etc/nsswitch.conf is: | |
657 | ||
658 | hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname | |
659 | ||
660 | * A new setting CtrlAltDelBurstAction= has been added to | |
661 | /etc/systemd/system.conf which may be used to configure the precise | |
662 | behaviour if the user on the console presses Ctrl-Alt-Del more often | |
663 | than 7 times in 2s. Previously this would unconditionally result in | |
664 | an expedited, immediate reboot. With this new setting the precise | |
665 | operation may be configured in more detail, and also turned off | |
666 | entirely. | |
667 | ||
668 | * In .netdev files two new settings RemoteChecksumTx= and | |
669 | RemoteChecksumRx= are now understood that permit configuring the | |
670 | remote checksumming logic for VXLAN networks. | |
671 | ||
672 | * The service manager learnt a new "invocation ID" concept for invoked | |
673 | services. Each runtime cycle of a service will get a new invocation | |
674 | ID (a 128bit random UUID) assigned that identifies the current | |
675 | run of the service uniquely and globally. A new invocation ID | |
676 | is generated each time a service starts up. The journal will store | |
677 | the invocation ID of a service along with any logged messages, thus | |
678 | making the invocation ID useful for matching the online runtime of a | |
679 | service with the offline log data it generated in a safe way without | |
680 | relying on synchronized timestamps. In many ways this new service | |
681 | invocation ID concept is similar to the kernel's boot ID concept that | |
682 | uniquely and globally identifies the runtime of each boot. The | |
683 | invocation ID of a service is passed to the service itself via an | |
684 | environment variable ($INVOCATION_ID). A new bus call | |
685 | GetUnitByInvocationID() has been added that is similar to GetUnit() | |
686 | but instead of retrieving the bus path for a unit by its name | |
687 | retrieves it by its invocation ID. The returned path is valid only as | |
688 | long as the passed invocation ID is current. | |
689 | ||
690 | * systemd-resolved gained a new "DNSStubListener" setting in | |
691 | resolved.conf. It either takes a boolean value or the special values | |
692 | "udp" and "tcp", and configures whether to enable the stub DNS | |
693 | listener on 127.0.0.53:53. | |
694 | ||
695 | * IP addresses configured via networkd may now carry additional | |
696 | configuration settings supported by the kernel. New options include: | |
697 | HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=, | |
698 | PrefixRoute=, AutoJoin=. | |
699 | ||
700 | * The PAM configuration fragment file for "user@.service" shipped with | |
701 | systemd (i.e. the --user instance of systemd) has been stripped to | |
702 | the minimum necessary to make the system boot. Previously, it | |
703 | contained Fedora-specific stanzas that did not apply to other | |
704 | distributions. It is expected that downstream distributions add | |
705 | additional configuration lines, matching their needs to this file, | |
706 | using it only as rough template of what systemd itself needs. Note | |
707 | that this reduced fragment does not even include an invocation of | |
708 | pam_limits which most distributions probably want to add, even though | |
709 | systemd itself does not need it. (There's also the new build time | |
710 | option --with-pamconfdir=no to disable installation of the PAM | |
711 | fragment entirely.) | |
712 | ||
713 | * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO | |
714 | capability is now also dropped from its set (in addition to | |
715 | CAP_SYS_MKNOD as before). | |
716 | ||
717 | * In service unit files it is now possible to connect a specific named | |
718 | file descriptor with stdin/stdout/stdout of an executed service. The | |
719 | name may be specified in matching .socket units using the | |
720 | FileDescriptorName= setting. | |
721 | ||
722 | * A number of journal settings may now be configured on the kernel | |
723 | command line. Specifically, the following options are now understood: | |
724 | systemd.journald.max_level_console=, | |
725 | systemd.journald.max_level_store=, | |
726 | systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=, | |
727 | systemd.journald.max_level_wall=. | |
728 | ||
729 | * "systemctl is-enabled --full" will now show by which symlinks a unit | |
730 | file is enabled in the unit dependency tree. | |
731 | ||
b4eed568 LP |
732 | * Support for VeraCrypt encrypted partitions has been added to the |
733 | "cryptsetup" logic and /etc/crypttab. | |
734 | ||
735 | * systemd-detect-virt gained support for a new --private-users switch | |
736 | that checks whether the invoking processes are running inside a user | |
737 | namespace. Similar, a new special value "private-users" for the | |
738 | existing ConditionVirtualization= setting has been added, permitting | |
739 | skipping of specific units in user namespace environments. | |
740 | ||
07393b6e LP |
741 | Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John, |
742 | Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin | |
743 | Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner, | |
744 | Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez, | |
745 | Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick, | |
746 | Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg, | |
747 | Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric | |
748 | Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang, | |
749 | Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke, | |
750 | Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan | |
751 | Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker, | |
752 | Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski, | |
753 | Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering, | |
754 | Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš | |
755 | Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou, | |
756 | Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej | |
757 | Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy, | |
758 | Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike | |
759 | Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer, | |
760 | Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny | |
761 | Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant | |
762 | Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit, | |
763 | Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut | |
764 | Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann | |
765 | E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew | |
766 | Jędrzejewski-Szmek, Zeal Jagannatha | |
767 | ||
54b24597 | 768 | — Santa Fe, 2016-11-03 |
07393b6e | 769 | |
5cd118ba MP |
770 | CHANGES WITH 231: |
771 | ||
fcd30826 LP |
772 | * In service units the various ExecXYZ= settings have been extended |
773 | with an additional special character as first argument of the | |
43eb109a | 774 | assigned value: if the character '+' is used the specified command |
fcd30826 LP |
775 | line it will be run with full privileges, regardless of User=, |
776 | Group=, CapabilityBoundingSet= and similar options. The effect is | |
777 | similar to the existing PermissionsStartOnly= option, but allows | |
778 | configuration of this concept for each executed command line | |
779 | independently. | |
780 | ||
781 | * Services may now alter the service watchdog timeout at runtime by | |
782 | sending a WATCHDOG_USEC= message via sd_notify(). | |
783 | ||
784 | * MemoryLimit= and related unit settings now optionally take percentage | |
785 | specifications. The percentage is taken relative to the amount of | |
786 | physical memory in the system (or in case of containers, the assigned | |
787 | amount of memory). This allows scaling service resources neatly with | |
771de3f5 | 788 | the amount of RAM available on the system. Similarly, systemd-logind's |
fcd30826 LP |
789 | RuntimeDirectorySize= option now also optionally takes percentage |
790 | values. | |
791 | ||
792 | * In similar fashion TasksMax= takes percentage values now, too. The | |
793 | value is taken relative to the configured maximum number of processes | |
794 | on the system. The per-service task maximum has been changed to 15% | |
795 | using this functionality. (Effectively this is an increase of 512 → | |
796 | 4915 for service units, given the kernel's default pid_max setting.) | |
797 | ||
798 | * Calendar time specifications in .timer units now understand a ".." | |
799 | syntax for time ranges. Example: "4..7:10" may now be used for | |
800 | defining a timer that is triggered at 4:10am, 5:10am, 6:10am and | |
801 | 7:10am every day. | |
802 | ||
803 | * The InaccessableDirectories=, ReadOnlyDirectories= and | |
804 | ReadWriteDirectories= unit file settings have been renamed to | |
805 | InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be | |
806 | applied to all kinds of file nodes, and not just directories, with | |
807 | the exception of symlinks. Specifically these settings may now be | |
808 | used on block and character device nodes, UNIX sockets and FIFOS as | |
809 | well as regular files. The old names of these settings remain | |
810 | available for compatibility. | |
811 | ||
812 | * systemd will now log about all service processes it kills forcibly | |
813 | (using SIGKILL) because they remained after the clean shutdown phase | |
814 | of the service completed. This should help identifying services that | |
815 | shut down uncleanly. Moreover if KillUserProcesses= is enabled in | |
816 | systemd-logind's configuration a similar log message is generated for | |
817 | processes killed at the end of each session due to this setting. | |
818 | ||
819 | * systemd will now set the $JOURNAL_STREAM environment variable for all | |
820 | services whose stdout/stderr are connected to the Journal (which | |
821 | effectively means by default: all services). The variable contains | |
822 | the device and inode number of the file descriptor used for | |
823 | stdout/stderr. This may be used by invoked programs to detect whether | |
824 | their stdout/stderr is connected to the Journal, in which case they | |
825 | can switch over to direct Journal communication, thus being able to | |
826 | pass extended, structured metadata along with their log messages. As | |
827 | one example, this is now used by glib's logging primitives. | |
828 | ||
829 | * When using systemd's default tmp.mount unit for /tmp, the mount point | |
830 | will now be established with the "nosuid" and "nodev" options. This | |
831 | avoids privilege escalation attacks that put traps and exploits into | |
832 | /tmp. However, this might cause problems if you e. g. put container | |
5cd118ba MP |
833 | images or overlays into /tmp; if you need this, override tmp.mount's |
834 | "Options=" with a drop-in, or mount /tmp from /etc/fstab with your | |
835 | desired options. | |
836 | ||
fcd30826 LP |
837 | * systemd now supports the "memory" cgroup controller also on |
838 | cgroupsv2. | |
839 | ||
840 | * The systemd-cgtop tool now optionally takes a control group path as | |
841 | command line argument. If specified, the control group list shown is | |
842 | limited to subgroups of that group. | |
843 | ||
844 | * The SystemCallFilter= unit file setting gained support for | |
845 | pre-defined, named system call filter sets. For example | |
846 | SystemCallFilter=@clock is now an effective way to make all clock | |
771de3f5 | 847 | changing-related system calls unavailable to a service. A number of |
fcd30826 LP |
848 | similar pre-defined groups are defined. Writing system call filters |
849 | for system services is simplified substantially with this new | |
850 | concept. Accordingly, all of systemd's own, long-running services now | |
851 | enable system call filtering based on this, by default. | |
852 | ||
853 | * A new service setting MemoryDenyWriteExecute= has been added, taking | |
854 | a boolean value. If turned on, a service may no longer create memory | |
855 | mappings that are writable and executable at the same time. This | |
856 | enhances security for services where this is enabled as it becomes | |
857 | harder to dynamically write and then execute memory in exploited | |
858 | service processes. This option has been enabled for all of systemd's | |
859 | own long-running services. | |
860 | ||
861 | * A new RestrictRealtime= service setting has been added, taking a | |
862 | boolean argument. If set the service's processes may no longer | |
863 | acquire realtime scheduling. This improves security as realtime | |
864 | scheduling may otherwise be used to easily freeze the system. | |
865 | ||
866 | * systemd-nspawn gained a new switch --notify-ready= taking a boolean | |
867 | value. This may be used for requesting that the system manager inside | |
868 | of the container reports start-up completion to nspawn which then | |
869 | propagates this notification further to the service manager | |
870 | supervising nspawn itself. A related option NotifyReady= in .nspawn | |
871 | files has been added too. This functionality allows ordering of the | |
872 | start-up of multiple containers using the usual systemd ordering | |
873 | primitives. | |
874 | ||
875 | * machinectl gained a new command "stop" that is an alias for | |
876 | "terminate". | |
877 | ||
878 | * systemd-resolved gained support for contacting DNS servers on | |
879 | link-local IPv6 addresses. | |
880 | ||
881 | * If systemd-resolved receives the SIGUSR2 signal it will now flush all | |
882 | its caches. A method call for requesting the same operation has been | |
883 | added to the bus API too, and is made available via "systemd-resolve | |
884 | --flush-caches". | |
885 | ||
771de3f5 | 886 | * systemd-resolve gained a new --status switch. If passed a brief |
fcd30826 LP |
887 | summary of the used DNS configuration with per-interface information |
888 | is shown. | |
889 | ||
890 | * resolved.conf gained a new Cache= boolean option, defaulting to | |
891 | on. If turned off local DNS caching is disabled. This comes with a | |
892 | performance penalty in particular when DNSSEC is enabled. Note that | |
771de3f5 | 893 | resolved disables its internal caching implicitly anyway, when the |
fcd30826 LP |
894 | configured DNS server is on a host-local IP address such as ::1 or |
895 | 127.0.0.1, thus automatically avoiding double local caching. | |
896 | ||
897 | * systemd-resolved now listens on the local IP address 127.0.0.53:53 | |
898 | for DNS requests. This improves compatibility with local programs | |
899 | that do not use the libc NSS or systemd-resolved's bus APIs for name | |
900 | resolution. This minimal DNS service is only available to local | |
901 | programs and does not implement the full DNS protocol, but enough to | |
902 | cover local DNS clients. A new, static resolv.conf file, listing just | |
903 | this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is | |
904 | now recommended to make /etc/resolv.conf a symlink to this file in | |
905 | order to route all DNS lookups to systemd-resolved, regardless if | |
906 | done via NSS, the bus API or raw DNS packets. Note that this local | |
907 | DNS service is not as fully featured as the libc NSS or | |
908 | systemd-resolved's bus APIs. For example, as unicast DNS cannot be | |
909 | used to deliver link-local address information (as this implies | |
910 | sending a local interface index along), LLMNR/mDNS support via this | |
911 | interface is severely restricted. It is thus strongly recommended for | |
912 | all applications to use the libc NSS API or native systemd-resolved | |
913 | bus API instead. | |
914 | ||
915 | * systemd-networkd's bridge support learned a new setting | |
916 | VLANFiltering= for controlling VLAN filtering. Moreover a new section | |
917 | in .network files has been added for configuring VLAN bridging in | |
918 | more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN]. | |
919 | ||
920 | * systemd-networkd's IPv6 Router Advertisement code now makes use of | |
921 | the DNSSL and RDNSS options. This means IPv6 DNS configuration may | |
922 | now be acquired without relying on DHCPv6. Two new options | |
923 | UseDomains= and UseDNS= have been added to configure this behaviour. | |
924 | ||
925 | * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been | |
926 | renamed IPv6AcceptRA=, without altering its behaviour. The old | |
927 | setting name remains available for compatibility reasons. | |
928 | ||
929 | * The systemd-networkd VTI/VTI6 tunneling support gained new options | |
930 | Key=, InputKey= and OutputKey=. | |
931 | ||
932 | * systemd-networkd gained support for VRF ("Virtual Routing Function") | |
933 | interface configuration. | |
934 | ||
935 | * "systemctl edit" may now be used to create new unit files by | |
936 | specifying the --force switch. | |
937 | ||
938 | * sd-event gained a new function sd_event_get_iteration() for | |
939 | requesting the current iteration counter of the event loop. It starts | |
940 | at zero and is increased by one with each event loop iteration. | |
941 | ||
43a569a1 ZJS |
942 | * A new rpm macro %systemd_ordering is provided by the macros.systemd |
943 | file. It can be used in lieu of %systemd_requires in packages which | |
944 | don't use any systemd functionality and are intended to be installed | |
945 | in minimal containers without systemd present. This macro provides | |
ce830873 | 946 | ordering dependencies to ensure that if the package is installed in |
43a569a1 ZJS |
947 | the same rpm transaction as systemd, systemd will be installed before |
948 | the scriptlets for the package are executed, allowing unit presets | |
949 | to be handled. | |
950 | ||
951 | New macros %_systemdgeneratordir and %_systemdusergeneratordir have | |
952 | been added to simplify packaging of generators. | |
953 | ||
954 | * The os-release file gained VERSION_CODENAME field for the | |
955 | distribution nickname (e.g. VERSION_CODENAME=woody). | |
956 | ||
957 | * New udev property UDEV_DISABLE_PERSISTENT_STORAGE_RULES_FLAG=1 | |
958 | can be set to disable parsing of metadata and the creation | |
959 | of persistent symlinks for that device. | |
960 | ||
0f1da52b LP |
961 | * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess" |
962 | to make them available to logged-in users has been reverted. | |
963 | ||
964 | * Much of the common code of the various systemd components is now | |
965 | built into an internal shared library libsystemd-shared-231.so | |
966 | (incorporating the systemd version number in the name, to be updated | |
967 | with future releases) that the components link to. This should | |
968 | decrease systemd footprint both in memory during runtime and on | |
969 | disk. Note that the shared library is not for public use, and is | |
970 | neither API not ABI stable, but is likely to change with every new | |
1ecbf32f ZJS |
971 | released update. Packagers need to make sure that binaries |
972 | linking to libsystemd-shared.so are updated in step with the | |
973 | library. | |
43a569a1 | 974 | |
fcd30826 LP |
975 | * Configuration for "mkosi" is now part of the systemd |
976 | repository. mkosi is a tool to easily build legacy-free OS images, | |
977 | and is available on github: https://github.com/systemd/mkosi. If | |
978 | "mkosi" is invoked in the build tree a new raw OS image is generated | |
979 | incorporating the systemd sources currently being worked on and a | |
980 | clean, fresh distribution installation. The generated OS image may be | |
ce830873 | 981 | booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical |
fcd30826 LP |
982 | UEFI PC. This functionality is particularly useful to easily test |
983 | local changes made to systemd in a pristine, defined environment. See | |
984 | HACKING for details. | |
ceeddf79 | 985 | |
4ffe2479 ZJS |
986 | * configure learned the --with-support-url= option to specify the |
987 | distribution's bugtracker. | |
988 | ||
38b383d9 LP |
989 | Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor |
990 | Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika | |
991 | Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar | |
992 | Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse, | |
993 | Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David | |
994 | Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias | |
995 | Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler, | |
996 | Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan | |
997 | Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke | |
998 | Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart | |
999 | Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel | |
1000 | Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov, | |
1001 | Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz, | |
1002 | Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran, | |
1003 | Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier, | |
1004 | Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas | |
771de3f5 ZJS |
1005 | Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan, |
1006 | Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič, | |
38b383d9 | 1007 | WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek |
5cd118ba | 1008 | |
38b383d9 | 1009 | — Berlin, 2016-07-25 |
5cd118ba | 1010 | |
46e40fab | 1011 | CHANGES WITH 230: |
7f6e8043 | 1012 | |
61ecb465 LP |
1013 | * DNSSEC is now turned on by default in systemd-resolved (in |
1014 | "allow-downgrade" mode), but may be turned off during compile time by | |
1015 | passing "--with-default-dnssec=no" to "configure" (and of course, | |
1016 | during runtime with DNSSEC= in resolved.conf). We recommend | |
1017 | downstreams to leave this on at least during development cycles and | |
1018 | report any issues with the DNSSEC logic upstream. We are very | |
1019 | interested in collecting feedback about the DNSSEC validator and its | |
1020 | limitations in the wild. Note however, that DNSSEC support is | |
1021 | probably nothing downstreams should turn on in stable distros just | |
96d49011 | 1022 | yet, as it might create incompatibilities with a few DNS servers and |
61ecb465 LP |
1023 | networks. We tried hard to make sure we downgrade to non-DNSSEC mode |
1024 | automatically whenever we detect such incompatible setups, but there | |
1025 | might be systems we do not cover yet. Hence: please help us testing | |
1026 | the DNSSEC code, leave this on where you can, report back, but then | |
1027 | again don't consider turning this on in your stable, LTS or | |
e40a326c LP |
1028 | production release just yet. (Note that you have to enable |
1029 | nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved | |
1030 | and its DNSSEC mode for host name resolution from local | |
1031 | applications.) | |
61ecb465 | 1032 | |
96515dbf | 1033 | * systemd-resolve conveniently resolves DANE records with the --tlsa |
e40a326c | 1034 | option and OPENPGPKEY records with the --openpgp option. It also |
e75690c3 | 1035 | supports dumping raw DNS record data via the new --raw= switch. |
96515dbf | 1036 | |
97e5530c ZJS |
1037 | * systemd-logind will now by default terminate user processes that are |
1038 | part of the user session scope unit (session-XX.scope) when the user | |
977f2bea | 1039 | logs out. This behavior is controlled by the KillUserProcesses= |
e40a326c LP |
1040 | setting in logind.conf, and the previous default of "no" is now |
1041 | changed to "yes". This means that user sessions will be properly | |
1042 | cleaned up after, but additional steps are necessary to allow | |
1043 | intentionally long-running processes to survive logout. | |
97e5530c ZJS |
1044 | |
1045 | While the user is logged in at least once, user@.service is running, | |
1046 | and any service that should survive the end of any individual login | |
1047 | session can be started at a user service or scope using systemd-run. | |
e40a326c | 1048 | systemd-run(1) man page has been extended with an example which shows |
8951eaec | 1049 | how to run screen in a scope unit underneath user@.service. The same |
e40a326c | 1050 | command works for tmux. |
97e5530c ZJS |
1051 | |
1052 | After the user logs out of all sessions, user@.service will be | |
1053 | terminated too, by default, unless the user has "lingering" enabled. | |
1054 | To effectively allow users to run long-term tasks even if they are | |
152199f2 ZJS |
1055 | logged out, lingering must be enabled for them. See loginctl(1) for |
1056 | details. The default polkit policy was modified to allow users to | |
1057 | set lingering for themselves without authentication. | |
7f6e8043 | 1058 | |
95365a57 | 1059 | Previous defaults can be restored at compile time by the |
e40a326c | 1060 | --without-kill-user-processes option to "configure". |
7f6e8043 | 1061 | |
e75690c3 ZJS |
1062 | * systemd-logind gained new configuration settings SessionsMax= and |
1063 | InhibitorsMax=, both with a default of 8192. It will not register new | |
188d3082 | 1064 | user sessions or inhibitors above this limit. |
e75690c3 ZJS |
1065 | |
1066 | * systemd-logind will now reload configuration on SIGHUP. | |
1067 | ||
96515dbf | 1068 | * The unified cgroup hierarchy added in Linux 4.5 is now supported. |
e40a326c | 1069 | Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to |
25b0e6cb LP |
1070 | enable. Also, support for the "io" cgroup controller in the unified |
1071 | hierarchy has been added, so that the "memory", "pids" and "io" are | |
1072 | now the controllers that are supported on the unified hierarchy. | |
e40a326c | 1073 | |
96515dbf ZJS |
1074 | WARNING: it is not possible to use previous systemd versions with |
1075 | systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it | |
1076 | is necessary to also update systemd in the initramfs if using the | |
e40a326c | 1077 | unified hierarchy. An updated SELinux policy is also required. |
96515dbf | 1078 | |
e40a326c LP |
1079 | * LLDP support has been extended, and both passive (receive-only) and |
1080 | active (sender) modes are supported. Passive mode ("routers-only") is | |
8951eaec ZJS |
1081 | enabled by default in systemd-networkd. Active LLDP mode is enabled |
1082 | by default for containers on the internal network. The "networkctl | |
e40a326c LP |
1083 | lldp" command may be used to list information gathered. "networkctl |
1084 | status" will also show basic LLDP information on connected peers now. | |
96515dbf | 1085 | |
e40a326c LP |
1086 | * The IAID and DUID unique identifier sent in DHCP requests may now be |
1087 | configured for the system and each .network file managed by | |
e75690c3 ZJS |
1088 | systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options. |
1089 | ||
1090 | * systemd-networkd gained support for configuring proxy ARP support for | |
1091 | each interface, via the ProxyArp= setting in .network files. It also | |
1092 | gained support for configuring the multicast querier feature of | |
1093 | bridge devices, via the new MulticastQuerier= setting in .netdev | |
1094 | files. Similarly, snooping on the IGMP traffic can be controlled | |
1095 | via the new setting MulticastSnooping=. | |
1096 | ||
1097 | A new setting PreferredLifetime= has been added for addresses | |
1098 | configured in .network file to configure the lifetime intended for an | |
1099 | address. | |
1100 | ||
1101 | The systemd-networkd DHCP server gained the option EmitRouter=, which | |
1102 | defaults to yes, to configure whether the DHCP Option 3 (Router) | |
1103 | should be emitted. | |
96515dbf | 1104 | |
e40a326c | 1105 | * The testing tool /usr/lib/systemd/systemd-activate is renamed to |
97e5530c ZJS |
1106 | systemd-socket-activate and installed into /usr/bin. It is now fully |
1107 | supported. | |
1108 | ||
e40a326c LP |
1109 | * systemd-journald now uses separate threads to flush changes to disk |
1110 | when closing journal files, thus reducing impact of slow disk I/O on | |
1111 | logging performance. | |
96515dbf | 1112 | |
e75690c3 ZJS |
1113 | * The sd-journal API gained two new calls |
1114 | sd_journal_open_directory_fd() and sd_journal_open_files_fd() which | |
1115 | can be used to open journal files using file descriptors instead of | |
1116 | file or directory paths. sd_journal_open_container() has been | |
1117 | deprecated, sd_journal_open_directory_fd() should be used instead | |
1118 | with the flag SD_JOURNAL_OS_ROOT. | |
1119 | ||
1120 | * journalctl learned a new output mode "-o short-unix" that outputs log | |
1121 | lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970 | |
1122 | UTC). It also gained support for a new --no-hostname setting to | |
1123 | suppress the hostname column in the family of "short" output modes. | |
1124 | ||
e40a326c LP |
1125 | * systemd-ask-password now optionally skips printing of the password to |
1126 | stdout with --no-output which can be useful in scripts. | |
96515dbf ZJS |
1127 | |
1128 | * Framebuffer devices (/dev/fb*) and 3D printers and scanners | |
1129 | (devices tagged with ID_MAKER_TOOL) are now tagged with | |
1130 | "uaccess" and are available to logged in users. | |
1131 | ||
e75690c3 | 1132 | * The DeviceAllow= unit setting now supports specifiers (with "%"). |
e40a326c LP |
1133 | |
1134 | * "systemctl show" gained a new --value switch, which allows print a | |
1135 | only the contents of a specific unit property, without also printing | |
8951eaec ZJS |
1136 | the property's name. Similar support was added to "show*" verbs |
1137 | of loginctl and machinectl that output "key=value" lists. | |
e40a326c | 1138 | |
e75690c3 ZJS |
1139 | * A new unit type "generated" was added for files dynamically generated |
1140 | by generator tools. Similarly, a new unit type "transient" is used | |
1141 | for unit files created using the runtime API. "systemctl enable" will | |
1142 | refuse to operate on such files. | |
1143 | ||
e40a326c LP |
1144 | * A new command "systemctl revert" has been added that may be used to |
1145 | revert to the vendor version of a unit file, in case local changes | |
1146 | have been made by adding drop-ins or overriding the unit file. | |
1147 | ||
1148 | * "machinectl clean" gained a new verb to automatically remove all or | |
1149 | just hidden container images. | |
1150 | ||
e40a326c LP |
1151 | * systemd-tmpfiles gained support for a new line type "e" for emptying |
1152 | directories, if they exist, without creating them if they don't. | |
1153 | ||
e40a326c LP |
1154 | * systemd-nspawn gained support for automatically patching the UID/GIDs |
1155 | of the owners and the ACLs of all files and directories in a | |
1156 | container tree to match the UID/GID user namespacing range selected | |
1157 | for the container invocation. This mode is enabled via the new | |
24597ee0 ZJS |
1158 | --private-users-chown switch. It also gained support for |
1159 | automatically choosing a free, previously unused UID/GID range when | |
1160 | starting a container, via the new --private-users=pick setting (which | |
1161 | implies --private-users-chown). Together, these options for the first | |
1162 | time make user namespacing for nspawn containers fully automatic and | |
1163 | thus deployable. The systemd-nspawn@.service template unit file has | |
1164 | been changed to use this functionality by default. | |
e40a326c | 1165 | |
25b0e6cb LP |
1166 | * systemd-nspawn gained a new --network-zone= switch, that allows |
1167 | creating ad-hoc virtual Ethernet links between multiple containers, | |
1168 | that only exist as long as at least one container referencing them is | |
1169 | running. This allows easy connecting of multiple containers with a | |
1170 | common link that implements an Ethernet broadcast domain. Each of | |
1171 | these network "zones" may be named relatively freely by the user, and | |
1172 | may be referenced by any number of containers, but each container may | |
1173 | only reference one of these "zones". On the lower level, this is | |
1174 | implemented by an automatically managed bridge network interface for | |
1175 | each zone, that is created when the first container referencing its | |
1176 | zone is created and removed when the last one referencing its zone | |
1177 | terminates. | |
1178 | ||
e40a326c | 1179 | * The default start timeout may now be configured on the kernel command |
8951eaec ZJS |
1180 | line via systemd.default_timeout_start_sec=. It was already |
1181 | configurable via the DefaultTimeoutStartSec= option in | |
1182 | /etc/systemd/system.conf. | |
e40a326c | 1183 | |
030bd839 | 1184 | * Socket units gained a new TriggerLimitIntervalSec= and |
e40a326c LP |
1185 | TriggerLimitBurst= setting to configure a limit on the activation |
1186 | rate of the socket unit. | |
1187 | ||
1188 | * The LimitNICE= setting now optionally takes normal UNIX nice values | |
1189 | in addition to the raw integer limit value. If the specified | |
1190 | parameter is prefixed with "+" or "-" and is in the range -20..19 the | |
1191 | value is understood as UNIX nice value. If not prefixed like this it | |
1192 | is understood as raw RLIMIT_NICE limit. | |
1193 | ||
999a43f8 LP |
1194 | * Note that the effect of the PrivateDevices= unit file setting changed |
1195 | slightly with this release: the per-device /dev file system will be | |
1196 | mounted read-only from this version on, and will have "noexec" | |
188d3082 | 1197 | set. This (minor) change of behavior might cause some (exceptional) |
999a43f8 LP |
1198 | legacy software to break, when PrivateDevices=yes is set for its |
1199 | service. Please leave PrivateDevices= off if you run into problems | |
1200 | with this. | |
1201 | ||
e75690c3 ZJS |
1202 | * systemd-bootchart has been split out to a separate repository: |
1203 | https://github.com/systemd/systemd-bootchart | |
1204 | ||
1205 | * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be | |
1206 | merged into the kernel in its current form. | |
1207 | ||
1208 | * The compatibility libraries libsystemd-daemon.so, | |
1209 | libsystemd-journal.so, libsystemd-id128.so, and libsystemd-login.so | |
1210 | which have been deprecated since systemd-209 have been removed along | |
1211 | with the corresponding pkg-config files. All symbols provided by | |
1212 | those libraries are provided by libsystemd.so. | |
1213 | ||
1214 | * The Capabilities= unit file setting has been removed (it is ignored | |
1215 | for backwards compatibility). AmbientCapabilities= and | |
1216 | CapabilityBoundingSet= should be used instead. | |
1217 | ||
4f9020fa DR |
1218 | * A new special target has been added, initrd-root-device.target, |
1219 | which creates a synchronization point for dependencies of the root | |
1220 | device in early userspace. Initramfs builders must ensure that this | |
1221 | target is now included in early userspace. | |
1222 | ||
e75690c3 ZJS |
1223 | Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov, |
1224 | Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin | |
1225 | Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens | |
1226 | Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh, | |
1227 | Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David | |
1228 | R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny | |
1229 | Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck | |
1230 | Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik | |
77ff6022 CG |
1231 | Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo |
1232 | Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth, | |
e75690c3 ZJS |
1233 | John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos |
1234 | Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir | |
1235 | Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt, | |
77ff6022 CG |
1236 | Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný, |
1237 | Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin, | |
1238 | mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween, | |
e75690c3 ZJS |
1239 | Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern, |
1240 | Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert | |
1241 | Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan | |
1242 | Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain | |
1243 | Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller, | |
1244 | Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen, | |
46e40fab ZJS |
1245 | Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso, |
1246 | Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev, | |
1247 | Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew | |
1248 | Jędrzejewski-Szmek | |
e40a326c | 1249 | |
46e40fab | 1250 | — Fairfax, 2016-05-21 |
96515dbf | 1251 | |
61f32bff MP |
1252 | CHANGES WITH 229: |
1253 | ||
d5f8b295 LP |
1254 | * The systemd-resolved DNS resolver service has gained a substantial |
1255 | set of new features, most prominently it may now act as a DNSSEC | |
1256 | validating stub resolver. DNSSEC mode is currently turned off by | |
ed5f8840 ZJS |
1257 | default, but is expected to be turned on by default in one of the |
1258 | next releases. For now, we invite everybody to test the DNSSEC logic | |
1259 | by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The | |
1260 | service also gained a full set of D-Bus interfaces, including calls | |
1261 | to configure DNS and DNSSEC settings per link (for use by external | |
1262 | network management software). systemd-resolved and systemd-networkd | |
1263 | now distinguish between "search" and "routing" domains. The former | |
1264 | are used to qualify single-label names, the latter are used purely | |
1265 | for routing lookups within certain domains to specific links. | |
1266 | resolved now also synthesizes RRs for all entries from /etc/hosts. | |
d5f8b295 LP |
1267 | |
1268 | * The systemd-resolve tool (which is a client utility for | |
ed5f8840 ZJS |
1269 | systemd-resolved) has been improved considerably and is now fully |
1270 | supported and documented. Hence it has moved from /usr/lib/systemd to | |
1271 | /usr/bin. | |
d5f8b295 LP |
1272 | |
1273 | * /dev/disk/by-path/ symlink support has been (re-)added for virtio | |
1274 | devices. | |
1275 | ||
a7c723c0 LP |
1276 | * The coredump collection logic has been reworked: when a coredump is |
1277 | collected it is now written to disk, compressed and processed | |
1278 | (including stacktrace extraction) from a new instantiated service | |
1279 | systemd-coredump@.service, instead of directly from the | |
1280 | /proc/sys/kernel/core_pattern hook we provide. This is beneficial as | |
1281 | processing large coredumps can take up a substantial amount of | |
1282 | resources and time, and this previously happened entirely outside of | |
1283 | systemd's service supervision. With the new logic the core_pattern | |
1284 | hook only does minimal metadata collection before passing off control | |
1285 | to the new instantiated service, which is configured with a time | |
1286 | limit, a nice level and other settings to minimize negative impact on | |
1287 | the rest of the system. Also note that the new logic will honour the | |
1288 | RLIMIT_CORE setting of the crashed process, which now allows users | |
1289 | and processes to turn off coredumping for their processes by setting | |
1290 | this limit. | |
1291 | ||
1292 | * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1 | |
1293 | and all forked processes by default. Previously, PID 1 would leave | |
1294 | the setting at "0" for all processes, as set by the kernel. Note that | |
1295 | the resource limit traditionally has no effect on the generated | |
1296 | coredumps on the system if the /proc/sys/kernel/core_pattern hook | |
1297 | logic is used. Since the limit is now honoured (see above) its | |
1298 | default has been changed so that the coredumping logic is enabled by | |
1299 | default for all processes, while allowing specific opt-out. | |
1300 | ||
1301 | * When the stacktrace is extracted from processes of system users, this | |
1302 | is now done as "systemd-coredump" user, in order to sandbox this | |
1303 | potentially security sensitive parsing operation. (Note that when | |
1304 | processing coredumps of normal users this is done under the user ID | |
1305 | of process that crashed, as before.) Packagers should take notice | |
1306 | that it is now necessary to create the "systemd-coredump" system user | |
1307 | and group at package installation time. | |
1308 | ||
d5f8b295 LP |
1309 | * The systemd-activate socket activation testing tool gained support |
1310 | for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram | |
1311 | and --seqpacket switches. It also has been extended to support both | |
1312 | new-style and inetd-style file descriptor passing. Use the new | |
1313 | --inetd switch to request inetd-style file descriptor passing. | |
1314 | ||
8968aea0 MP |
1315 | * Most systemd tools now honor a new $SYSTEMD_COLORS environment |
1316 | variable, which takes a boolean value. If set to false, ANSI color | |
d5f8b295 LP |
1317 | output is disabled in the tools even when run on a terminal that |
1318 | supports it. | |
1319 | ||
1320 | * The VXLAN support in networkd now supports two new settings | |
1321 | DestinationPort= and PortRange=. | |
1322 | ||
1323 | * A new systemd.machine_id= kernel command line switch has been added, | |
1324 | that may be used to set the machine ID in /etc/machine-id if it is | |
1325 | not initialized yet. This command line option has no effect if the | |
1326 | file is already initialized. | |
1327 | ||
1328 | * systemd-nspawn gained a new --as-pid2 switch that invokes any | |
1329 | specified command line as PID 2 rather than PID 1 in the | |
ed5f8840 ZJS |
1330 | container. In this mode PID 1 is a minimal stub init process that |
1331 | implements the special POSIX and Linux semantics of PID 1 regarding | |
1332 | signal and child process management. Note that this stub init process | |
1333 | is implemented in nspawn itself and requires no support from the | |
1334 | container image. This new logic is useful to support running | |
1335 | arbitrary commands in the container, as normal processes are | |
d5f8b295 LP |
1336 | generally not prepared to run as PID 1. |
1337 | ||
1338 | * systemd-nspawn gained a new --chdir= switch for setting the current | |
1339 | working directory for the process started in the container. | |
1340 | ||
ed5f8840 ZJS |
1341 | * "journalctl /dev/sda" will now output all kernel log messages for |
1342 | specified device from the current boot, in addition to all devices | |
1343 | that are parents of it. This should make log output about devices | |
1344 | pretty useful, as long as kernel drivers attach enough metadata to | |
1345 | the log messages. (The usual SATA drivers do.) | |
d5f8b295 LP |
1346 | |
1347 | * The sd-journal API gained two new calls | |
1348 | sd_journal_has_runtime_files() and sd_journal_has_persistent_files() | |
1349 | that report whether log data from /run or /var has been found. | |
1350 | ||
1351 | * journalctl gained a new switch "--fields" that prints all journal | |
1352 | record field names currently in use in the journal. This is backed | |
1353 | by two new sd-journal API calls sd_journal_enumerate_fields() and | |
1354 | sd_journal_restart_fields(). | |
1355 | ||
1356 | * Most configurable timeouts in systemd now expect an argument of | |
8968aea0 MP |
1357 | "infinity" to turn them off, instead of "0" as before. The semantics |
1358 | from now on is that a timeout of "0" means "now", and "infinity" | |
1359 | means "never". To maintain backwards compatibility, "0" continues to | |
1360 | turn off previously existing timeout settings. | |
d5f8b295 LP |
1361 | |
1362 | * "systemctl reload-or-try-restart" has been renamed to "systemctl | |
8968aea0 MP |
1363 | try-reload-or-restart" to clarify what it actually does: the "try" |
1364 | logic applies to both reloading and restarting, not just restarting. | |
1365 | The old name continues to be accepted for compatibility. | |
1366 | ||
1367 | * On boot-up, when PID 1 detects that the system clock is behind the | |
1368 | release date of the systemd version in use, the clock is now set | |
1369 | to the latter. Previously, this was already done in timesyncd, in order | |
d5f8b295 LP |
1370 | to avoid running with clocks set to the various clock epochs such as |
1371 | 1902, 1938 or 1970. With this change the logic is now done in PID 1 | |
1372 | in addition to timesyncd during early boot-up, so that it is enforced | |
1373 | before the first process is spawned by systemd. Note that the logic | |
1374 | in timesyncd remains, as it is more comprehensive and ensures | |
ed5f8840 | 1375 | clock monotonicity by maintaining a persistent timestamp file in |
d5f8b295 LP |
1376 | /var. Since /var is generally not available in earliest boot or the |
1377 | initrd, this part of the logic remains in timesyncd, and is not done | |
1378 | by PID 1. | |
1379 | ||
50f48ad3 DM |
1380 | * Support for tweaking details in net_cls.class_id through the |
1381 | NetClass= configuration directive has been removed, as the kernel | |
1382 | people have decided to deprecate that controller in cgroup v2. | |
1383 | Userspace tools such as nftables are moving over to setting rules | |
1384 | that are specific to the full cgroup path of a task, which obsoletes | |
1385 | these controllers anyway. The NetClass= directive is kept around for | |
1386 | legacy compatibility reasons. For a more in-depth description of the | |
1387 | kernel change, please refer to the respective upstream commit: | |
1388 | ||
1389 | https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671 | |
1390 | ||
d5f8b295 | 1391 | * A new service setting RuntimeMaxSec= has been added that may be used |
8968aea0 | 1392 | to specify a maximum runtime for a service. If the timeout is hit, the |
d5f8b295 LP |
1393 | service is terminated and put into a failure state. |
1394 | ||
8968aea0 MP |
1395 | * A new service setting AmbientCapabilities= has been added. It allows |
1396 | configuration of additional Linux process capabilities that are | |
1397 | passed to the activated processes. This is only available on very | |
d5f8b295 LP |
1398 | recent kernels. |
1399 | ||
1400 | * The process resource limit settings in service units may now be used | |
1401 | to configure hard and soft limits individually. | |
1402 | ||
8968aea0 | 1403 | * The various libsystemd APIs such as sd-bus or sd-event now publicly |
ed5f8840 ZJS |
1404 | expose support for gcc's __attribute__((cleanup())) C extension. |
1405 | Specifically, for many object destructor functions alternative | |
1406 | versions have been added that have names suffixed with "p" and take a | |
1407 | pointer to a pointer to the object to destroy, instead of just a | |
1408 | pointer to the object itself. This is useful because these destructor | |
1409 | functions may be used directly as parameters to the cleanup | |
1410 | construct. Internally, systemd has been a heavy user of this GCC | |
1411 | extension for a long time, and with this change similar support is | |
1412 | now available to consumers of the library outside of systemd. Note | |
8968aea0 | 1413 | that by using this extension in your sources compatibility with old |
ed5f8840 ZJS |
1414 | and strictly ANSI compatible C compilers is lost. However, all gcc or |
1415 | LLVM versions of recent years support this extension. | |
d5f8b295 LP |
1416 | |
1417 | * Timer units gained support for a new setting RandomizedDelaySec= that | |
8968aea0 MP |
1418 | allows configuring some additional randomized delay to the configured |
1419 | time. This is useful to spread out timer events to avoid load peaks in | |
1420 | clusters or larger setups. | |
d5f8b295 LP |
1421 | |
1422 | * Calendar time specifications now support sub-second accuracy. | |
1423 | ||
1424 | * Socket units now support listening on SCTP and UDP-lite protocol | |
1425 | sockets. | |
1426 | ||
1427 | * The sd-event API now comes with a full set of man pages. | |
1428 | ||
1429 | * Older versions of systemd contained experimental support for | |
1430 | compressing journal files and coredumps with the LZ4 compressor that | |
1431 | was not compatible with the lz4 binary (due to API limitations of the | |
1432 | lz4 library). This support has been removed; only support for files | |
1433 | compatible with the lz4 binary remains. This LZ4 logic is now | |
1434 | officially supported and no longer considered experimental. | |
1435 | ||
1436 | * The dkr image import logic has been removed again from importd. dkr's | |
1437 | micro-services focus doesn't fit into the machine image focus of | |
1438 | importd, and quickly got out of date with the upstream dkr API. | |
1439 | ||
1440 | * Creation of the /run/lock/lockdev/ directory was dropped from | |
1441 | tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have | |
61f32bff MP |
1442 | been available for many years. If you still need this, you need to |
1443 | create your own tmpfiles.d config file with: | |
d5f8b295 LP |
1444 | |
1445 | d /run/lock/lockdev 0775 root lock - | |
61f32bff | 1446 | |
3545ab35 LP |
1447 | Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander |
1448 | Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov, | |
1449 | Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler, | |
1450 | Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan | |
1451 | Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack, | |
1452 | David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman, | |
1453 | Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen, | |
1454 | Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen, | |
1455 | Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub | |
1456 | Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek, | |
1457 | Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos, | |
1458 | lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel | |
1459 | Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer, | |
1460 | Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils | |
1461 | Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz, | |
1462 | Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant | |
1463 | Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel | |
1464 | Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito | |
1465 | Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek | |
1466 | ||
ccddd104 | 1467 | — Berlin, 2016-02-11 |
61f32bff | 1468 | |
a11c7ea5 LP |
1469 | CHANGES WITH 228: |
1470 | ||
a11c7ea5 LP |
1471 | * A number of properties previously only settable in unit |
1472 | files are now also available as properties to set when | |
1473 | creating transient units programmatically via the bus, as it | |
1474 | is exposed with systemd-run's --property= | |
1475 | setting. Specifically, these are: SyslogIdentifier=, | |
1476 | SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=, | |
1477 | EnvironmentFile=, ReadWriteDirectories=, | |
1478 | ReadOnlyDirectories=, InaccessibleDirectories=, | |
1479 | ProtectSystem=, ProtectHome=, RuntimeDirectory=. | |
1480 | ||
28c85daf LP |
1481 | * When creating transient services via the bus API it is now |
1482 | possible to pass in a set of file descriptors to use as | |
1483 | STDIN/STDOUT/STDERR for the invoked process. | |
a11c7ea5 | 1484 | |
f1f8a5a5 LP |
1485 | * Slice units may now be created transiently via the bus APIs, |
1486 | similar to the way service and scope units may already be | |
1487 | created transiently. | |
1488 | ||
a11c7ea5 LP |
1489 | * Wherever systemd expects a calendar timestamp specification |
1490 | (like in journalctl's --since= and --until= switches) UTC | |
1491 | timestamps are now supported. Timestamps suffixed with "UTC" | |
1492 | are now considered to be in Universal Time Coordinated | |
1493 | instead of the local timezone. Also, timestamps may now | |
815bb5bd | 1494 | optionally be specified with sub-second accuracy. Both of |
a11c7ea5 LP |
1495 | these additions also apply to recurring calendar event |
1496 | specification, such as OnCalendar= in timer units. | |
1497 | ||
28c85daf LP |
1498 | * journalctl gained a new "--sync" switch that asks the |
1499 | journal daemon to write all so far unwritten log messages to | |
1500 | disk and sync the files, before returning. | |
1501 | ||
a11c7ea5 LP |
1502 | * systemd-tmpfiles learned two new line types "q" and "Q" that |
1503 | operate like "v", but also set up a basic btrfs quota | |
1504 | hierarchy when used on a btrfs file system with quota | |
1505 | enabled. | |
1506 | ||
f1f8a5a5 LP |
1507 | * tmpfiles' "v", "q" and "Q" will now create a plain directory |
1508 | instead of a subvolume (even on a btrfs file system) if the | |
1509 | root directory is a plain directory, and not a | |
1510 | subvolume. This should simplify things with certain chroot() | |
1511 | environments which are not aware of the concept of btrfs | |
1512 | subvolumes. | |
1513 | ||
a11c7ea5 LP |
1514 | * systemd-detect-virt gained a new --chroot switch to detect |
1515 | whether execution takes place in a chroot() environment. | |
1516 | ||
28c85daf | 1517 | * CPUAffinity= now takes CPU index ranges in addition to |
a11c7ea5 LP |
1518 | individual indexes. |
1519 | ||
28c85daf LP |
1520 | * The various memory-related resource limit settings (such as |
1521 | LimitAS=) now understand the usual K, M, G, ... suffixes to | |
1522 | the base of 1024 (IEC). Similar, the time-related resource | |
1523 | limit settings understand the usual min, h, day, ... | |
1524 | suffixes now. | |
1525 | ||
f1f8a5a5 LP |
1526 | * There's a new system.conf setting DefaultTasksMax= to |
1527 | control the default TasksMax= setting for services and | |
1528 | scopes running on the system. (TasksMax= is the primary | |
1529 | setting that exposes the "pids" cgroup controller on systemd | |
1530 | and was introduced in the previous systemd release.) The | |
1531 | setting now defaults to 512, which means services that are | |
1532 | not explicitly configured otherwise will only be able to | |
1533 | create 512 processes or threads at maximum, from this | |
1534 | version on. Note that this means that thread- or | |
1535 | process-heavy services might need to be reconfigured to set | |
1536 | TasksMax= to a higher value. It is sufficient to set | |
1537 | TasksMax= in these specific unit files to a higher value, or | |
1538 | even "infinity". Similar, there's now a logind.conf setting | |
1539 | UserTasksMax= that defaults to 4096 and limits the total | |
1540 | number of processes or tasks each user may own | |
1541 | concurrently. nspawn containers also have the TasksMax= | |
1542 | value set by default now, to 8192. Note that all of this | |
1543 | only has an effect if the "pids" cgroup controller is | |
1544 | enabled in the kernel. The general benefit of these changes | |
1545 | should be a more robust and safer system, that provides a | |
1546 | certain amount of per-service fork() bomb protection. | |
1547 | ||
28c85daf LP |
1548 | * systemd-nspawn gained the new --network-veth-extra= switch |
1549 | to define additional and arbitrarily-named virtual Ethernet | |
1550 | links between the host and the container. | |
1551 | ||
1552 | * A new service execution setting PassEnvironment= has been | |
1553 | added that allows importing select environment variables | |
1554 | from PID1's environment block into the environment block of | |
1555 | the service. | |
1556 | ||
ddb4b0d3 | 1557 | * Timer units gained support for a new RemainAfterElapse= |
595bfe7d | 1558 | setting which takes a boolean argument. It defaults to on, |
ddb4b0d3 LP |
1559 | exposing behaviour unchanged to previous releases. If set to |
1560 | off, timer units are unloaded after they elapsed if they | |
1561 | cannot elapse again. This is particularly useful for | |
1562 | transient timer units, which shall not stay around longer | |
1563 | than until they first elapse. | |
1564 | ||
a11c7ea5 | 1565 | * systemd will now bump the net.unix.max_dgram_qlen to 512 by |
28c85daf LP |
1566 | default now (the kernel default is 16). This is beneficial |
1567 | for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it | |
a11c7ea5 LP |
1568 | allows substantially larger numbers of queued |
1569 | datagrams. This should increase the capability of systemd to | |
1570 | parallelize boot-up, as logging and sd_notify() are unlikely | |
1571 | to stall execution anymore. If you need to change the value | |
1572 | from the new defaults, use the usual sysctl.d/ snippets. | |
1573 | ||
28c85daf LP |
1574 | * The compression framing format used by the journal or |
1575 | coredump processing has changed to be in line with what the | |
1576 | official LZ4 tools generate. LZ4 compression support in | |
1577 | systemd was considered unsupported previously, as the format | |
1578 | was not compatible with the normal tools. With this release | |
1579 | this has changed now, and it is hence safe for downstream | |
1580 | distributions to turn it on. While not compressing as well | |
815bb5bd | 1581 | as the XZ, LZ4 is substantially faster, which makes |
28c85daf LP |
1582 | it a good default choice for the compression logic in the |
1583 | journal and in coredump handling. | |
a11c7ea5 | 1584 | |
28c85daf LP |
1585 | * Any reference to /etc/mtab has been dropped from |
1586 | systemd. The file has been obsolete since a while, but | |
1587 | systemd refused to work on systems where it was incorrectly | |
815bb5bd | 1588 | set up (it should be a symlink or non-existent). Please make |
28c85daf LP |
1589 | sure to update to util-linux 2.27.1 or newer in conjunction |
1590 | with this systemd release, which also drops any reference to | |
1591 | /etc/mtab. If you maintain a distribution make sure that no | |
1592 | software you package still references it, as this is a | |
1593 | likely source of bugs. There's also a glibc bug pending, | |
1594 | asking for removal of any reference to this obsolete file: | |
1595 | ||
1596 | https://sourceware.org/bugzilla/show_bug.cgi?id=19108 | |
a11c7ea5 | 1597 | |
d5bd92bb LP |
1598 | Note that only util-linux versions built with |
1599 | --enable-libmount-force-mountinfo are supported. | |
1600 | ||
a11c7ea5 LP |
1601 | * Support for the ".snapshot" unit type has been removed. This |
1602 | feature turned out to be little useful and little used, and | |
1603 | has now been removed from the core and from systemctl. | |
1604 | ||
b9e2f7eb LP |
1605 | * The dependency types RequiresOverridable= and |
1606 | RequisiteOverridable= have been removed from systemd. They | |
1607 | have been used only very sparingly to our knowledge and | |
1608 | other options that provide a similar effect (such as | |
1609 | systemctl --mode=ignore-dependencies) are much more useful | |
1610 | and commonly used. Moreover, they were only half-way | |
1611 | implemented as the option to control behaviour regarding | |
1612 | these dependencies was never added to systemctl. By removing | |
1613 | these dependency types the execution engine becomes a bit | |
1614 | simpler. Unit files that use these dependencies should be | |
1615 | changed to use the non-Overridable dependency types | |
1616 | instead. In fact, when parsing unit files with these | |
1617 | options, that's what systemd will automatically convert them | |
1618 | too, but it will also warn, asking users to fix the unit | |
1619 | files accordingly. Removal of these dependency types should | |
1620 | only affect a negligible number of unit files in the wild. | |
1621 | ||
1622 | * Behaviour of networkd's IPForward= option changed | |
1623 | (again). It will no longer maintain a per-interface setting, | |
1624 | but propagate one way from interfaces where this is enabled | |
1625 | to the global kernel setting. The global setting will be | |
1626 | enabled when requested by a network that is set up, but | |
1627 | never be disabled again. This change was made to make sure | |
1628 | IPv4 and IPv6 behaviour regarding packet forwarding is | |
1629 | similar (as the Linux IPv6 stack does not support | |
1630 | per-interface control of this setting) and to minimize | |
1631 | surprises. | |
1632 | ||
28c85daf LP |
1633 | * In unit files the behaviour of %u, %U, %h, %s has |
1634 | changed. These specifiers will now unconditionally resolve | |
1635 | to the various user database fields of the user that the | |
1636 | systemd instance is running as, instead of the user | |
1637 | configured in the specific unit via User=. Note that this | |
1638 | effectively doesn't change much, as resolving of these | |
1639 | specifiers was already turned off in the --system instance | |
1640 | of systemd, as we cannot do NSS lookups from PID 1. In the | |
1641 | --user instance of systemd these specifiers where correctly | |
1642 | resolved, but hardly made any sense, since the user instance | |
1643 | lacks privileges to do user switches anyway, and User= is | |
ce830873 | 1644 | hence useless. Moreover, even in the --user instance of |
28c85daf LP |
1645 | systemd behaviour was awkward as it would only take settings |
1646 | from User= assignment placed before the specifier into | |
1647 | account. In order to unify and simplify the logic around | |
1648 | this the specifiers will now always resolve to the | |
1649 | credentials of the user invoking the manager (which in case | |
1650 | of PID 1 is the root user). | |
1651 | ||
1652 | Contributions from: Andrew Jones, Beniamino Galvani, Boyuan | |
1653 | Yang, Daniel Machon, Daniel Mack, David Herrmann, David | |
1654 | Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin, | |
f1f8a5a5 LP |
1655 | Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo |
1656 | Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan | |
1657 | Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers, | |
1658 | Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel | |
1659 | Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark | |
1660 | Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich, | |
1661 | Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens, | |
1662 | Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer, | |
1663 | Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden, | |
1664 | Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen, | |
1665 | Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew | |
1666 | Jędrzejewski-Szmek | |
28c85daf | 1667 | |
ccddd104 | 1668 | — Berlin, 2015-11-18 |
a11c7ea5 | 1669 | |
c97e586d DM |
1670 | CHANGES WITH 227: |
1671 | ||
1672 | * systemd now depends on util-linux v2.27. More specifically, | |
1673 | the newly added mount monitor feature in libmount now | |
1674 | replaces systemd's former own implementation. | |
1675 | ||
1676 | * libmount mandates /etc/mtab not to be regular file, and | |
1677 | systemd now enforces this condition at early boot. | |
1678 | /etc/mtab has been deprecated and warned about for a very | |
1679 | long time, so systems running systemd should already have | |
1680 | stopped having this file around as anything else than a | |
1681 | symlink to /proc/self/mounts. | |
1682 | ||
d046fb93 LP |
1683 | * Support for the "pids" cgroup controller has been added. It |
1684 | allows accounting the number of tasks in a cgroup and | |
c97e586d DM |
1685 | enforcing limits on it. This adds two new setting |
1686 | TasksAccounting= and TasksMax= to each unit, as well as a | |
6fd5517b | 1687 | global option DefaultTasksAccounting=. |
c97e586d DM |
1688 | |
1689 | * Support for the "net_cls" cgroup controller has been added. | |
fe08a30b LP |
1690 | It allows assigning a net class ID to each task in the |
1691 | cgroup, which can then be used in firewall rules and traffic | |
1692 | shaping configurations. Note that the kernel netfilter net | |
1693 | class code does not currently work reliably for ingress | |
1694 | packets on unestablished sockets. | |
c97e586d DM |
1695 | |
1696 | This adds a new config directive called NetClass= to CGroup | |
6fd5517b | 1697 | enabled units. Allowed values are positive numbers for fixed |
c97e586d DM |
1698 | assignments and "auto" for picking a free value |
1699 | automatically. | |
1700 | ||
21d86c61 DM |
1701 | * 'systemctl is-system-running' now returns 'offline' if the |
1702 | system is not booted with systemd. This command can now be | |
1703 | used as a substitute for 'systemd-notify --booted'. | |
1704 | ||
1705 | * Watchdog timeouts have been increased to 3 minutes for all | |
1706 | in-tree service files. Apparently, disk IO issues are more | |
1707 | frequent than we hoped, and user reported >1 minute waiting | |
1708 | for disk IO. | |
1709 | ||
1710 | * 'machine-id-commit' functionality has been merged into | |
1711 | 'machine-id-setup --commit'. The separate binary has been | |
1712 | removed. | |
1713 | ||
d046fb93 LP |
1714 | * The WorkingDirectory= directive in unit files may now be set |
1715 | to the special value '~'. In this case, the working | |
1716 | directory is set to the home directory of the user | |
1717 | configured in User=. | |
21d86c61 | 1718 | |
fe08a30b LP |
1719 | * "machinectl shell" will now open the shell in the home |
1720 | directory of the selected user by default. | |
1721 | ||
21d86c61 | 1722 | * The CrashChVT= configuration file setting is renamed to |
d046fb93 LP |
1723 | CrashChangeVT=, following our usual logic of not |
1724 | abbreviating unnecessarily. The old directive is still | |
1725 | supported for compat reasons. Also, this directive now takes | |
1726 | an integer value between 1 and 63, or a boolean value. The | |
1727 | formerly supported '-1' value for disabling stays around for | |
1728 | compat reasons. | |
21d86c61 | 1729 | |
fe08a30b | 1730 | * The PrivateTmp=, PrivateDevices=, PrivateNetwork=, |
8b5f9d15 | 1731 | NoNewPrivileges=, TTYPath=, WorkingDirectory= and |
fe08a30b LP |
1732 | RootDirectory= properties can now be set for transient |
1733 | units. | |
1734 | ||
1735 | * The systemd-analyze tool gained a new "set-log-target" verb | |
1736 | to change the logging target the system manager logs to | |
1737 | dynamically during runtime. This is similar to how | |
1738 | "systemd-analyze set-log-level" already changes the log | |
1739 | level. | |
1740 | ||
1741 | * In nspawn /sys is now mounted as tmpfs, with only a selected | |
1742 | set of subdirectories mounted in from the real sysfs. This | |
1743 | enhances security slightly, and is useful for ensuring user | |
1744 | namespaces work correctly. | |
1745 | ||
1746 | * Support for USB FunctionFS activation has been added. This | |
1747 | allows implementation of USB gadget services that are | |
1748 | activated as soon as they are requested, so that they don't | |
595bfe7d | 1749 | have to run continuously, similar to classic socket |
fe08a30b LP |
1750 | activation. |
1751 | ||
1752 | * The "systemctl exit" command now optionally takes an | |
1753 | additional parameter that sets the exit code to return from | |
1754 | the systemd manager when exiting. This is only relevant when | |
1755 | running the systemd user instance, or when running the | |
1756 | system instance in a container. | |
1757 | ||
1758 | * sd-bus gained the new API calls sd_bus_path_encode_many() | |
1759 | and sd_bus_path_decode_many() that allow easy encoding and | |
1760 | decoding of multiple identifier strings inside a D-Bus | |
1761 | object path. Another new call sd_bus_default_flush_close() | |
1762 | has been added to flush and close per-thread default | |
1763 | connections. | |
1764 | ||
1765 | * systemd-cgtop gained support for a -M/--machine= switch to | |
1766 | show the control groups within a certain container only. | |
1767 | ||
1768 | * "systemctl kill" gained support for an optional --fail | |
1769 | switch. If specified the requested operation will fail of no | |
1770 | processes have been killed, because the unit had no | |
1771 | processes attached, or similar. | |
1772 | ||
bdba9227 DM |
1773 | * A new systemd.crash_reboot=1 kernel command line option has |
1774 | been added that triggers a reboot after crashing. This can | |
1775 | also be set through CrashReboot= in systemd.conf. | |
1776 | ||
1777 | * The RuntimeDirectory= setting now understands unit | |
1778 | specifiers like %i or %f. | |
1779 | ||
ce830873 | 1780 | * A new (still internal) library API sd-ipv4acd has been added, |
fe08a30b LP |
1781 | that implements address conflict detection for IPv4. It's |
1782 | based on code from sd-ipv4ll, and will be useful for | |
1783 | detecting DHCP address conflicts. | |
1784 | ||
bdba9227 DM |
1785 | * File descriptors passed during socket activation may now be |
1786 | named. A new API sd_listen_fds_with_names() is added to | |
a8eaaee7 | 1787 | access the names. The default names may be overridden, |
bdba9227 DM |
1788 | either in the .socket file using the FileDescriptorName= |
1789 | parameter, or by passing FDNAME= when storing the file | |
1790 | descriptors using sd_notify(). | |
fe08a30b | 1791 | |
d046fb93 LP |
1792 | * systemd-networkd gained support for: |
1793 | ||
0053598f | 1794 | - Setting the IPv6 Router Advertisement settings via |
edf4126f | 1795 | IPv6AcceptRouterAdvertisements= in .network files. |
d046fb93 LP |
1796 | |
1797 | - Configuring the HelloTimeSec=, MaxAgeSec= and | |
1798 | ForwardDelaySec= bridge parameters in .netdev files. | |
1799 | ||
1800 | - Configuring PreferredSource= for static routes in | |
edf4126f | 1801 | .network files. |
fe08a30b | 1802 | |
bdba9227 DM |
1803 | * The "ask-password" framework used to query for LUKS harddisk |
1804 | passwords or SSL passwords during boot gained support for | |
1805 | caching passwords in the kernel keyring, if it is | |
1806 | available. This makes sure that the user only has to type in | |
1807 | a passphrase once if there are multiple objects to unlock | |
1808 | with the same one. Previously, such password caching was | |
1809 | available only when Plymouth was used; this moves the | |
1810 | caching logic into the systemd codebase itself. The | |
1811 | "systemd-ask-password" utility gained a new --keyname= | |
1812 | switch to control which kernel keyring key to use for | |
1813 | caching a password in. This functionality is also useful for | |
1814 | enabling display managers such as gdm to automatically | |
1815 | unlock the user's GNOME keyring if its passphrase, the | |
1816 | user's password and the harddisk password are the same, if | |
1817 | gdm-autologin is used. | |
fe08a30b LP |
1818 | |
1819 | * When downloading tar or raw images using "machinectl | |
1820 | pull-tar" or "machinectl pull-raw", a matching ".nspawn" | |
1821 | file is now also downloaded, if it is available and stored | |
1822 | next to the image file. | |
c97e586d | 1823 | |
91d0d699 LP |
1824 | * Units of type ".socket" gained a new boolean setting |
1825 | Writable= which is only useful in conjunction with | |
1826 | ListenSpecial=. If true, enables opening the specified | |
1827 | special file in O_RDWR mode rather than O_RDONLY mode. | |
1828 | ||
1829 | * systemd-rfkill has been reworked to become a singleton | |
1830 | service that is activated through /dev/rfkill on each rfkill | |
1831 | state change and saves the settings to disk. This way, | |
1832 | systemd-rfkill is now compatible with devices that exist | |
1833 | only intermittendly, and even restores state if the previous | |
1834 | system shutdown was abrupt rather than clean. | |
1835 | ||
d046fb93 LP |
1836 | * The journal daemon gained support for vacuuming old journal |
1837 | files controlled by the number of files that shall remain, | |
1838 | in addition to the already existing control by size and by | |
1839 | date. This is useful as journal interleaving performance | |
6dd6a9c4 | 1840 | degrades with too many separate journal files, and allows |
d046fb93 LP |
1841 | putting an effective limit on them. The new setting defaults |
1842 | to 100, but this may be changed by setting SystemMaxFiles= | |
1843 | and RuntimeMaxFiles= in journald.conf. Also, the | |
1844 | "journalctl" tool gained the new --vacuum-files= switch to | |
1845 | manually vacuum journal files to leave only the specified | |
1846 | number of files in place. | |
c48eb61f | 1847 | |
bdba9227 DM |
1848 | * udev will now create /dev/disk/by-path links for ATA devices |
1849 | on kernels where that is supported. | |
c30f086f | 1850 | |
efce0ffe | 1851 | * Galician, Serbian, Turkish and Korean translations were added. |
c97e586d | 1852 | |
61e6771c LP |
1853 | Contributions from: Aaro Koskinen, Alban Crequy, Beniamino |
1854 | Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao | |
1855 | (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David | |
1856 | Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin, | |
1857 | Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel | |
1858 | de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner, | |
1859 | Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen, | |
1860 | Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir | |
1861 | Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart | |
1862 | Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, | |
1863 | Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl, | |
1864 | Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike | |
1865 | Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma, | |
1866 | Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer, | |
1867 | Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani, | |
1868 | Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel | |
1869 | Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich, | |
1870 | Zbigniew Jędrzejewski-Szmek, Марко М. Костић | |
1871 | ||
ccddd104 | 1872 | — Berlin, 2015-10-07 |
c97e586d | 1873 | |
c9912c5e DH |
1874 | CHANGES WITH 226: |
1875 | ||
5e8d4254 LP |
1876 | * The DHCP implementation of systemd-networkd gained a set of |
1877 | new features: | |
1878 | ||
1879 | - The DHCP server now supports emitting DNS and NTP | |
1880 | information. It may be enabled and configured via | |
1881 | EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS | |
1882 | and NTP information is enabled, but no servers are | |
1883 | configured, the corresponding uplink information (if there | |
1884 | is any) is propagated. | |
1885 | ||
1886 | - Server and client now support transmission and reception | |
1887 | of timezone information. It can be configured via the | |
1888 | newly introduced network options UseTimezone=, | |
1889 | EmitTimezone=, and Timezone=. Transmission of timezone | |
1890 | information is enabled between host and containers by | |
1891 | default now: the container will change its local timezone | |
1892 | to what the host has set. | |
1893 | ||
1894 | - Lease timeouts can now be configured via | |
1895 | MaxLeaseTimeSec= and DefaultLeaseTimeSec=. | |
1896 | ||
1897 | - The DHCP server improved on the stability of | |
1898 | leases. Clients are more likely to get the same lease | |
1899 | information back, even if the server loses state. | |
1900 | ||
1901 | - The DHCP server supports two new configuration options to | |
1902 | control the lease address pool metrics, PoolOffset= and | |
1903 | PoolSize=. | |
1904 | ||
1905 | * The encapsulation limit of tunnels in systemd-networkd may | |
1906 | now be configured via 'EncapsulationLimit='. It allows | |
1907 | modifying the maximum additional levels of encapsulation | |
1908 | that are permitted to be prepended to a packet. | |
1909 | ||
1910 | * systemd now supports the concept of user buses replacing | |
1911 | session buses, if used with dbus-1.10 (and enabled via dbus | |
1912 | --enable-user-session). It previously only supported this on | |
1913 | kdbus-enabled systems, and this release expands this to | |
1914 | 'dbus-daemon' systems. | |
1915 | ||
1916 | * systemd-networkd now supports predictable interface names | |
1917 | for virtio devices. | |
1918 | ||
1919 | * systemd now optionally supports the new Linux kernel | |
1920 | "unified" control group hierarchy. If enabled via the kernel | |
1921 | command-line option 'systemd.unified_cgroup_hierarchy=1', | |
1922 | systemd will try to mount the unified cgroup hierarchy | |
1923 | directly on /sys/fs/cgroup. If not enabled, or not | |
1924 | available, systemd will fall back to the legacy cgroup | |
1925 | hierarchy setup, as before. Host system and containers can | |
1926 | mix and match legacy and unified hierarchies as they | |
856ca72b | 1927 | wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY |
5e8d4254 LP |
1928 | environment variable to individually select the hierarchy to |
1929 | use for executed containers. By default, nspawn will use the | |
1930 | unified hierarchy for the containers if the host uses the | |
1931 | unified hierarchy, and the legacy hierarchy otherwise. | |
1932 | Please note that at this point the unified hierarchy is an | |
1933 | experimental kernel feature and is likely to change in one | |
1934 | of the next kernel releases. Therefore, it should not be | |
1935 | enabled by default in downstream distributions yet. The | |
1936 | minimum required kernel version for the unified hierarchy to | |
1937 | work is 4.2. Note that when the unified hierarchy is used | |
1938 | for the first time delegated access to controllers is | |
1939 | safe. Because of this systemd-nspawn containers will get | |
1940 | access to controllers now, as will systemd user | |
1941 | sessions. This means containers and user sessions may now | |
1942 | manage their own resources, partitioning up what the system | |
1943 | grants them. | |
1944 | ||
1945 | * A new special scope unit "init.scope" has been introduced | |
1946 | that encapsulates PID 1 of the system. It may be used to | |
1947 | determine resource usage and enforce resource limits on PID | |
1948 | 1 itself. PID 1 hence moved out of the root of the control | |
1949 | group tree. | |
1950 | ||
1951 | * The cgtop tool gained support for filtering out kernel | |
1952 | threads when counting tasks in a control group. Also, the | |
1953 | count of processes is now recursively summed up by | |
1954 | default. Two options -k and --recursive= have been added to | |
1955 | revert to old behaviour. The tool has also been updated to | |
1956 | work correctly in containers now. | |
1957 | ||
1958 | * systemd-nspawn's --bind= and --bind-ro= options have been | |
1959 | extended to allow creation of non-recursive bind mounts. | |
1960 | ||
c626bf1d DM |
1961 | * libsystemd gained two new calls sd_pid_get_cgroup() and |
1962 | sd_peer_get_cgroup() which return the control group path of | |
5e8d4254 LP |
1963 | a process or peer of a connected AF_UNIX socket. This |
1964 | function call is particularly useful when implementing | |
1965 | delegated subtrees support in the control group hierarchy. | |
1966 | ||
1967 | * The "sd-event" event loop API of libsystemd now supports | |
1968 | correct dequeuing of real-time signals, without losing | |
1969 | signal events. | |
1970 | ||
1971 | * When systemd requests a PolicyKit decision when managing | |
1972 | units it will now add additional fields to the request, | |
1973 | including unit name and desired operation. This enables more | |
1974 | powerful PolicyKit policies, that make decisions depending | |
1975 | on these parameters. | |
c9912c5e | 1976 | |
47f5a38c LP |
1977 | * nspawn learnt support for .nspawn settings files, that may |
1978 | accompany the image files or directories of containers, and | |
1979 | may contain additional settings for the container. This is | |
1980 | an alternative to configuring container parameters via the | |
1981 | nspawn command line. | |
1982 | ||
2f77decc LP |
1983 | Contributions from: Cristian Rodríguez, Daniel Mack, David |
1984 | Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe | |
1985 | Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan | |
1986 | Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel | |
1987 | Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal | |
1988 | Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin | |
1989 | Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel | |
23d08d1b | 1990 | Andersen, Tom Gundersen, Torstein Husebø |
c9912c5e | 1991 | |
ccddd104 | 1992 | — Berlin, 2015-09-08 |
c9912c5e | 1993 | |
ec5249a2 DM |
1994 | CHANGES WITH 225: |
1995 | ||
5e8d4254 LP |
1996 | * machinectl gained a new verb 'shell' which opens a fresh |
1997 | shell on the target container or the host. It is similar to | |
1998 | the existing 'login' command of machinectl, but spawns the | |
1999 | shell directly without prompting for username or | |
2000 | password. The pseudo machine '.host' now refers to the local | |
2001 | host and is used by default. Hence, 'machinectl shell' can | |
2002 | be used as replacement for 'su -' which spawns a session as | |
2003 | a fresh systemd unit in a way that is fully isolated from | |
2004 | the originating session. | |
2005 | ||
2006 | * systemd-networkd learned to cope with private-zone DHCP | |
2007 | options and allows other programs to query the values. | |
2008 | ||
2009 | * SELinux access control when enabling/disabling units is no | |
2010 | longer enforced with this release. The previous | |
2011 | implementation was incorrect, and a new corrected | |
2012 | implementation is not yet available. As unit file operations | |
2013 | are still protected via PolicyKit and D-Bus policy this is | |
2014 | not a security problem. Yet, distributions which care about | |
2015 | optimal SELinux support should probably not stabilize on | |
2016 | this release. | |
2017 | ||
2018 | * sd-bus gained support for matches of type "arg0has=", that | |
2019 | test for membership of strings in string arrays sent in bus | |
2020 | messages. | |
2021 | ||
2022 | * systemd-resolved now dumps the contents of its DNS and LLMNR | |
2023 | caches to the logs on reception of the SIGUSR1 signal. This | |
2024 | is useful to debug DNS behaviour. | |
2025 | ||
2026 | * The coredumpctl tool gained a new --directory= option to | |
2027 | operate on journal files in a specific directory. | |
2028 | ||
2029 | * "systemctl reboot" and related commands gained a new | |
2030 | "--message=" option which may be used to set a free-text | |
2031 | wall message when shutting down or rebooting the | |
2032 | system. This message is also logged, which is useful for | |
2033 | figuring out the reason for a reboot or shutdown a | |
2034 | posteriori. | |
2035 | ||
2036 | * The "systemd-resolve-host" tool's -i switch now takes | |
2037 | network interface numbers as alternative to interface names. | |
2038 | ||
2039 | * A new unit file setting for services has been introduced: | |
2040 | UtmpMode= allows configuration of how precisely systemd | |
2041 | handles utmp and wtmp entries for the service if this is | |
2042 | enabled. This allows writing services that appear similar to | |
2043 | user sessions in the output of the "w", "who", "last" and | |
2044 | "lastlog" tools. | |
2045 | ||
2046 | * systemd-resolved will now locally synthesize DNS resource | |
2047 | records for the "localhost" and "gateway" domains as well as | |
2048 | the local hostname. This should ensure that clients querying | |
2049 | RRs via resolved will get similar results as those going via | |
2050 | NSS, if nss-myhostname is enabled. | |
2051 | ||
2052 | Contributions from: Alastair Hughes, Alex Crawford, Daniel | |
2053 | Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski, | |
2054 | Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan | |
2055 | Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers, | |
2056 | Kefeng Wang, Lennart Poettering, Major Hayden, Marcel | |
2057 | Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt | |
2058 | Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim, | |
2059 | Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer, | |
2060 | reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings, | |
2061 | Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe | |
2062 | Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts, | |
2063 | WaLyong Cho, Zbigniew Jędrzejewski-Szmek | |
e1439a14 | 2064 | |
ccddd104 | 2065 | — Berlin, 2015-08-27 |
ec5249a2 | 2066 | |
11811e85 DH |
2067 | CHANGES WITH 224: |
2068 | ||
10fa421c DH |
2069 | * The systemd-efi-boot-generator functionality was merged into |
2070 | systemd-gpt-auto-generator. | |
2071 | ||
5e8d4254 LP |
2072 | * systemd-networkd now supports Group Policy for vxlan |
2073 | devices. It can be enabled via the new boolean configuration | |
2074 | option called 'GroupPolicyExtension='. | |
10fa421c | 2075 | |
11811e85 DH |
2076 | Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David |
2077 | Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart | |
2078 | Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen | |
2079 | ||
ccddd104 | 2080 | — Berlin, 2015-07-31 |
11811e85 | 2081 | |
e57eaef8 DH |
2082 | CHANGES WITH 223: |
2083 | ||
2084 | * The python-systemd code has been removed from the systemd repository. | |
2085 | A new repository has been created which accommodates the code from | |
2086 | now on, and we kindly ask distributions to create a separate package | |
2087 | for this: https://github.com/systemd/python-systemd | |
2088 | ||
01608bc8 | 2089 | * The systemd daemon will now reload its main configuration |
e57eaef8 DH |
2090 | (/etc/systemd/system.conf) on daemon-reload. |
2091 | ||
2092 | * sd-dhcp now exposes vendor specific extensions via | |
2093 | sd_dhcp_lease_get_vendor_specific(). | |
2094 | ||
931618d0 DM |
2095 | * systemd-networkd gained a number of new configuration options. |
2096 | ||
2097 | - A new boolean configuration option for TAP devices called | |
37d54b93 | 2098 | 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the |
931618d0 DM |
2099 | device, thus allowing to send and receive GSO packets. |
2100 | ||
2101 | - A new tunnel configuration option called 'CopyDSCP='. | |
2102 | If enabled, the DSCP field of ip6 tunnels is copied into the | |
2103 | decapsulated packet. | |
2104 | ||
2105 | - A set of boolean bridge configuration options were added. | |
2106 | 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=', | |
2107 | and 'UnicastFlood=' are now parsed by networkd and applied to the | |
2108 | respective bridge link device via the respective IFLA_BRPORT_* | |
2109 | netlink attribute. | |
2110 | ||
2111 | - A new string configuration option to override the hostname sent | |
2112 | to a DHCP server, called 'Hostname='. If set and 'SendHostname=' | |
2113 | is true, networkd will use the configured hostname instead of the | |
2114 | system hostname when sending DHCP requests. | |
2115 | ||
2116 | - A new tunnel configuration option called 'IPv6FlowLabel='. If set, | |
2117 | networkd will configure the IPv6 flow-label of the tunnel device | |
2118 | according to RFC2460. | |
e57eaef8 | 2119 | |
f5f113f6 DH |
2120 | - The 'macvtap' virtual network devices are now supported, similar to |
2121 | the already supported 'macvlan' devices. | |
2122 | ||
e57eaef8 | 2123 | * systemd-resolved now implements RFC5452 to improve resilience against |
01608bc8 | 2124 | cache poisoning. Additionally, source port randomization is enabled |
e57eaef8 DH |
2125 | by default to further protect against DNS spoofing attacks. |
2126 | ||
2127 | * nss-mymachines now supports translating UIDs and GIDs of running | |
2128 | containers with user-namespaces enabled. If a container 'foo' | |
2129 | translates a host uid 'UID' to the container uid 'TUID', then | |
2130 | nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID' | |
2131 | (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are | |
2132 | mapped as 'vg-foo-TGID'. | |
2133 | ||
2134 | Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel | |
e4e66993 DH |
2135 | Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov, |
2136 | HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig), | |
2137 | Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers, | |
2138 | Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael | |
2139 | Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim, | |
2140 | Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo, | |
2141 | Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom | |
2142 | Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo, | |
2143 | Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek | |
2144 | ||
ccddd104 | 2145 | — Berlin, 2015-07-29 |
e57eaef8 | 2146 | |
0db83ad7 | 2147 | CHANGES WITH 222: |
5541c889 | 2148 | |
861b02eb KS |
2149 | * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules. |
2150 | There are no known issues with current sysfs, and udev does not need | |
2151 | or should be used to work around such bugs. | |
2152 | ||
2153 | * udev does no longer enable USB HID power management. Several reports | |
2154 | indicate, that some devices cannot handle that setting. | |
0db83ad7 DH |
2155 | |
2156 | * The udev accelerometer helper was removed. The functionality | |
2157 | is now fully included in iio-sensor-proxy. But this means, | |
2158 | older iio-sensor-proxy versions will no longer provide | |
2159 | accelerometer/orientation data with this systemd version. | |
2160 | Please upgrade iio-sensor-proxy to version 1.0. | |
2161 | ||
5541c889 DH |
2162 | * networkd gained a new configuration option IPv6PrivacyExtensions= |
2163 | which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions | |
2164 | for Stateless Address") on selected networks. | |
2165 | ||
9b361114 DM |
2166 | * For the sake of fewer build-time dependencies and less code in the |
2167 | main repository, the python bindings are about to be removed in the | |
2168 | next release. A new repository has been created which accommodates | |
2169 | the code from now on, and we kindly ask distributions to create a | |
2170 | separate package for this. The removal will take place in v223. | |
2171 | ||
2172 | https://github.com/systemd/python-systemd | |
2173 | ||
0db83ad7 DH |
2174 | Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera, |
2175 | Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack, | |
2176 | daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric | |
2177 | Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario, | |
5541c889 DH |
2178 | Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens |
2179 | (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering, | |
0db83ad7 DH |
2180 | Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal |
2181 | Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne, | |
2d1ca112 DH |
2182 | Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein |
2183 | Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek | |
0db83ad7 | 2184 | |
ccddd104 | 2185 | — Berlin, 2015-07-07 |
0db83ad7 | 2186 | |
0f0467e6 MP |
2187 | CHANGES WITH 221: |
2188 | ||
470e72d4 | 2189 | * The sd-bus.h and sd-event.h APIs have now been declared |
5f92d24f | 2190 | stable and have been added to the official interface of |
470e72d4 LP |
2191 | libsystemd.so. sd-bus implements an alternative D-Bus client |
2192 | library, that is relatively easy to use, very efficient and | |
2193 | supports both classic D-Bus as well as kdbus as transport | |
2194 | backend. sd-event is a generic event loop abstraction that | |
2195 | is built around Linux epoll, but adds features such as event | |
0aee49d5 | 2196 | prioritization or efficient timer handling. Both APIs are good |
470e72d4 LP |
2197 | choices for C programs looking for a bus and/or event loop |
2198 | implementation that is minimal and does not have to be | |
5f92d24f | 2199 | portable to other kernels. |
0f0467e6 | 2200 | |
470e72d4 LP |
2201 | * kdbus support is no longer compile-time optional. It is now |
2202 | always built-in. However, it can still be disabled at | |
2203 | runtime using the kdbus=0 kernel command line setting, and | |
c6551464 | 2204 | that setting may be changed to default to off, by specifying |
470e72d4 LP |
2205 | --disable-kdbus at build-time. Note though that the kernel |
2206 | command line setting has no effect if the kdbus.ko kernel | |
2207 | module is not installed, in which case kdbus is (obviously) | |
2208 | also disabled. We encourage all downstream distributions to | |
0aee49d5 | 2209 | begin testing kdbus by adding it to the kernel images in the |
470e72d4 LP |
2210 | development distributions, and leaving kdbus support in |
2211 | systemd enabled. | |
0f0467e6 | 2212 | |
470e72d4 LP |
2213 | * The minimal required util-linux version has been bumped to |
2214 | 2.26. | |
2215 | ||
2216 | * Support for chkconfig (--enable-chkconfig) was removed in | |
0aee49d5 | 2217 | favor of calling an abstraction tool |
470e72d4 LP |
2218 | /lib/systemd/systemd-sysv-install. This needs to be |
2219 | implemented for your distribution. See "SYSV INIT.D SCRIPTS" | |
2220 | in README for details. | |
2221 | ||
2222 | * If there's a systemd unit and a SysV init script for the | |
2223 | same service name, and the user executes "systemctl enable" | |
2224 | for it (or a related call), then this will now enable both | |
2225 | (or execute the related operation on both), not just the | |
2226 | unit. | |
2227 | ||
2228 | * The libudev API documentation has been converted from gtkdoc | |
2229 | into man pages. | |
2230 | ||
2231 | * gudev has been removed from the systemd tree, it is now an | |
2232 | external project. | |
2233 | ||
2234 | * The systemd-cgtop tool learnt a new --raw switch to generate | |
0aee49d5 | 2235 | "raw" (machine parsable) output. |
470e72d4 LP |
2236 | |
2237 | * networkd's IPForwarding= .network file setting learnt the | |
2238 | new setting "kernel", which ensures that networkd does not | |
2239 | change the IP forwarding sysctl from the default kernel | |
2240 | state. | |
2241 | ||
2242 | * The systemd-logind bus API now exposes a new boolean | |
2243 | property "Docked" that reports whether logind considers the | |
2244 | system "docked", i.e. connected to a docking station or not. | |
2245 | ||
2246 | Contributions from: Alex Crawford, Andreas Pokorny, Andrei | |
2247 | Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez, | |
2248 | Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann, | |
2249 | David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed | |
2250 | Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario, | |
2251 | Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek, | |
2252 | Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang, | |
2253 | Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart | |
2254 | Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario | |
2255 | Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich, | |
2256 | Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes, | |
b912e251 LP |
2257 | Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip |
2258 | Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani, | |
2259 | Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein | |
2260 | Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner | |
2261 | Fink, Zbigniew Jędrzejewski-Szmek | |
470e72d4 | 2262 | |
ccddd104 | 2263 | — Berlin, 2015-06-19 |
0f0467e6 | 2264 | |
481a0aa2 LP |
2265 | CHANGES WITH 220: |
2266 | ||
f7a73a25 DH |
2267 | * The gudev library has been extracted into a separate repository |
2268 | available at: https://git.gnome.org/browse/libgudev/ | |
2269 | It is now managed as part of the Gnome project. Distributions | |
2270 | are recommended to pass --disable-gudev to systemd and use | |
2271 | gudev from the Gnome project instead. gudev is still included | |
2272 | in systemd, for now. It will be removed soon, though. Please | |
2273 | also see the announcement-thread on systemd-devel: | |
56cadcb6 | 2274 | https://lists.freedesktop.org/archives/systemd-devel/2015-May/032070.html |
f7a73a25 | 2275 | |
481a0aa2 LP |
2276 | * systemd now exposes a CPUUsageNSec= property for each |
2277 | service unit on the bus, that contains the overall consumed | |
2278 | CPU time of a service (the sum of what each process of the | |
2279 | service consumed). This value is only available if | |
2280 | CPUAccounting= is turned on for a service, and is then shown | |
2281 | in the "systemctl status" output. | |
2282 | ||
2283 | * Support for configuring alternative mappings of the old SysV | |
2284 | runlevels to systemd targets has been removed. They are now | |
29d1fcb4 | 2285 | hardcoded in a way that runlevels 2, 3, 4 all map to |
481a0aa2 LP |
2286 | multi-user.target and 5 to graphical.target (which |
2287 | previously was already the default behaviour). | |
2288 | ||
2289 | * The auto-mounter logic gained support for mount point | |
2290 | expiry, using a new TimeoutIdleSec= setting in .automount | |
2291 | units. (Also available as x-systemd.idle-timeout= in /etc/fstab). | |
2292 | ||
2293 | * The EFI System Partition (ESP) as mounted to /boot by | |
2294 | systemd-efi-boot-generator will now be unmounted | |
29d1fcb4 | 2295 | automatically after 2 minutes of not being used. This should |
481a0aa2 LP |
2296 | minimize the risk of ESP corruptions. |
2297 | ||
2298 | * New /etc/fstab options x-systemd.requires= and | |
2299 | x-systemd.requires-mounts-for= are now supported to express | |
2300 | additional dependencies for mounts. This is useful for | |
2301 | journalling file systems that support external journal | |
2302 | devices or overlay file systems that require underlying file | |
2303 | systems to be mounted. | |
2304 | ||
2305 | * systemd does not support direct live-upgrades (via systemctl | |
2306 | daemon-reexec) from versions older than v44 anymore. As no | |
2307 | distribution we are aware of shipped such old versions in a | |
2308 | stable release this should not be problematic. | |
2309 | ||
2310 | * When systemd forks off a new per-connection service instance | |
2311 | it will now set the $REMOTE_ADDR environment variable to the | |
2312 | remote IP address, and $REMOTE_PORT environment variable to | |
2313 | the remote IP port. This behaviour is similar to the | |
2314 | corresponding environment variables defined by CGI. | |
2315 | ||
2316 | * systemd-networkd gained support for uplink failure | |
2317 | detection. The BindCarrier= option allows binding interface | |
2318 | configuration dynamically to the link sense of other | |
2319 | interfaces. This is useful to achieve behaviour like in | |
2320 | network switches. | |
2321 | ||
2322 | * systemd-networkd gained support for configuring the DHCP | |
2323 | client identifier to use when requesting leases. | |
2324 | ||
2325 | * systemd-networkd now has a per-network UseNTP= option to | |
2326 | configure whether NTP server information acquired via DHCP | |
2327 | is passed on to services like systemd-timesyncd. | |
2328 | ||
2329 | * systemd-networkd gained support for vti6 tunnels. | |
2330 | ||
1579dd2c LP |
2331 | * Note that systemd-networkd manages the sysctl variable |
2332 | /proc/sys/net/ipv[46]/conf/*/forwarding for each interface | |
2333 | it is configured for since v219. The variable controls IP | |
2334 | forwarding, and is a per-interface alternative to the global | |
2335 | /proc/sys/net/ipv[46]/ip_forward. This setting is | |
2336 | configurable in the IPForward= option, which defaults to | |
2337 | "no". This means if networkd is used for an interface it is | |
2338 | no longer sufficient to set the global sysctl option to turn | |
2339 | on IP forwarding! Instead, the .network file option | |
2340 | IPForward= needs to be turned on! Note that the | |
2341 | implementation of this behaviour was broken in v219 and has | |
2342 | been fixed in v220. | |
2343 | ||
481a0aa2 LP |
2344 | * Many bonding and vxlan options are now configurable in |
2345 | systemd-networkd. | |
2346 | ||
2347 | * systemd-nspawn gained a new --property= setting to set unit | |
2348 | properties for the container scope. This is useful for | |
ce830873 | 2349 | setting resource parameters (e.g. "CPUShares=500") on |
481a0aa2 LP |
2350 | containers started from the command line. |
2351 | ||
2352 | * systemd-nspawn gained a new --private-users= switch to make | |
2353 | use of user namespacing available on recent Linux kernels. | |
2354 | ||
2355 | * systemd-nspawn may now be called as part of a shell pipeline | |
2356 | in which case the pipes used for stdin and stdout are passed | |
2357 | directly to the process invoked in the container, without | |
2358 | indirection via a pseudo tty. | |
2359 | ||
2360 | * systemd-nspawn gained a new switch to control the UNIX | |
2361 | signal to use when killing the init process of the container | |
2362 | when shutting down. | |
2363 | ||
2364 | * systemd-nspawn gained a new --overlay= switch for mounting | |
2365 | overlay file systems into the container using the new kernel | |
2366 | overlayfs support. | |
2367 | ||
2368 | * When a container image is imported via systemd-importd and | |
2369 | the host file system is not btrfs, a loopback block device | |
2370 | file is created in /var/lib/machines.raw with a btrfs file | |
2371 | system inside. It is then mounted to /var/lib/machines to | |
2372 | enable btrfs features for container management. The loopback | |
2373 | file and btrfs file system is grown as needed when container | |
2374 | images are imported via systemd-importd. | |
2375 | ||
2376 | * systemd-machined/systemd-importd gained support for btrfs | |
2377 | quota, to enforce container disk space limits on disk. This | |
2378 | is exposed in "machinectl set-limit". | |
2379 | ||
2380 | * systemd-importd now can import containers from local .tar, | |
2381 | .raw and .qcow2 images, and export them to .tar and .raw. It | |
2382 | can also import dkr v2 images now from the network (on top | |
2383 | of v1 as before). | |
2384 | ||
2385 | * systemd-importd gained support for verifying downloaded | |
2386 | images with gpg2 (previously only gpg1 was supported). | |
2387 | ||
2388 | * systemd-machined, systemd-logind, systemd: most bus calls | |
2389 | are now accessible to unprivileged processes via | |
2390 | PolicyKit. Also, systemd-logind will now allow users to kill | |
2391 | their own sessions without further privileges or | |
2392 | authorization. | |
2393 | ||
2394 | * systemd-shutdownd has been removed. This service was | |
2395 | previously responsible for implementing scheduled shutdowns | |
2396 | as exposed in /usr/bin/shutdown's time parameter. This | |
2397 | functionality has now been moved into systemd-logind and is | |
2398 | accessible via a bus interface. | |
2399 | ||
2400 | * "systemctl reboot" gained a new switch --firmware-setup that | |
2401 | can be used to reboot into the EFI firmware setup, if that | |
2402 | is available. systemd-logind now exposes an API on the bus | |
2403 | to trigger such reboots, in case graphical desktop UIs want | |
2404 | to cover this functionality. | |
2405 | ||
2406 | * "systemctl enable", "systemctl disable" and "systemctl mask" | |
1579dd2c | 2407 | now support a new "--now" switch. If specified the units |
481a0aa2 LP |
2408 | that are enabled will also be started, and the ones |
2409 | disabled/masked also stopped. | |
2410 | ||
2411 | * The Gummiboot EFI boot loader tool has been merged into | |
1a2d5fbe DH |
2412 | systemd, and renamed to "systemd-boot". The bootctl tool has been |
2413 | updated to support systemd-boot. | |
481a0aa2 LP |
2414 | |
2415 | * An EFI kernel stub has been added that may be used to create | |
2416 | kernel EFI binaries that contain not only the actual kernel, | |
2417 | but also an initrd, boot splash, command line and OS release | |
2418 | information. This combined binary can then be signed as a | |
2419 | single image, so that the firmware can verify it all in one | |
1a2d5fbe | 2420 | step. systemd-boot has special support for EFI binaries created |
481a0aa2 LP |
2421 | like this and can extract OS release information from them |
2422 | and show them in the boot menu. This functionality is useful | |
2423 | to implement cryptographically verified boot schemes. | |
2424 | ||
2425 | * Optional support has been added to systemd-fsck to pass | |
2426 | fsck's progress report to an AF_UNIX socket in the file | |
2427 | system. | |
2428 | ||
2429 | * udev will no longer create device symlinks for all block | |
2430 | devices by default. A blacklist for excluding special block | |
2431 | devices from this logic has been turned into a whitelist | |
2432 | that requires picking block devices explicitly that require | |
2433 | device symlinks. | |
2434 | ||
2435 | * A new (currently still internal) API sd-device.h has been | |
2436 | added to libsystemd. This modernized API is supposed to | |
2437 | replace libudev eventually. In fact, already much of libudev | |
2438 | is now just a wrapper around sd-device.h. | |
2439 | ||
2440 | * A new hwdb database for storing metadata about pointing | |
2441 | stick devices has been added. | |
2442 | ||
2443 | * systemd-tmpfiles gained support for setting file attributes | |
2444 | similar to the "chattr" tool with new 'h' and 'H' lines. | |
2445 | ||
2446 | * systemd-journald will no longer unconditionally set the | |
2447 | btrfs NOCOW flag on new journal files. This is instead done | |
2448 | with tmpfiles snippet using the new 'h' line type. This | |
2449 | allows easy disabling of this logic, by masking the | |
2450 | journal-nocow.conf tmpfiles file. | |
2451 | ||
2452 | * systemd-journald will now translate audit message types to | |
2453 | human readable identifiers when writing them to the | |
2454 | journal. This should improve readability of audit messages. | |
2455 | ||
2456 | * The LUKS logic gained support for the offset= and skip= | |
2457 | options in /etc/crypttab, as previously implemented by | |
2458 | Debian. | |
2459 | ||
2460 | * /usr/lib/os-release gained a new optional field VARIANT= for | |
2461 | distributions that support multiple variants (such as a | |
2462 | desktop edition, a server edition, ...) | |
2463 | ||
2464 | Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy, | |
2465 | Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin | |
2466 | Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel, | |
2467 | Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž | |
2468 | Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian | |
2469 | Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel | |
2470 | Mack, Daniel Mustieles, daurnimator, Davide Bettio, David | |
2471 | Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov, | |
2472 | Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke, | |
2473 | Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López | |
2474 | Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan | |
2475 | Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John | |
2476 | Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay | |
2477 | Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas | |
2478 | De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz | |
2479 | Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel | |
2480 | Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett, | |
2481 | Michael Biebl, Michael Marineau, Michael Olbrich, Michal | |
2482 | Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik | |
2483 | Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter | |
2484 | Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny | |
2485 | Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick, | |
2486 | Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker, | |
2487 | Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas | |
2488 | Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom | |
2489 | Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will | |
2490 | Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek | |
2491 | ||
ccddd104 | 2492 | — Berlin, 2015-05-22 |
481a0aa2 | 2493 | |
615aaf41 LP |
2494 | CHANGES WITH 219: |
2495 | ||
615aaf41 LP |
2496 | * Introduce a new API "sd-hwdb.h" for querying the hardware |
2497 | metadata database. With this minimal interface one can query | |
2498 | and enumerate the udev hwdb, decoupled from the old libudev | |
2499 | library. libudev's interface for this is now only a wrapper | |
2500 | around sd-hwdb. A new tool systemd-hwdb has been added to | |
2501 | interface with and update the database. | |
2502 | ||
2503 | * When any of systemd's tools copies files (for example due to | |
2504 | tmpfiles' C lines) a btrfs reflink will attempted first, | |
2505 | before bytewise copying is done. | |
2506 | ||
2507 | * systemd-nspawn gained a new --ephemeral switch. When | |
2508 | specified a btrfs snapshot is taken of the container's root | |
2509 | directory, and immediately removed when the container | |
2510 | terminates again. Thus, a container can be started whose | |
2511 | changes never alter the container's root directory, and are | |
2512 | lost on container termination. This switch can also be used | |
2513 | for starting a container off the root file system of the | |
2514 | host without affecting the host OS. This switch is only | |
2515 | available on btrfs file systems. | |
2516 | ||
2517 | * systemd-nspawn gained a new --template= switch. It takes the | |
2518 | path to a container tree to use as template for the tree | |
7edecf21 | 2519 | specified via --directory=, should that directory be |
615aaf41 LP |
2520 | missing. This allows instantiating containers dynamically, |
2521 | on first run. This switch is only available on btrfs file | |
2522 | systems. | |
2523 | ||
2524 | * When a .mount unit refers to a mount point on which multiple | |
2525 | mounts are stacked, and the .mount unit is stopped all of | |
2526 | the stacked mount points will now be unmounted until no | |
2527 | mount point remains. | |
2528 | ||
2529 | * systemd now has an explicit notion of supported and | |
2530 | unsupported unit types. Jobs enqueued for unsupported unit | |
2531 | types will now fail with an "unsupported" error code. More | |
2532 | specifically .swap, .automount and .device units are not | |
2533 | supported in containers, .busname units are not supported on | |
2534 | non-kdbus systems. .swap and .automount are also not | |
2535 | supported if their respective kernel compile time options | |
2536 | are disabled. | |
2537 | ||
2538 | * machinectl gained support for two new "copy-from" and | |
2539 | "copy-to" commands for copying files from a running | |
2540 | container to the host or vice versa. | |
2541 | ||
2542 | * machinectl gained support for a new "bind" command to bind | |
2543 | mount host directories into local containers. This is | |
2544 | currently only supported for nspawn containers. | |
2545 | ||
2546 | * networkd gained support for configuring bridge forwarding | |
2547 | database entries (fdb) from .network files. | |
2548 | ||
2549 | * A new tiny daemon "systemd-importd" has been added that can | |
2550 | download container images in tar, raw, qcow2 or dkr formats, | |
2551 | and make them available locally in /var/lib/machines, so | |
2552 | that they can run as nspawn containers. The daemon can GPG | |
2553 | verify the downloads (not supported for dkr, since it has no | |
2554 | provisions for verifying downloads). It will transparently | |
2555 | decompress bz2, xz, gzip compressed downloads if necessary, | |
2556 | and restore sparse files on disk. The daemon uses privilege | |
2557 | separation to ensure the actual download logic runs with | |
94e5ba37 | 2558 | fewer privileges than the daemon itself. machinectl has |
615aaf41 LP |
2559 | gained new commands "pull-tar", "pull-raw" and "pull-dkr" to |
2560 | make the functionality of importd available to the | |
2561 | user. With this in place the Fedora and Ubuntu "Cloud" | |
2562 | images can be downloaded and booted as containers unmodified | |
2563 | (the Fedora images lack the appropriate GPG signature files | |
2564 | currently, so they cannot be verified, but this will change | |
2565 | soon, hopefully). Note that downloading images is currently | |
2566 | only fully supported on btrfs. | |
2567 | ||
2568 | * machinectl is now able to list container images found in | |
2569 | /var/lib/machines, along with some metadata about sizes of | |
2570 | disk and similar. If the directory is located on btrfs and | |
2571 | quota is enabled, this includes quota display. A new command | |
2572 | "image-status" has been added that shows additional | |
2573 | information about images. | |
2574 | ||
2575 | * machinectl is now able to clone container images | |
2576 | efficiently, if the underlying file system (btrfs) supports | |
f59dba26 | 2577 | it, with the new "machinectl clone" command. It also |
615aaf41 LP |
2578 | gained commands for renaming and removing images, as well as |
2579 | marking them read-only or read-write (supported also on | |
2580 | legacy file systems). | |
2581 | ||
2582 | * networkd gained support for collecting LLDP network | |
2583 | announcements, from hardware that supports this. This is | |
2584 | shown in networkctl output. | |
2585 | ||
2586 | * systemd-run gained support for a new -t (--pty) switch for | |
2587 | invoking a binary on a pty whose input and output is | |
2588 | connected to the invoking terminal. This allows executing | |
2589 | processes as system services while interactively | |
2590 | communicating with them via the terminal. Most interestingly | |
2591 | this is supported across container boundaries. Invoking | |
2592 | "systemd-run -t /bin/bash" is an alternative to running a | |
2593 | full login session, the difference being that the former | |
2594 | will not register a session, nor go through the PAM session | |
2595 | setup. | |
2596 | ||
2597 | * tmpfiles gained support for a new "v" line type for creating | |
2598 | btrfs subvolumes. If the underlying file system is a legacy | |
2599 | file system, this automatically degrades to creating a | |
2600 | normal directory. Among others /var/lib/machines is now | |
2601 | created like this at boot, should it be missing. | |
2602 | ||
2603 | * The directory /var/lib/containers/ has been deprecated and | |
2604 | been replaced by /var/lib/machines. The term "machines" has | |
2605 | been used in the systemd context as generic term for both | |
2606 | VMs and containers, and hence appears more appropriate for | |
2607 | this, as the directory can also contain raw images bootable | |
2608 | via qemu/kvm. | |
2609 | ||
2610 | * systemd-nspawn when invoked with -M but without --directory= | |
2611 | or --image= is now capable of searching for the container | |
2612 | root directory, subvolume or disk image automatically, in | |
2613 | /var/lib/machines. systemd-nspawn@.service has been updated | |
2614 | to make use of this, thus allowing it to be used for raw | |
2615 | disk images, too. | |
2616 | ||
2617 | * A new machines.target unit has been introduced that is | |
2618 | supposed to group all containers/VMs invoked as services on | |
2619 | the system. systemd-nspawn@.service has been updated to | |
2620 | integrate with that. | |
2621 | ||
2622 | * machinectl gained a new "start" command, for invoking a | |
2623 | container as a service. "machinectl start foo" is mostly | |
2624 | equivalent to "systemctl start systemd-nspawn@foo.service", | |
2625 | but handles escaping in a nicer way. | |
2626 | ||
2627 | * systemd-nspawn will now mount most of the cgroupfs tree | |
2628 | read-only into each container, with the exception of the | |
2629 | container's own subtree in the name=systemd hierarchy. | |
2630 | ||
2631 | * journald now sets the special FS_NOCOW file flag for its | |
2632 | journal files. This should improve performance on btrfs, by | |
2633 | avoiding heavy fragmentation when journald's write-pattern | |
2634 | is used on COW file systems. It degrades btrfs' data | |
2635 | integrity guarantees for the files to the same levels as for | |
2636 | ext3/ext4 however. This should be OK though as journald does | |
2637 | its own data integrity checks and all its objects are | |
2638 | checksummed on disk. Also, journald should handle btrfs disk | |
2639 | full events a lot more gracefully now, by processing SIGBUS | |
2640 | errors, and not relying on fallocate() anymore. | |
2641 | ||
2642 | * When journald detects that journal files it is writing to | |
2643 | have been deleted it will immediately start new journal | |
2644 | files. | |
2645 | ||
2646 | * systemd now provides a way to store file descriptors | |
4c37970d | 2647 | per-service in PID 1. This is useful for daemons to ensure |
615aaf41 | 2648 | that fds they require are not lost during a daemon |
94e5ba37 | 2649 | restart. The fds are passed to the daemon on the next |
615aaf41 LP |
2650 | invocation in the same way socket activation fds are |
2651 | passed. This is now used by journald to ensure that the | |
2652 | various sockets connected to all the system's stdout/stderr | |
2653 | are not lost when journald is restarted. File descriptors | |
2654 | may be stored in PID 1 via the sd_pid_notify_with_fds() API, | |
2655 | an extension to sd_notify(). Note that a limit is enforced | |
2656 | on the number of fds a service can store in PID 1, and it | |
2657 | defaults to 0, so that no fds may be stored, unless this is | |
2658 | explicitly turned on. | |
2659 | ||
2660 | * The default TERM variable to use for units connected to a | |
2661 | terminal, when no other value is explicitly is set is now | |
2662 | vt220 rather than vt102. This should be fairly safe still, | |
2663 | but allows PgUp/PgDn work. | |
2664 | ||
2665 | * The /etc/crypttab option header= as known from Debian is now | |
2666 | supported. | |
2667 | ||
2668 | * "loginctl user-status" and "loginctl session-status" will | |
2669 | now show the last 10 lines of log messages of the | |
2670 | user/session following the status output. Similar, | |
2671 | "machinectl status" will show the last 10 log lines | |
2672 | associated with a virtual machine or container | |
2673 | service. (Note that this is usually not the log messages | |
2674 | done in the VM/container itself, but simply what the | |
2675 | container manager logs. For nspawn this includes all console | |
2676 | output however.) | |
2677 | ||
2678 | * "loginctl session-status" without further argument will now | |
2679 | show the status of the session of the caller. Similar, | |
2680 | "lock-session", "unlock-session", "activate", | |
2681 | "enable-linger", "disable-linger" may now be called without | |
2682 | session/user parameter in which case they apply to the | |
2683 | caller's session/user. | |
2684 | ||
2685 | * An X11 session scriptlet is now shipped that uploads | |
2686 | $DISPLAY and $XAUTHORITY into the environment of the systemd | |
2687 | --user daemon if a session begins. This should improve | |
2688 | compatibility with X11 enabled applications run as systemd | |
2689 | user services. | |
2690 | ||
2691 | * Generators are now subject to masking via /etc and /run, the | |
2692 | same way as unit files. | |
2693 | ||
2694 | * networkd .network files gained support for configuring | |
2695 | per-link IPv4/IPv6 packet forwarding as well as IPv4 | |
2696 | masquerading. This is by default turned on for veth links to | |
2697 | containers, as registered by systemd-nspawn. This means that | |
2698 | nspawn containers run with --network-veth will now get | |
2699 | automatic routed access to the host's networks without any | |
2700 | further configuration or setup, as long as networkd runs on | |
2701 | the host. | |
2702 | ||
2703 | * systemd-nspawn gained the --port= (-p) switch to expose TCP | |
2704 | or UDP posts of a container on the host. With this in place | |
2705 | it is possible to run containers with private veth links | |
2706 | (--network-veth), and have their functionality exposed on | |
2707 | the host as if their services were running directly on the | |
2708 | host. | |
2709 | ||
dd2fd155 | 2710 | * systemd-nspawn's --network-veth switch now gained a short |
615aaf41 LP |
2711 | version "-n", since with the changes above it is now truly |
2712 | useful out-of-the-box. The systemd-nspawn@.service has been | |
2713 | updated to make use of it too by default. | |
2714 | ||
2715 | * systemd-nspawn will now maintain a per-image R/W lock, to | |
2716 | ensure that the same image is not started more than once | |
2717 | writable. (It's OK to run an image multiple times | |
2718 | simultaneously in read-only mode.) | |
2719 | ||
2720 | * systemd-nspawn's --image= option is now capable of | |
2721 | dissecting and booting MBR and GPT disk images that contain | |
2722 | only a single active Linux partition. Previously it | |
2723 | supported only GPT disk images with proper GPT type | |
2724 | IDs. This allows running cloud images from major | |
2725 | distributions directly with systemd-nspawn, without | |
2726 | modification. | |
2727 | ||
2728 | * In addition to collecting mouse dpi data in the udev | |
2729 | hardware database, there's now support for collecting angle | |
2730 | information for mouse scroll wheels. The database is | |
7edecf21 | 2731 | supposed to guarantee similar scrolling behavior on mice |
615aaf41 LP |
2732 | that it knows about. There's also support for collecting |
2733 | information about Touchpad types. | |
2734 | ||
2735 | * udev's input_id built-in will now also collect touch screen | |
2736 | dimension data and attach it to probed devices. | |
2737 | ||
2738 | * /etc/os-release gained support for a Distribution Privacy | |
2739 | Policy link field. | |
2740 | ||
2741 | * networkd gained support for creating "ipvlan", "gretap", | |
2742 | "ip6gre", "ip6gretap" and "ip6tnl" network devices. | |
2743 | ||
2744 | * systemd-tmpfiles gained support for "a" lines for setting | |
2745 | ACLs on files. | |
2746 | ||
2747 | * systemd-nspawn will now mount /tmp in the container to | |
2748 | tmpfs, automatically. | |
2749 | ||
2750 | * systemd now exposes the memory.usage_in_bytes cgroup | |
2751 | attribute and shows it for each service in the "systemctl | |
2752 | status" output, if available. | |
2753 | ||
2754 | * When the user presses Ctrl-Alt-Del more than 7x within 2s an | |
2755 | immediate reboot is triggered. This useful if shutdown is | |
2756 | hung and is unable to complete, to expedite the | |
2757 | operation. Note that this kind of reboot will still unmount | |
2758 | all file systems, and hence should not result in fsck being | |
2759 | run on next reboot. | |
2760 | ||
2761 | * A .device unit for an optical block device will now be | |
2762 | considered active only when a medium is in the drive. Also, | |
2763 | mount units are now bound to their backing devices thus | |
2764 | triggering automatic unmounting when devices become | |
2765 | unavailable. With this in place systemd will now | |
2766 | automatically unmount left-over mounts when a CD-ROM is | |
2767 | ejected or an USB stick is yanked from the system. | |
2768 | ||
2769 | * networkd-wait-online now has support for waiting for | |
2770 | specific interfaces only (with globbing), and for giving up | |
2771 | after a configurable timeout. | |
2772 | ||
2773 | * networkd now exits when idle. It will be automatically | |
2774 | restarted as soon as interfaces show up, are removed or | |
2775 | change state. networkd will stay around as long as there is | |
2776 | at least one DHCP state machine or similar around, that keep | |
2777 | it non-idle. | |
2778 | ||
2779 | * networkd may now configure IPv6 link-local addressing in | |
2780 | addition to IPv4 link-local addressing. | |
2781 | ||
2782 | * The IPv6 "token" for use in SLAAC may now be configured for | |
2783 | each .network interface in networkd. | |
2784 | ||
2785 | * Routes configured with networkd may now be assigned a scope | |
2786 | in .network files. | |
2787 | ||
2788 | * networkd's [Match] sections now support globbing and lists | |
2789 | of multiple space-separated matches per item. | |
2790 | ||
11ea2781 | 2791 | Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser, |
d2c643c6 LP |
2792 | Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos |
2793 | Morata Castillo, Chris Atkinson, Chris J. Arges, Christian | |
2794 | Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie, | |
2795 | Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack, | |
2796 | Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald, | |
2797 | Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de | |
2798 | Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan | |
2799 | Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas | |
2800 | Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken | |
2801 | Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian, | |
2802 | Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas, | |
2803 | Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko | |
2804 | Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl, | |
2805 | Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas | |
11ea2781 LP |
2806 | Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul |
2807 | Martin, Peter Hutterer, Peter Mattern, Philippe De Swert, | |
2808 | Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny | |
2809 | Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick, | |
2810 | Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain | |
2811 | Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom | |
2812 | Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar | |
d2c643c6 LP |
2813 | Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland |
2814 | Hoffmann, Zbigniew Jędrzejewski-Szmek | |
11ea2781 | 2815 | |
ccddd104 | 2816 | — Berlin, 2015-02-16 |
11ea2781 | 2817 | |
d4f5a1f4 DH |
2818 | CHANGES WITH 218: |
2819 | ||
f9e00a9f LP |
2820 | * When querying unit file enablement status (for example via |
2821 | "systemctl is-enabled"), a new state "indirect" is now known | |
2822 | which indicates that a unit might not be enabled itself, but | |
c7683ffb | 2823 | another unit listed in its Also= setting might be. |
f9e00a9f LP |
2824 | |
2825 | * Similar to the various existing ConditionXYZ= settings for | |
b938cb90 | 2826 | units, there are now matching AssertXYZ= settings. While |
f9e00a9f LP |
2827 | failing conditions cause a unit to be skipped, but its job |
2828 | to succeed, failing assertions declared like this will cause | |
2829 | a unit start operation and its job to fail. | |
2830 | ||
2831 | * hostnamed now knows a new chassis type "embedded". | |
2832 | ||
2833 | * systemctl gained a new "edit" command. When used on a unit | |
b938cb90 | 2834 | file, this allows extending unit files with .d/ drop-in |
f9e00a9f LP |
2835 | configuration snippets or editing the full file (after |
2836 | copying it from /usr/lib to /etc). This will invoke the | |
2837 | user's editor (as configured with $EDITOR), and reload the | |
2838 | modified configuration after editing. | |
2839 | ||
2840 | * "systemctl status" now shows the suggested enablement state | |
2841 | for a unit, as declared in the (usually vendor-supplied) | |
2842 | system preset files. | |
2843 | ||
2844 | * nss-myhostname will now resolve the single-label host name | |
2845 | "gateway" to the locally configured default IP routing | |
2846 | gateways, ordered by their metrics. This assigns a stable | |
2847 | name to the used gateways, regardless which ones are | |
2848 | currently configured. Note that the name will only be | |
2849 | resolved after all other name sources (if nss-myhostname is | |
2850 | configured properly) and should hence not negatively impact | |
2851 | systems that use the single-label host name "gateway" in | |
2852 | other contexts. | |
2853 | ||
2854 | * systemd-inhibit now allows filtering by mode when listing | |
2855 | inhibitors. | |
2856 | ||
122676c9 | 2857 | * Scope and service units gained a new "Delegate" boolean |
b938cb90 | 2858 | property, which, when set, allows processes running inside the |
122676c9 LP |
2859 | unit to further partition resources. This is primarily |
2860 | useful for systemd user instances as well as container | |
2861 | managers. | |
f9e00a9f LP |
2862 | |
2863 | * journald will now pick up audit messages directly from | |
2864 | the kernel, and log them like any other log message. The | |
2865 | audit fields are split up and fully indexed. This means that | |
2866 | journalctl in many ways is now a (nicer!) alternative to | |
2867 | ausearch, the traditional audit client. Note that this | |
b938cb90 | 2868 | implements only a minimal audit client. If you want the |
f9e00a9f LP |
2869 | special audit modes like reboot-on-log-overflow, please use |
2870 | the traditional auditd instead, which can be used in | |
2871 | parallel to journald. | |
2872 | ||
2873 | * The ConditionSecurity= unit file option now understands the | |
2874 | special string "audit" to check whether auditing is | |
2875 | available. | |
2876 | ||
2877 | * journalctl gained two new commands --vacuum-size= and | |
2878 | --vacuum-time= to delete old journal files until the | |
a8eaaee7 | 2879 | remaining ones take up no more than the specified size on disk, |
f9e00a9f LP |
2880 | or are not older than the specified time. |
2881 | ||
2882 | * A new, native PPPoE library has been added to sd-network, | |
2883 | systemd's library of light-weight networking protocols. This | |
2884 | library will be used in a future version of networkd to | |
2885 | enable PPPoE communication without an external pppd daemon. | |
2886 | ||
2887 | * The busctl tool now understands a new "capture" verb that | |
2888 | works similar to "monitor", but writes a packet capture | |
2889 | trace to STDOUT that can be redirected to a file which is | |
2890 | compatible with libcap's capture file format. This can then | |
2891 | be loaded in Wireshark and similar tools to inspect bus | |
2892 | communication. | |
2893 | ||
2894 | * The busctl tool now understands a new "tree" verb that shows | |
2895 | the object trees of a specific service on the bus, or of all | |
2896 | services. | |
2897 | ||
2898 | * The busctl tool now understands a new "introspect" verb that | |
2899 | shows all interfaces and members of objects on the bus, | |
2900 | including their signature and values. This is particularly | |
2901 | useful to get more information about bus objects shown by | |
2902 | the new "busctl tree" command. | |
2903 | ||
2904 | * The busctl tool now understands new verbs "call", | |
2905 | "set-property" and "get-property" for invoking bus method | |
2906 | calls, setting and getting bus object properties in a | |
2907 | friendly way. | |
2908 | ||
2909 | * busctl gained a new --augment-creds= argument that controls | |
2910 | whether the tool shall augment credential information it | |
2911 | gets from the bus with data from /proc, in a possibly | |
2912 | race-ful way. | |
2913 | ||
2914 | * nspawn's --link-journal= switch gained two new values | |
2915 | "try-guest" and "try-host" that work like "guest" and | |
17c29493 | 2916 | "host", but do not fail if the host has no persistent |
f9e00a9f LP |
2917 | journalling enabled. -j is now equivalent to |
2918 | --link-journal=try-guest. | |
2919 | ||
2920 | * macvlan network devices created by nspawn will now have | |
2921 | stable MAC addresses. | |
2922 | ||
2923 | * A new SmackProcessLabel= unit setting has been added, which | |
2924 | controls the SMACK security label processes forked off by | |
2925 | the respective unit shall use. | |
2926 | ||
d4f5a1f4 DH |
2927 | * If compiled with --enable-xkbcommon, systemd-localed will |
2928 | verify x11 keymap settings by compiling the given keymap. It | |
2929 | will spew out warnings if the compilation fails. This | |
2930 | requires libxkbcommon to be installed. | |
2931 | ||
b938cb90 | 2932 | * When a coredump is collected, a larger number of metadata |
f9e00a9f | 2933 | fields is now collected and included in the journal records |
b938cb90 | 2934 | created for it. More specifically, control group membership, |
f9e00a9f LP |
2935 | environment variables, memory maps, working directory, |
2936 | chroot directory, /proc/$PID/status, and a list of open file | |
2937 | descriptors is now stored in the log entry. | |
2938 | ||
17c29493 | 2939 | * The udev hwdb now contains DPI information for mice. For |
f9e00a9f LP |
2940 | details see: |
2941 | ||
2942 | http://who-t.blogspot.de/2014/12/building-a-dpi-database-for-mice.html | |
2943 | ||
2944 | * All systemd programs that read standalone configuration | |
2945 | files in /etc now also support a corresponding series of | |
997b2b43 JT |
2946 | .conf.d configuration directories in /etc/, /run/, |
2947 | /usr/local/lib/, /usr/lib/, and (if configured with | |
2948 | --enable-split-usr) /lib/. In particular, the following | |
2949 | configuration files now have corresponding configuration | |
2950 | directories: system.conf user.conf, logind.conf, | |
2951 | journald.conf, sleep.conf, bootchart.conf, coredump.conf, | |
2952 | resolved.conf, timesyncd.conf, journal-remote.conf, and | |
2953 | journal-upload.conf. Note that distributions should use the | |
2954 | configuration directories in /usr/lib/; the directories in | |
2955 | /etc/ are reserved for the system administrator. | |
2956 | ||
f9e00a9f LP |
2957 | * systemd-rfkill will no longer take the rfkill device name |
2958 | into account when storing rfkill state on disk, as the name | |
2959 | might be dynamically assigned and not stable. Instead, the | |
2960 | ID_PATH udev variable combined with the rfkill type (wlan, | |
2961 | bluetooth, ...) is used. | |
2962 | ||
2963 | * A new service systemd-machine-id-commit.service has been | |
2964 | added. When used on systems where /etc is read-only during | |
2965 | boot, and /etc/machine-id is not initialized (but an empty | |
2966 | file), this service will copy the temporary machine ID | |
2967 | created as replacement into /etc after the system is fully | |
2968 | booted up. This is useful for systems that are freshly | |
2969 | installed with a non-initialized machine ID, but should get | |
2970 | a fixed machine ID for subsequent boots. | |
2971 | ||
2972 | * networkd's .netdev files now provide a large set of | |
a8eaaee7 | 2973 | configuration parameters for VXLAN devices. Similarly, the |
f9e00a9f LP |
2974 | bridge port cost parameter is now configurable in .network |
2975 | files. There's also new support for configuring IP source | |
2976 | routing. networkd .link files gained support for a new | |
2977 | OriginalName= match that is useful to match against the | |
2978 | original interface name the kernel assigned. .network files | |
2979 | may include MTU= and MACAddress= fields for altering the MTU | |
2980 | and MAC address while being connected to a specific network | |
2981 | interface. | |
2982 | ||
2983 | * The LUKS logic gained supported for configuring | |
2984 | UUID-specific key files. There's also new support for naming | |
2985 | LUKS device from the kernel command line, using the new | |
2986 | luks.name= argument. | |
2987 | ||
2988 | * Timer units may now be transiently created via the bus API | |
2989 | (this was previously already available for scope and service | |
2990 | units). In addition it is now possible to create multiple | |
2991 | transient units at the same time with a single bus call. The | |
2992 | "systemd-run" tool has been updated to make use of this for | |
2993 | running commands on a specified time, in at(1)-style. | |
2994 | ||
2995 | * tmpfiles gained support for "t" lines, for assigning | |
2996 | extended attributes to files. Among other uses this may be | |
2997 | used to assign SMACK labels to files. | |
2998 | ||
13e92f39 LP |
2999 | Contributions from: Alin Rauta, Alison Chaiken, Andrej |
3000 | Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris | |
3001 | Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez, | |
3002 | Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave | |
3003 | Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin | |
3004 | Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan | |
3005 | Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe | |
3006 | Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering, | |
3007 | Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas | |
3008 | Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi, | |
3009 | Michael Biebl, Michael Chapman, Michael Marineau, Michal | |
7da81d33 LP |
3010 | Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter |
3011 | Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode, | |
3012 | Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross | |
3013 | Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani, | |
3014 | Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen, | |
3015 | Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert | |
3016 | Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek | |
13e92f39 | 3017 | |
ccddd104 | 3018 | — Berlin, 2014-12-10 |
f9e00a9f | 3019 | |
b62a309a ZJS |
3020 | CHANGES WITH 217: |
3021 | ||
78b6b7ce LP |
3022 | * journalctl gained the new options -t/--identifier= to match |
3023 | on the syslog identifier (aka "tag"), as well as --utc to | |
3024 | show log timestamps in the UTC timezone. journalctl now also | |
3025 | accepts -n/--lines=all to disable line capping in a pager. | |
b62a309a | 3026 | |
a65b8245 ZJS |
3027 | * journalctl gained a new switch, --flush, that synchronously |
3028 | flushes logs from /run/log/journal to /var/log/journal if | |
3029 | persistent storage is enabled. systemd-journal-flush.service | |
3030 | now waits until the operation is complete. | |
2a97b03b | 3031 | |
b62a309a ZJS |
3032 | * Services can notify the manager before they start a reload |
3033 | (by sending RELOADING=1) or shutdown (by sending | |
4bdc60cb LP |
3034 | STOPPING=1). This allows the manager to track and show the |
3035 | internal state of daemons and closes a race condition when | |
78b6b7ce | 3036 | the process is still running but has closed its D-Bus |
4bdc60cb | 3037 | connection. |
b62a309a | 3038 | |
78b6b7ce LP |
3039 | * Services with Type=oneshot do not have to have any ExecStart |
3040 | commands anymore. | |
b62a309a ZJS |
3041 | |
3042 | * User units are now loaded also from | |
3043 | $XDG_RUNTIME_DIR/systemd/user/. This is similar to the | |
3044 | /run/systemd/user directory that was already previously | |
3045 | supported, but is under the control of the user. | |
3046 | ||
4ffd29fd LP |
3047 | * Job timeouts (i.e. time-outs on the time a job that is |
3048 | queued stays in the run queue) can now optionally result in | |
3049 | immediate reboot or power-off actions (JobTimeoutAction= and | |
3050 | JobTimeoutRebootArgument=). This is useful on ".target" | |
3051 | units, to limit the maximum time a target remains | |
3052 | undispatched in the run queue, and to trigger an emergency | |
3053 | operation in such a case. This is now used by default to | |
3054 | turn off the system if boot-up (as defined by everything in | |
3055 | basic.target) hangs and does not complete for at least | |
3056 | 15min. Also, if power-off or reboot hang for at least 30min | |
3057 | an immediate power-off/reboot operation is triggered. This | |
3058 | functionality is particularly useful to increase reliability | |
3059 | on embedded devices, but also on laptops which might | |
3060 | accidentally get powered on when carried in a backpack and | |
3061 | whose boot stays stuck in a hard disk encryption passphrase | |
3062 | question. | |
3063 | ||
b62a309a ZJS |
3064 | * systemd-logind can be configured to also handle lid switch |
3065 | events even when the machine is docked or multiple displays | |
3066 | are attached (HandleLidSwitchDocked= option). | |
3067 | ||
3068 | * A helper binary and a service have been added which can be | |
3069 | used to resume from hibernation in the initramfs. A | |
3070 | generator will parse the resume= option on the kernel | |
81c7dd89 | 3071 | command line to trigger resume. |
b62a309a | 3072 | |
78b6b7ce LP |
3073 | * A user console daemon systemd-consoled has been |
3074 | added. Currently, it is a preview, and will so far open a | |
3075 | single terminal on each session of the user marked as | |
09077149 | 3076 | Desktop=systemd-console. |
b62a309a ZJS |
3077 | |
3078 | * Route metrics can be specified for DHCP routes added by | |
3079 | systemd-networkd. | |
3080 | ||
ba8df74b | 3081 | * The SELinux context of socket-activated services can be set |
78b6b7ce | 3082 | from the information provided by the networking stack |
b62a309a ZJS |
3083 | (SELinuxContextFromNet= option). |
3084 | ||
3085 | * Userspace firmware loading support has been removed and | |
3086 | the minimum supported kernel version is thus bumped to 3.7. | |
3087 | ||
3088 | * Timeout for udev workers has been increased from 1 to 3 | |
3089 | minutes, but a warning will be printed after 1 minute to | |
3090 | help diagnose kernel modules that take a long time to load. | |
3091 | ||
78b6b7ce | 3092 | * Udev rules can now remove tags on devices with TAG-="foobar". |
b62a309a | 3093 | |
4bdc60cb | 3094 | * systemd's readahead implementation has been removed. In many |
f6d1de85 | 3095 | circumstances it didn't give expected benefits even for |
b62a309a | 3096 | rotational disk drives and was becoming less relevant in the |
78b6b7ce LP |
3097 | age of SSDs. As none of the developers has been using |
3098 | rotating media anymore, and nobody stepped up to actively | |
3099 | maintain this component of systemd it has now been removed. | |
b62a309a | 3100 | |
c4ac9900 | 3101 | * Swap units can use Options= to specify discard options. |
b62a309a ZJS |
3102 | Discard options specified for swaps in /etc/fstab are now |
3103 | respected. | |
3104 | ||
3105 | * Docker containers are now detected as a separate type of | |
3106 | virtualization. | |
3107 | ||
3108 | * The Password Agent protocol gained support for queries where | |
ba8df74b | 3109 | the user input is shown, useful e.g. for user names. |
78b6b7ce LP |
3110 | systemd-ask-password gained a new --echo option to turn that |
3111 | on. | |
b62a309a | 3112 | |
e6c253e3 MS |
3113 | * The default sysctl.d/ snippets will now set: |
3114 | ||
3115 | net.core.default_qdisc = fq_codel | |
3116 | ||
ba8df74b KS |
3117 | This selects Fair Queuing Controlled Delay as the default |
3118 | queuing discipline for network interfaces. fq_codel helps | |
e6c253e3 MS |
3119 | fight the network bufferbloat problem. It is believed to be |
3120 | a good default with no tuning required for most workloads. | |
3121 | Downstream distributions may override this choice. On 10Gbit | |
3122 | servers that do not do forwarding, "fq" may perform better. | |
3123 | Systems without a good clocksource should use "pfifo_fast". | |
3124 | ||
4bdc60cb LP |
3125 | * If kdbus is enabled during build a new option BusPolicy= is |
3126 | available for service units, that allows locking all service | |
3127 | processes into a stricter bus policy, in order to limit | |
3128 | access to various bus services, or even hide most of them | |
3129 | from the service's view entirely. | |
3130 | ||
3131 | * networkctl will now show the .network and .link file | |
3132 | networkd has applied to a specific interface. | |
3133 | ||
3134 | * sd-login gained a new API call sd_session_get_desktop() to | |
3135 | query which desktop environment has been selected for a | |
3136 | session. | |
3137 | ||
3138 | * UNIX utmp support is now compile-time optional to support | |
3139 | legacy-free systems. | |
3140 | ||
78b6b7ce LP |
3141 | * systemctl gained two new commands "add-wants" and |
3142 | "add-requires" for pulling in units from specific targets | |
3143 | easily. | |
3144 | ||
3145 | * If the word "rescue" is specified on the kernel command line | |
3146 | the system will now boot into rescue mode (aka | |
3147 | rescue.target), which was previously available only by | |
3148 | specifying "1" or "systemd.unit=rescue.target" on the kernel | |
3149 | command line. This new kernel command line option nicely | |
3150 | mirrors the already existing "emergency" kernel command line | |
3151 | option. | |
3152 | ||
3153 | * New kernel command line options mount.usr=, mount.usrflags=, | |
d4474c41 | 3154 | mount.usrfstype= have been added that match root=, rootflags=, |
78b6b7ce LP |
3155 | rootfstype= but allow mounting a specific file system to |
3156 | /usr. | |
3157 | ||
f6d1de85 | 3158 | * The $NOTIFY_SOCKET is now also passed to control processes of |
78b6b7ce LP |
3159 | services, not only the main process. |
3160 | ||
3161 | * This version reenables support for fsck's -l switch. This | |
3162 | means at least version v2.25 of util-linux is required for | |
3163 | operation, otherwise dead-locks on device nodes may | |
3164 | occur. Again: you need to update util-linux to at least | |
3165 | v2.25 when updating systemd to v217. | |
3166 | ||
3769415e TT |
3167 | * The "multi-seat-x" tool has been removed from systemd, as |
3168 | its functionality has been integrated into X servers 1.16, | |
3169 | and the tool is hence redundant. It is recommended to update | |
3170 | display managers invoking this tool to simply invoke X | |
3171 | directly from now on, again. | |
3172 | ||
fae9332b LP |
3173 | * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus |
3174 | message flag has been added for all of systemd's PolicyKit | |
3175 | authenticated method calls has been added. In particular | |
3176 | this now allows optional interactive authorization via | |
ba8df74b | 3177 | PolicyKit for many of PID1's privileged operations such as |
fae9332b LP |
3178 | unit file enabling and disabling. |
3179 | ||
cfa1571b LP |
3180 | * "udevadm hwdb --update" learnt a new switch "--usr" for |
3181 | placing the rebuilt hardware database in /usr instead of | |
3182 | /etc. When used only hardware database entries stored in | |
3183 | /usr will be used, and any user database entries in /etc are | |
3184 | ignored. This functionality is useful for vendors to ship a | |
3185 | pre-built database on systems where local configuration is | |
3186 | unnecessary or unlikely. | |
3187 | ||
7e63dd10 LP |
3188 | * Calendar time specifications in .timer units now also |
3189 | understand the strings "semi-annually", "quarterly" and | |
ba8df74b | 3190 | "minutely" as shortcuts (in addition to the preexisting |
7e63dd10 LP |
3191 | "anually", "hourly", ...). |
3192 | ||
d4474c41 TG |
3193 | * systemd-tmpfiles will now correctly create files in /dev |
3194 | at boot which are marked for creation only at boot. It is | |
3195 | recommended to always create static device nodes with 'c!' | |
3196 | and 'b!', so that they are created only at boot and not | |
3197 | overwritten at runtime. | |
3198 | ||
3b187c5c LP |
3199 | * When the watchdog logic is used for a service (WatchdogSec=) |
3200 | and the watchdog timeout is hit the service will now be | |
3201 | terminated with SIGABRT (instead of just SIGTERM), in order | |
3202 | to make sure a proper coredump and backtrace is | |
3203 | generated. This ensures that hanging services will result in | |
3204 | similar coredump/backtrace behaviour as services that hit a | |
3205 | segmentation fault. | |
3206 | ||
4b08dd87 LP |
3207 | Contributions from: Andreas Henriksson, Andrei Borzenkov, |
3208 | Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L. | |
3209 | Black, Christian Hesse, Cristian Rodríguez, Daniel Buch, | |
3210 | Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David | |
3211 | Herrmann, David Sommerseth, David Strauss, Emil Renner | |
3212 | Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger, | |
3213 | Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo | |
3214 | Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan | |
3215 | Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus | |
3216 | Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz | |
3217 | Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann, | |
3218 | Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl, | |
3219 | Michael Marineau, Michael Olbrich, Michael Scherer, Michal | |
3220 | Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt, | |
3221 | Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard | |
3222 | Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof, | |
3223 | Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd | |
3224 | Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant | |
3225 | Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen, | |
3226 | Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein | |
3227 | Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew | |
13e92f39 | 3228 | Jędrzejewski-Szmek |
4b08dd87 | 3229 | |
ccddd104 | 3230 | — Berlin, 2014-10-28 |
4b08dd87 | 3231 | |
b72ddf0f | 3232 | CHANGES WITH 216: |
b2ca0d63 LP |
3233 | |
3234 | * timedated no longer reads NTP implementation unit names from | |
b72ddf0f | 3235 | /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP |
b2ca0d63 LP |
3236 | implementations should add a |
3237 | ||
b72ddf0f | 3238 | Conflicts=systemd-timesyncd.service |
b2ca0d63 LP |
3239 | |
3240 | to their unit files to take over and replace systemd's NTP | |
3241 | default functionality. | |
3242 | ||
3243 | * systemd-sysusers gained a new line type "r" for configuring | |
3244 | which UID/GID ranges to allocate system users/groups | |
3245 | from. Lines of type "u" may now add an additional column | |
3246 | that specifies the home directory for the system user to be | |
3247 | created. Also, systemd-sysusers may now optionally read user | |
3248 | information from STDIN instead of a file. This is useful for | |
3249 | invoking it from RPM preinst scriptlets that need to create | |
3250 | users before the first RPM file is installed since these | |
3251 | files might need to be owned by them. A new | |
3252 | %sysusers_create_inline RPM macro has been introduced to do | |
3253 | just that. systemd-sysusers now updates the shadow files as | |
3254 | well as the user/group databases, which should enhance | |
3255 | compatibility with certain tools like grpck. | |
3256 | ||
3257 | * A number of bus APIs of PID 1 now optionally consult | |
5f02e26c | 3258 | PolicyKit to permit access for otherwise unprivileged |
b2ca0d63 LP |
3259 | clients under certain conditions. Note that this currently |
3260 | doesn't support interactive authentication yet, but this is | |
3261 | expected to be added eventually, too. | |
3262 | ||
3263 | * /etc/machine-info now has new fields for configuring the | |
3264 | deployment environment of the machine, as well as the | |
3265 | location of the machine. hostnamectl has been updated with | |
3266 | new command to update these fields. | |
3267 | ||
3268 | * systemd-timesyncd has been updated to automatically acquire | |
3269 | NTP server information from systemd-networkd, which might | |
3270 | have been discovered via DHCP. | |
3271 | ||
3272 | * systemd-resolved now includes a caching DNS stub resolver | |
3273 | and a complete LLMNR name resolution implementation. A new | |
daa05349 AB |
3274 | NSS module "nss-resolve" has been added which can be used |
3275 | instead of glibc's own "nss-dns" to resolve hostnames via | |
b2ca0d63 LP |
3276 | systemd-resolved. Hostnames, addresses and arbitrary RRs may |
3277 | be resolved via systemd-resolved D-Bus APIs. In contrast to | |
3278 | the glibc internal resolver systemd-resolved is aware of | |
3279 | multi-homed system, and keeps DNS server and caches separate | |
5f02e26c | 3280 | and per-interface. Queries are sent simultaneously on all |
b2ca0d63 LP |
3281 | interfaces that have DNS servers configured, in order to |
3282 | properly handle VPNs and local LANs which might resolve | |
3283 | separate sets of domain names. systemd-resolved may acquire | |
a1a4a25e | 3284 | DNS server information from systemd-networkd automatically, |
b2ca0d63 LP |
3285 | which in turn might have discovered them via DHCP. A tool |
3286 | "systemd-resolve-host" has been added that may be used to | |
3287 | query the DNS logic in resolved. systemd-resolved implements | |
3288 | IDNA and automatically uses IDNA or UTF-8 encoding depending | |
3289 | on whether classic DNS or LLMNR is used as transport. In the | |
3290 | next releases we intend to add a DNSSEC and mDNS/DNS-SD | |
3291 | implementation to systemd-resolved. | |
3292 | ||
3293 | * A new NSS module nss-mymachines has been added, that | |
3294 | automatically resolves the names of all local registered | |
3295 | containers to their respective IP addresses. | |
3296 | ||
3297 | * A new client tool "networkctl" for systemd-networkd has been | |
3298 | added. It currently is entirely passive and will query | |
3299 | networking configuration from udev, rtnetlink and networkd, | |
5f02e26c | 3300 | and present it to the user in a very friendly |
b2ca0d63 LP |
3301 | way. Eventually, we hope to extend it to become a full |
3302 | control utility for networkd. | |
3303 | ||
3304 | * .socket units gained a new DeferAcceptSec= setting that | |
3305 | controls the kernels' TCP_DEFER_ACCEPT sockopt for | |
a8eaaee7 | 3306 | TCP. Similarly, support for controlling TCP keep-alive |
b2ca0d63 LP |
3307 | settings has been added (KeepAliveTimeSec=, |
3308 | KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for | |
3309 | turning off Nagle's algorithm on TCP has been added | |
3310 | (NoDelay=). | |
3311 | ||
a1a4a25e | 3312 | * logind learned a new session type "web", for use in projects |
b2ca0d63 LP |
3313 | like Cockpit which register web clients as PAM sessions. |
3314 | ||
3315 | * timer units with at least one OnCalendar= setting will now | |
3316 | be started only after timer-sync.target has been | |
3317 | reached. This way they will not elapse before the system | |
3318 | clock has been corrected by a local NTP client or | |
3319 | similar. This is particular useful on RTC-less embedded | |
3320 | machines, that come up with an invalid system clock. | |
3321 | ||
3322 | * systemd-nspawn's --network-veth= switch should now result in | |
3323 | stable MAC addresses for both the outer and the inner side | |
3324 | of the link. | |
3325 | ||
3326 | * systemd-nspawn gained a new --volatile= switch for running | |
3327 | container instances with /etc or /var unpopulated. | |
3328 | ||
3329 | * The kdbus client code has been updated to use the new Linux | |
3330 | 3.17 memfd subsystem instead of the old kdbus-specific one. | |
3331 | ||
3332 | * systemd-networkd's DHCP client and server now support | |
01da80b1 LP |
3333 | FORCERENEW. There are also new configuration options to |
3334 | configure the vendor client identifier and broadcast mode | |
3335 | for DHCP. | |
b2ca0d63 LP |
3336 | |
3337 | * systemd will no longer inform the kernel about the current | |
3338 | timezone, as this is necessarily incorrect and racy as the | |
3339 | kernel has no understanding of DST and similar | |
3340 | concepts. This hence means FAT timestamps will be always | |
3341 | considered UTC, similar to what Android is already | |
3342 | doing. Also, when the RTC is configured to the local time | |
3343 | (rather than UTC) systemd will never synchronize back to it, | |
3344 | as this might confuse Windows at a later boot. | |
3345 | ||
3346 | * systemd-analyze gained a new command "verify" for offline | |
3347 | validation of unit files. | |
3348 | ||
3349 | * systemd-networkd gained support for a couple of additional | |
3350 | settings for bonding networking setups. Also, the metric for | |
3351 | statically configured routes may now be configured. For | |
3352 | network interfaces where this is appropriate the peer IP | |
3353 | address may now be configured. | |
3354 | ||
26568403 TG |
3355 | * systemd-networkd's DHCP client will no longer request |
3356 | broadcasting by default, as this tripped up some networks. | |
3357 | For hardware where broadcast is required the feature should | |
3358 | be switched back on using RequestBroadcast=yes. | |
3359 | ||
3360 | * systemd-networkd will now set up IPv4LL addresses (when | |
3361 | enabled) even if DHCP is configured successfully. | |
3362 | ||
3363 | * udev will now default to respect network device names given | |
3364 | by the kernel when the kernel indicates that these are | |
3365 | predictable. This behavior can be tweaked by changing | |
3366 | NamePolicy= in the relevant .link file. | |
3367 | ||
b2ca0d63 LP |
3368 | * A new library systemd-terminal has been added that |
3369 | implements full TTY stream parsing and rendering. This | |
3370 | library is supposed to be used later on for implementing a | |
3371 | full userspace VT subsystem, replacing the current kernel | |
3372 | implementation. | |
3373 | ||
3374 | * A new tool systemd-journal-upload has been added to push | |
3375 | journal data to a remote system running | |
3376 | systemd-journal-remote. | |
3377 | ||
3378 | * journald will no longer forward all local data to another | |
3379 | running syslog daemon. This change has been made because | |
3380 | rsyslog (which appears to be the most commonly used syslog | |
3381 | implementation these days) no longer makes use of this, and | |
3382 | instead pulls the data out of the journal on its own. Since | |
5f02e26c | 3383 | forwarding the messages to a non-existent syslog server is |
b2ca0d63 LP |
3384 | more expensive than we assumed we have now turned this |
3385 | off. If you run a syslog server that is not a recent rsyslog | |
3386 | version, you have to turn this option on again | |
3387 | (ForwardToSyslog= in journald.conf). | |
3388 | ||
3389 | * journald now optionally supports the LZ4 compressor for | |
3390 | larger journal fields. This compressor should perform much | |
3391 | better than XZ which was the previous default. | |
3392 | ||
3393 | * machinectl now shows the IP addresses of local containers, | |
3394 | if it knows them, plus the interface name of the container. | |
3395 | ||
3396 | * A new tool "systemd-escape" has been added that makes it | |
3397 | easy to escape strings to build unit names and similar. | |
3398 | ||
3399 | * sd_notify() messages may now include a new ERRNO= field | |
3400 | which is parsed and collected by systemd and shown among the | |
3401 | "systemctl status" output for a service. | |
3402 | ||
3403 | * A new component "systemd-firstboot" has been added that | |
3404 | queries the most basic systemd information (timezone, | |
a1a4a25e | 3405 | hostname, root password) interactively on first |
b2ca0d63 LP |
3406 | boot. Alternatively it may also be used to provision these |
3407 | things offline on OS images installed into directories. | |
3408 | ||
01da80b1 LP |
3409 | * The default sysctl.d/ snippets will now set |
3410 | ||
3411 | net.ipv4.conf.default.promote_secondaries=1 | |
3412 | ||
3413 | This has the benefit of no flushing secondary IP addresses | |
3414 | when primary addresses are removed. | |
3415 | ||
b2ca0d63 LP |
3416 | Contributions from: Ansgar Burchardt, Bastien Nocera, Colin |
3417 | Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel | |
3418 | Mack, Dan Williams, Dave Reisner, David Herrmann, Denis | |
3419 | Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald | |
3420 | Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann | |
3421 | B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin | |
3422 | Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, | |
3423 | Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael | |
3424 | Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar, | |
3425 | Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert | |
3426 | Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef | |
3427 | Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas | |
3428 | Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets, | |
3429 | Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut | |
3430 | Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek | |
3431 | ||
ccddd104 | 3432 | — Berlin, 2014-08-19 |
b72ddf0f | 3433 | |
3dff3e00 | 3434 | CHANGES WITH 215: |
24a2bf4c LP |
3435 | |
3436 | * A new tool systemd-sysusers has been added. This tool | |
3437 | creates system users and groups in /etc/passwd and | |
3438 | /etc/group, based on static declarative system user/group | |
3439 | definitions in /usr/lib/sysusers.d/. This is useful to | |
3440 | enable factory resets and volatile systems that boot up with | |
3441 | an empty /etc directory, and thus need system users and | |
3442 | groups created during early boot. systemd now also ships | |
3443 | with two default sysusers.d/ files for the most basic | |
3444 | users and groups systemd and the core operating system | |
3445 | require. | |
3446 | ||
3447 | * A new tmpfiles snippet has been added that rebuilds the | |
3448 | essential files in /etc on boot, should they be missing. | |
3449 | ||
3450 | * A directive for ensuring automatic clean-up of | |
3451 | /var/cache/man/ has been removed from the default | |
3452 | configuration. This line should now be shipped by the man | |
3453 | implementation. The necessary change has been made to the | |
3454 | man-db implementation. Note that you need to update your man | |
3455 | implementation to one that ships this line, otherwise no | |
3456 | automatic clean-up of /var/cache/man will take place. | |
3457 | ||
3458 | * A new condition ConditionNeedsUpdate= has been added that | |
3459 | may conditionalize services to only run when /etc or /var | |
3460 | are "older" than the vendor operating system resources in | |
3461 | /usr. This is useful for reconstructing or updating /etc | |
3462 | after an offline update of /usr or a factory reset, on the | |
3463 | next reboot. Services that want to run once after such an | |
3464 | update or reset should use this condition and order | |
3465 | themselves before the new systemd-update-done.service, which | |
3466 | will mark the two directories as fully updated. A number of | |
3467 | service files have been added making use of this, to rebuild | |
3468 | the udev hardware database, the journald message catalog and | |
3469 | dynamic loader cache (ldconfig). The systemd-sysusers tool | |
3470 | described above also makes use of this now. With this in | |
3471 | place it is now possible to start up a minimal operating | |
ce1dde29 | 3472 | system with /etc empty cleanly. For more information on the |
24a2bf4c LP |
3473 | concepts involved see this recent blog story: |
3474 | ||
3475 | http://0pointer.de/blog/projects/stateless.html | |
3476 | ||
3477 | * A new system group "input" has been introduced, and all | |
3478 | input device nodes get this group assigned. This is useful | |
3479 | for system-level software to get access to input devices. It | |
3dff3e00 KS |
3480 | complements what is already done for "audio" and "video". |
3481 | ||
24a2bf4c LP |
3482 | * systemd-networkd learnt minimal DHCPv4 server support in |
3483 | addition to the existing DHCPv4 client support. It also | |
3484 | learnt DHCPv6 client and IPv6 Router Solicitation client | |
3485 | support. The DHCPv4 client gained support for static routes | |
3486 | passed in from the server. Note that the [DHCPv4] section | |
3487 | known in older systemd-networkd versions has been renamed to | |
3488 | [DHCP] and is now also used by the DHCPv6 client. Existing | |
c7435cc9 LP |
3489 | .network files using settings of this section should be |
3490 | updated, though compatibility is maintained. Optionally, the | |
3491 | client hostname may now be sent to the DHCP server. | |
24a2bf4c | 3492 | |
c7435cc9 LP |
3493 | * networkd gained support for vxlan virtual networks as well |
3494 | as tun/tap and dummy devices. | |
24a2bf4c LP |
3495 | |
3496 | * networkd gained support for automatic allocation of address | |
3497 | ranges for interfaces from a system-wide pool of | |
3498 | addresses. This is useful for dynamically managing a large | |
3499 | number of interfaces with a single network configuration | |
3500 | file. In particular this is useful to easily assign | |
3501 | appropriate IP addresses to the veth links of a large number | |
3502 | of nspawn instances. | |
3503 | ||
3504 | * RPM macros for processing sysusers, sysctl and binfmt | |
3505 | drop-in snippets at package installation time have been | |
3506 | added. | |
3507 | ||
3508 | * The /etc/os-release file should now be placed in | |
3509 | /usr/lib/os-release. The old location is automatically | |
3510 | created as symlink. /usr/lib is the more appropriate | |
3511 | location of this file, since it shall actually describe the | |
3512 | vendor operating system shipped in /usr, and not the | |
3513 | configuration stored in /etc. | |
3514 | ||
3515 | * .mount units gained a new boolean SloppyOptions= setting | |
3516 | that maps to mount(8)'s -s option which enables permissive | |
3517 | parsing of unknown mount options. | |
3518 | ||
3519 | * tmpfiles learnt a new "L+" directive which creates a symlink | |
3520 | but (unlike "L") deletes a pre-existing file first, should | |
3521 | it already exist and not already be the correct | |
a8eaaee7 | 3522 | symlink. Similarly, "b+", "c+" and "p+" directives have been |
24a2bf4c LP |
3523 | added as well, which create block and character devices, as |
3524 | well as fifos in the filesystem, possibly removing any | |
3525 | pre-existing files of different types. | |
3526 | ||
3527 | * For tmpfiles' "L", "L+", "C" and "C+" directives the final | |
3528 | 'argument' field (which so far specified the source to | |
ce1dde29 | 3529 | symlink/copy the files from) is now optional. If omitted the |
24a2bf4c LP |
3530 | same file os copied from /usr/share/factory/ suffixed by the |
3531 | full destination path. This is useful for populating /etc | |
3532 | with essential files, by copying them from vendor defaults | |
3533 | shipped in /usr/share/factory/etc. | |
3534 | ||
3535 | * A new command "systemctl preset-all" has been added that | |
3536 | applies the service preset settings to all installed unit | |
3537 | files. A new switch --preset-mode= has been added that | |
3538 | controls whether only enable or only disable operations | |
3539 | shall be executed. | |
3540 | ||
3541 | * A new command "systemctl is-system-running" has been added | |
3542 | that allows checking the overall state of the system, for | |
ce1dde29 | 3543 | example whether it is fully up and running. |
24a2bf4c LP |
3544 | |
3545 | * When the system boots up with an empty /etc, the equivalent | |
3546 | to "systemctl preset-all" is executed during early boot, to | |
3547 | make sure all default services are enabled after a factory | |
3548 | reset. | |
3549 | ||
3550 | * systemd now contains a minimal preset file that enables the | |
3551 | most basic services systemd ships by default. | |
3552 | ||
3553 | * Unit files' [Install] section gained a new DefaultInstance= | |
3554 | field for defining the default instance to create if a | |
3555 | template unit is enabled with no instance specified. | |
3556 | ||
3557 | * A new passive target cryptsetup-pre.target has been added | |
3558 | that may be used by services that need to make they run and | |
3559 | finish before the first LUKS cryptographic device is set up. | |
3560 | ||
3561 | * The /dev/loop-control and /dev/btrfs-control device nodes | |
3562 | are now owned by the "disk" group by default, opening up | |
3563 | access to this group. | |
3564 | ||
3565 | * systemd-coredump will now automatically generate a | |
3566 | stack trace of all core dumps taking place on the system, | |
3567 | based on elfutils' libdw library. This stack trace is logged | |
3568 | to the journal. | |
3569 | ||
3570 | * systemd-coredump may now optionally store coredumps directly | |
3571 | on disk (in /var/lib/systemd/coredump, possibly compressed), | |
3572 | instead of storing them unconditionally in the journal. This | |
3573 | mode is the new default. A new configuration file | |
3574 | /etc/systemd/coredump.conf has been added to configure this | |
3575 | and other parameters of systemd-coredump. | |
3576 | ||
3577 | * coredumpctl gained a new "info" verb to show details about a | |
3578 | specific coredump. A new switch "-1" has also been added | |
3579 | that makes sure to only show information about the most | |
3580 | recent entry instead of all entries. Also, as the tool is | |
3581 | generally useful now the "systemd-" prefix of the binary | |
3582 | name has been removed. Distributions that want to maintain | |
3583 | compatibility with the old name should add a symlink from | |
3584 | the old name to the new name. | |
3585 | ||
3586 | * journald's SplitMode= now defaults to "uid". This makes sure | |
ce1dde29 | 3587 | that unprivileged users can access their own coredumps with |
24a2bf4c LP |
3588 | coredumpctl without restrictions. |
3589 | ||
3590 | * New kernel command line options "systemd.wants=" (for | |
3591 | pulling an additional unit during boot), "systemd.mask=" | |
3592 | (for masking a specific unit for the boot), and | |
3593 | "systemd.debug-shell" (for enabling the debug shell on tty9) | |
3594 | have been added. This is implemented in the new generator | |
3595 | "systemd-debug-generator". | |
3596 | ||
3597 | * systemd-nspawn will now by default filter a couple of | |
3598 | syscalls for containers, among them those required for | |
3599 | kernel module loading, direct x86 IO port access, swap | |
3600 | management, and kexec. Most importantly though | |
3601 | open_by_handle_at() is now prohibited for containers, | |
3602 | closing a hole similar to a recently discussed vulnerability | |
3603 | in docker regarding access to files on file hierarchies the | |
b938cb90 JE |
3604 | container should normally not have access to. Note that, for |
3605 | nspawn, we generally make no security claims anyway (and | |
24a2bf4c LP |
3606 | this is explicitly documented in the man page), so this is |
3607 | just a fix for one of the most obvious problems. | |
3608 | ||
3609 | * A new man page file-hierarchy(7) has been added that | |
3610 | contains a minimized, modernized version of the file system | |
3611 | layout systemd expects, similar in style to the FHS | |
c7435cc9 LP |
3612 | specification or hier(5). A new tool systemd-path(1) has |
3613 | been added to query many of these paths for the local | |
3614 | machine and user. | |
24a2bf4c LP |
3615 | |
3616 | * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no | |
3617 | longer done. Since the directory now has a per-user size | |
3618 | limit, and is cleaned on logout this appears unnecessary, | |
3619 | in particular since this now brings the lifecycle of this | |
3620 | directory closer in line with how IPC objects are handled. | |
3621 | ||
3622 | * systemd.pc now exports a number of additional directories, | |
3623 | including $libdir (which is useful to identify the library | |
3624 | path for the primary architecture of the system), and a | |
3625 | couple of drop-in directories. | |
3626 | ||
3058e017 TLSC |
3627 | * udev's predictable network interface names now use the dev_port |
3628 | sysfs attribute, introduced in linux 3.15 instead of dev_id to | |
3629 | distinguish between ports of the same PCI function. dev_id should | |
3630 | only be used for ports using the same HW address, hence the need | |
3631 | for dev_port. | |
3632 | ||
c7435cc9 LP |
3633 | * machined has been updated to export the OS version of a |
3634 | container (read from /etc/os-release and | |
3635 | /usr/lib/os-release) on the bus. This is now shown in | |
3636 | "machinectl status" for a machine. | |
3637 | ||
3638 | * A new service setting RestartForceExitStatus= has been | |
3639 | added. If configured to a set of exit signals or process | |
3640 | return values, the service will be restarted when the main | |
3641 | daemon process exits with any of them, regardless of the | |
3642 | Restart= setting. | |
3643 | ||
3644 | * systemctl's -H switch for connecting to remote systemd | |
3645 | machines has been extended so that it may be used to | |
3646 | directly connect to a specific container on the | |
3647 | host. "systemctl -H root@foobar:waldi" will now connect as | |
3648 | user "root" to host "foobar", and then proceed directly to | |
3649 | the container named "waldi". Note that currently you have to | |
3650 | authenticate as user "root" for this to work, as entering | |
3651 | containers is a privileged operation. | |
3652 | ||
3653 | Contributions from: Andreas Henriksson, Benjamin Steinwender, | |
3654 | Carl Schaefer, Christian Hesse, Colin Ian King, Cristian | |
3655 | Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene | |
3656 | Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo | |
3657 | Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart | |
3658 | Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine | |
3659 | Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich, | |
3660 | Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le | |
3661 | Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan, | |
3662 | Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe | |
3663 | Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar | |
3664 | Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek | |
3665 | ||
ccddd104 | 3666 | — Berlin, 2014-07-03 |
c7435cc9 | 3667 | |
4196a3ea KS |
3668 | CHANGES WITH 214: |
3669 | ||
3670 | * As an experimental feature, udev now tries to lock the | |
3671 | disk device node (flock(LOCK_SH|LOCK_NB)) while it | |
3672 | executes events for the disk or any of its partitions. | |
3673 | Applications like partitioning programs can lock the | |
3674 | disk device node (flock(LOCK_EX)) and claim temporary | |
3675 | device ownership that way; udev will entirely skip all event | |
3676 | handling for this disk and its partitions. If the disk | |
3677 | was opened for writing, the close will trigger a partition | |
3678 | table rescan in udev's "watch" facility, and if needed | |
71449caf | 3679 | synthesize "change" events for the disk and all its partitions. |
8d0e0ddd | 3680 | This is now unconditionally enabled, and if it turns out to |
4196a3ea | 3681 | cause major problems, we might turn it on only for specific |
45df8656 | 3682 | devices, or might need to disable it entirely. Device Mapper |
4196a3ea KS |
3683 | devices are excluded from this logic. |
3684 | ||
04e91da2 LP |
3685 | * We temporarily dropped the "-l" switch for fsck invocations, |
3686 | since they collide with the flock() logic above. util-linux | |
3687 | upstream has been changed already to avoid this conflict, | |
3688 | and we will readd "-l" as soon as util-linux with this | |
3689 | change has been released. | |
3690 | ||
3691 | * The dependency on libattr has been removed. Since a long | |
8d0e0ddd | 3692 | time, the extended attribute calls have moved to glibc, and |
04e91da2 LP |
3693 | libattr is thus unnecessary. |
3694 | ||
ce830873 | 3695 | * Virtualization detection works without privileges now. This |
04e91da2 LP |
3696 | means the systemd-detect-virt binary no longer requires |
3697 | CAP_SYS_PTRACE file capabilities, and our daemons can run | |
71449caf | 3698 | with fewer privileges. |
04e91da2 LP |
3699 | |
3700 | * systemd-networkd now runs under its own "systemd-network" | |
3701 | user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE, | |
3702 | CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but | |
3703 | loses the ability to write to files owned by root this way. | |
3704 | ||
a8eaaee7 | 3705 | * Similarly, systemd-resolved now runs under its own |
04e91da2 LP |
3706 | "systemd-resolve" user with no capabilities remaining. |
3707 | ||
a8eaaee7 | 3708 | * Similarly, systemd-bus-proxyd now runs under its own |
04e91da2 LP |
3709 | "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining. |
3710 | ||
3711 | * systemd-networkd gained support for setting up "veth" | |
a8eaaee7 | 3712 | virtual Ethernet devices for container connectivity, as well |
04e91da2 LP |
3713 | as GRE and VTI tunnels. |
3714 | ||
3715 | * systemd-networkd will no longer automatically attempt to | |
3716 | manually load kernel modules necessary for certain tunnel | |
8d0e0ddd | 3717 | transports. Instead, it is assumed the kernel loads them |
04e91da2 LP |
3718 | automatically when required. This only works correctly on |
3719 | very new kernels. On older kernels, please consider adding | |
c54bed5d | 3720 | the kernel modules to /etc/modules-load.d/ as a work-around. |
04e91da2 | 3721 | |
cd14eda3 | 3722 | * The resolv.conf file systemd-resolved generates has been |
8d0e0ddd JE |
3723 | moved to /run/systemd/resolve/. If you have a symlink from |
3724 | /etc/resolv.conf, it might be necessary to correct it. | |
cd14eda3 | 3725 | |
ef392da6 | 3726 | * Two new service settings, ProtectHome= and ProtectSystem=, |
8d0e0ddd | 3727 | have been added. When enabled, they will make the user data |
04e91da2 LP |
3728 | (such as /home) inaccessible or read-only and the system |
3729 | (such as /usr) read-only, for specific services. This allows | |
3730 | very light-weight per-service sandboxing to avoid | |
3731 | modifications of user data or system files from | |
3732 | services. These two new switches have been enabled for all | |
3733 | of systemd's long-running services, where appropriate. | |
3734 | ||
3735 | * Socket units gained new SocketUser= and SocketGroup= | |
3736 | settings to set the owner user and group of AF_UNIX sockets | |
3737 | and FIFOs in the file system. | |
3738 | ||
8d0e0ddd | 3739 | * Socket units gained a new RemoveOnStop= setting. If enabled, |
04e91da2 LP |
3740 | all FIFOS and sockets in the file system will be removed |
3741 | when the specific socket unit is stopped. | |
3742 | ||
3743 | * Socket units gained a new Symlinks= setting. It takes a list | |
3744 | of symlinks to create to file system sockets or FIFOs | |
45df8656 | 3745 | created by the specific Unix sockets. This is useful to |
71449caf | 3746 | manage symlinks to socket nodes with the same life-cycle as |
04e91da2 LP |
3747 | the socket itself. |
3748 | ||
3749 | * The /dev/log socket and /dev/initctl FIFO have been moved to | |
3750 | /run, and have been replaced by symlinks. This allows | |
3751 | connecting to these facilities even if PrivateDevices=yes is | |
3752 | used for a service (which makes /dev/log itself unavailable, | |
3753 | but /run is left). This also has the benefit of ensuring | |
3754 | that /dev only contains device nodes, directories and | |
3755 | symlinks, and nothing else. | |
3756 | ||
3757 | * sd-daemon gained two new calls sd_pid_notify() and | |
3758 | sd_pid_notifyf(). They are similar to sd_notify() and | |
3759 | sd_notifyf(), but allow overriding of the source PID of | |
3760 | notification messages if permissions permit this. This is | |
3761 | useful to send notify messages on behalf of a different | |
3762 | process (for example, the parent process). The | |
3763 | systemd-notify tool has been updated to make use of this | |
3764 | when sending messages (so that notification messages now | |
3765 | originate from the shell script invoking systemd-notify and | |
3766 | not the systemd-notify process itself. This should minimize | |
3767 | a race where systemd fails to associate notification | |
3768 | messages to services when the originating process already | |
3769 | vanished. | |
3770 | ||
3771 | * A new "on-abnormal" setting for Restart= has been added. If | |
8d0e0ddd | 3772 | set, it will result in automatic restarts on all "abnormal" |
04e91da2 LP |
3773 | reasons for a process to exit, which includes unclean |
3774 | signals, core dumps, timeouts and watchdog timeouts, but | |
3775 | does not include clean and unclean exit codes or clean | |
3776 | signals. Restart=on-abnormal is an alternative for | |
3777 | Restart=on-failure for services that shall be able to | |
3778 | terminate and avoid restarts on certain errors, by | |
3779 | indicating so with an unclean exit code. Restart=on-failure | |
3780 | or Restart=on-abnormal is now the recommended setting for | |
3781 | all long-running services. | |
3782 | ||
3783 | * If the InaccessibleDirectories= service setting points to a | |
3784 | mount point (or if there are any submounts contained within | |
3785 | it), it is now attempted to completely unmount it, to make | |
3786 | the file systems truly unavailable for the respective | |
3787 | service. | |
3788 | ||
3789 | * The ReadOnlyDirectories= service setting and | |
3790 | systemd-nspawn's --read-only parameter are now recursively | |
3791 | applied to all submounts, too. | |
3792 | ||
3793 | * Mount units may now be created transiently via the bus APIs. | |
3794 | ||
3795 | * The support for SysV and LSB init scripts has been removed | |
3796 | from the systemd daemon itself. Instead, it is now | |
3797 | implemented as a generator that creates native systemd units | |
3798 | from these scripts when needed. This enables us to remove a | |
3799 | substantial amount of legacy code from PID 1, following the | |
3800 | fact that many distributions only ship a very small number | |
3801 | of LSB/SysV init scripts nowadays. | |
3802 | ||
cc98b302 | 3803 | * Privileged Xen (dom0) domains are not considered |
04e91da2 LP |
3804 | virtualization anymore by the virtualization detection |
3805 | logic. After all, they generally have unrestricted access to | |
71449caf | 3806 | the hardware and usually are used to manage the unprivileged |
04e91da2 LP |
3807 | (domU) domains. |
3808 | ||
3809 | * systemd-tmpfiles gained a new "C" line type, for copying | |
3810 | files or entire directories. | |
3811 | ||
3812 | * systemd-tmpfiles "m" lines are now fully equivalent to "z" | |
8d0e0ddd JE |
3813 | lines. So far, they have been non-globbing versions of the |
3814 | latter, and have thus been redundant. In future, it is | |
3815 | recommended to only use "z". "m" has hence been removed | |
04e91da2 LP |
3816 | from the documentation, even though it stays supported. |
3817 | ||
3818 | * A tmpfiles snippet to recreate the most basic structure in | |
3819 | /var has been added. This is enough to create the /var/run → | |
3820 | /run symlink and create a couple of structural | |
3821 | directories. This allows systems to boot up with an empty or | |
8d0e0ddd JE |
3822 | volatile /var. Of course, while with this change, the core OS |
3823 | now is capable with dealing with a volatile /var, not all | |
04e91da2 | 3824 | user services are ready for it. However, we hope that sooner |
8d0e0ddd | 3825 | or later, many service daemons will be changed upstream so |
04e91da2 LP |
3826 | that they are able to automatically create their necessary |
3827 | directories in /var at boot, should they be missing. This is | |
3828 | the first step to allow state-less systems that only require | |
3829 | the vendor image for /usr to boot. | |
3830 | ||
3831 | * systemd-nspawn has gained a new --tmpfs= switch to mount an | |
3832 | empty tmpfs instance to a specific directory. This is | |
3833 | particularly useful for making use of the automatic | |
3834 | reconstruction of /var (see above), by passing --tmpfs=/var. | |
3835 | ||
3836 | * Access modes specified in tmpfiles snippets may now be | |
3837 | prefixed with "~", which indicates that they shall be masked | |
daa05349 | 3838 | by whether the existing file or directory is currently |
8d0e0ddd | 3839 | writable, readable or executable at all. Also, if specified, |
04e91da2 LP |
3840 | the sgid/suid/sticky bits will be masked for all |
3841 | non-directories. | |
3842 | ||
3843 | * A new passive target unit "network-pre.target" has been | |
3844 | added which is useful for services that shall run before any | |
3845 | network is configured, for example firewall scripts. | |
3846 | ||
4c0d13bd LP |
3847 | * The "floppy" group that previously owned the /dev/fd* |
3848 | devices is no longer used. The "disk" group is now used | |
3849 | instead. Distributions should probably deprecate usage of | |
3850 | this group. | |
3851 | ||
dc1d6c02 LP |
3852 | Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian |
3853 | King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David | |
3854 | Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers, | |
3855 | Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny | |
3856 | Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel | |
3857 | Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew | |
3858 | Jędrzejewski-Szmek | |
3859 | ||
ccddd104 | 3860 | — Berlin, 2014-06-11 |
dc1d6c02 | 3861 | |
6936cd89 LP |
3862 | CHANGES WITH 213: |
3863 | ||
3864 | * A new "systemd-timesyncd" daemon has been added for | |
69beda1f | 3865 | synchronizing the system clock across the network. It |
6936cd89 | 3866 | implements an SNTP client. In contrast to NTP |
8d0e0ddd | 3867 | implementations such as chrony or the NTP reference server, |
6936cd89 | 3868 | this only implements a client side, and does not bother with |
c9679c65 LP |
3869 | the full NTP complexity, focusing only on querying time from |
3870 | one remote server and synchronizing the local clock to | |
6936cd89 | 3871 | it. Unless you intend to serve NTP to networked clients or |
8d0e0ddd | 3872 | want to connect to local hardware clocks, this simple NTP |
6936cd89 LP |
3873 | client should be more than appropriate for most |
3874 | installations. The daemon runs with minimal privileges, and | |
3875 | has been hooked up with networkd to only operate when | |
3876 | network connectivity is available. The daemon saves the | |
3877 | current clock to disk every time a new NTP sync has been | |
3878 | acquired, and uses this to possibly correct the system clock | |
69beda1f | 3879 | early at bootup, in order to accommodate for systems that |
6936cd89 | 3880 | lack an RTC such as the Raspberry Pi and embedded devices, |
8d0e0ddd | 3881 | and to make sure that time monotonically progresses on these |
c9679c65 | 3882 | systems, even if it is not always correct. To make use of |
8d0e0ddd | 3883 | this daemon, a new system user and group "systemd-timesync" |
c9679c65 | 3884 | needs to be created on installation of systemd. |
6936cd89 | 3885 | |
69beda1f KS |
3886 | * The queue "seqnum" interface of libudev has been disabled, as |
3887 | it was generally incompatible with device namespacing as | |
6936cd89 LP |
3888 | sequence numbers of devices go "missing" if the devices are |
3889 | part of a different namespace. | |
3890 | ||
3891 | * "systemctl list-timers" and "systemctl list-sockets" gained | |
3892 | a --recursive switch for showing units of these types also | |
499b604b ZJS |
3893 | for all local containers, similar in style to the already |
3894 | supported --recursive switch for "systemctl list-units". | |
6936cd89 LP |
3895 | |
3896 | * A new RebootArgument= setting has been added for service | |
3897 | units, which may be used to specify a kernel reboot argument | |
499b604b | 3898 | to use when triggering reboots with StartLimitAction=. |
6936cd89 LP |
3899 | |
3900 | * A new FailureAction= setting has been added for service | |
3901 | units which may be used to specify an operation to trigger | |
499b604b | 3902 | when a service fails. This works similarly to |
8d0e0ddd | 3903 | StartLimitAction=, but unlike it, controls what is done |
6936cd89 LP |
3904 | immediately rather than only after several attempts to |
3905 | restart the service in question. | |
3906 | ||
3907 | * hostnamed got updated to also expose the kernel name, | |
499b604b ZJS |
3908 | release, and version on the bus. This is useful for |
3909 | executing commands like hostnamectl with the -H switch. | |
3910 | systemd-analyze makes use of this to properly display | |
3911 | details when running non-locally. | |
6936cd89 LP |
3912 | |
3913 | * The bootchart tool can now show cgroup information in the | |
3914 | graphs it generates. | |
3915 | ||
3916 | * The CFS CPU quota cgroup attribute is now exposed for | |
3917 | services. The new CPUQuota= switch has been added for this | |
3918 | which takes a percentage value. Setting this will have the | |
3919 | result that a service may never get more CPU time than the | |
3920 | specified percentage, even if the machine is otherwise idle. | |
3921 | ||
3922 | * systemd-networkd learned IPIP and SIT tunnel support. | |
3923 | ||
3924 | * LSB init scripts exposing a dependency on $network will now | |
3925 | get a dependency on network-online.target rather than simply | |
3926 | network.target. This should bring LSB handling closer to | |
3927 | what it was on SysV systems. | |
3928 | ||
3929 | * A new fsck.repair= kernel option has been added to control | |
3930 | how fsck shall deal with unclean file systems at boot. | |
3931 | ||
3932 | * The (.ini) configuration file parser will now silently | |
3933 | ignore sections whose name begins with "X-". This may be | |
3934 | used to maintain application-specific extension sections in unit | |
3935 | files. | |
3936 | ||
3937 | * machined gained a new API to query the IP addresses of | |
3938 | registered containers. "machinectl status" has been updated | |
3939 | to show these addresses in its output. | |
3940 | ||
3941 | * A new call sd_uid_get_display() has been added to the | |
3942 | sd-login APIs for querying the "primary" session of a | |
3943 | user. The "primary" session of the user is elected from the | |
3944 | user's sessions and generally a graphical session is | |
3945 | preferred over a text one. | |
3946 | ||
3947 | * A minimal systemd-resolved daemon has been added. It | |
3948 | currently simply acts as a companion to systemd-networkd and | |
3949 | manages resolv.conf based on per-interface DNS | |
3950 | configuration, possibly supplied via DHCP. In the long run | |
3951 | we hope to extend this into a local DNSSEC enabled DNS and | |
3952 | mDNS cache. | |
3953 | ||
68dd0956 TG |
3954 | * The systemd-networkd-wait-online tool is now enabled by |
3955 | default. It will delay network-online.target until a network | |
3956 | connection has been configured. The tool primarily integrates | |
3957 | with networkd, but will also make a best effort to make sense | |
3958 | of network configuration performed in some other way. | |
3959 | ||
6936cd89 | 3960 | * Two new service options StartupCPUShares= and |
499b604b | 3961 | StartupBlockIOWeight= have been added that work similarly to |
6936cd89 | 3962 | CPUShares= and BlockIOWeight= however only apply during |
69beda1f | 3963 | system startup. This is useful to prioritize certain services |
6936cd89 LP |
3964 | differently during bootup than during normal runtime. |
3965 | ||
8e7acf67 LP |
3966 | * hostnamed has been changed to prefer the statically |
3967 | configured hostname in /etc/hostname (unless set to | |
3968 | 'localhost' or empty) over any dynamic one supplied by | |
8d0e0ddd | 3969 | dhcp. With this change, the rules for picking the hostname |
8e7acf67 LP |
3970 | match more closely the rules of other configuration settings |
3971 | where the local administrator's configuration in /etc always | |
3972 | overrides any other settings. | |
3973 | ||
3974 | Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van | |
6936cd89 LP |
3975 | den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch, |
3976 | Dan Kilman, Dave Reisner, David Härdeman, David Herrmann, | |
3977 | David Strauss, Dimitris Spingos, Djalal Harouni, Eelco | |
3978 | Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg | |
3979 | Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan | |
3980 | Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark, | |
3981 | Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas | |
3982 | Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas, | |
8e7acf67 LP |
3983 | Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael |
3984 | Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis | |
3985 | Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode, | |
3986 | Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter, | |
3987 | Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler, | |
3988 | Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar | |
3989 | Lindskog, WaLyong Cho, Will Woods, Zbigniew | |
6936cd89 LP |
3990 | Jędrzejewski-Szmek |
3991 | ||
ccddd104 | 3992 | — Beijing, 2014-05-28 |
6936cd89 | 3993 | |
51c61cda LP |
3994 | CHANGES WITH 212: |
3995 | ||
3996 | * When restoring the screen brightness at boot, stay away from | |
3997 | the darkest setting or from the lowest 5% of the available | |
3998 | range, depending on which is the larger value of both. This | |
3999 | should effectively protect the user from rebooting into a | |
4000 | black screen, should the brightness have been set to minimum | |
4001 | by accident. | |
4002 | ||
4003 | * sd-login gained a new sd_machine_get_class() call to | |
4004 | determine the class ("vm" or "container") of a machine | |
4005 | registered with machined. | |
4006 | ||
4007 | * sd-login gained new calls | |
4008 | sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(), | |
4009 | to query the identity of the peer of a local AF_UNIX | |
499b604b | 4010 | connection. They operate similarly to their sd_pid_get_xyz() |
51c61cda LP |
4011 | counterparts. |
4012 | ||
4013 | * PID 1 will now maintain a system-wide system state engine | |
4014 | with the states "starting", "running", "degraded", | |
4015 | "maintenance", "stopping". These states are bound to system | |
4016 | startup, normal runtime, runtime with at least one failed | |
4017 | service, rescue/emergency mode and system shutdown. This | |
4018 | state is shown in the "systemctl status" output when no unit | |
4019 | name is passed. It is useful to determine system state, in | |
4020 | particularly when doing so for many systems or containers at | |
4021 | once. | |
4022 | ||
4023 | * A new command "list-machines" has been added to "systemctl" | |
4024 | that lists all local OS containers and shows their system | |
4025 | state (see above), if systemd runs inside of them. | |
4026 | ||
4027 | * systemctl gained a new "-r" switch to recursively enumerate | |
4028 | units on all local containers, when used with the | |
4029 | "list-unit" command (which is the default one that is | |
4030 | executed when no parameters are specified). | |
4031 | ||
4032 | * The GPT automatic partition discovery logic will now honour | |
4033 | two GPT partition flags: one may be set on a partition to | |
4034 | cause it to be mounted read-only, and the other may be set | |
4035 | on a partition to ignore it during automatic discovery. | |
4036 | ||
4037 | * Two new GPT type UUIDs have been added for automatic root | |
70a44afe | 4038 | partition discovery, for 32-bit and 64-bit ARM. This is not |
51c61cda LP |
4039 | particularly useful for discovering the root directory on |
4040 | these architectures during bare-metal boots (since UEFI is | |
4041 | not common there), but still very useful to allow booting of | |
4042 | ARM disk images in nspawn with the -i option. | |
4043 | ||
4044 | * MAC addresses of interfaces created with nspawn's | |
4045 | --network-interface= switch will now be generated from the | |
4046 | machine name, and thus be stable between multiple invocations | |
4047 | of the container. | |
4048 | ||
4049 | * logind will now automatically remove all IPC objects owned | |
4050 | by a user if she or he fully logs out. This makes sure that | |
4051 | users who are logged out cannot continue to consume IPC | |
4052 | resources. This covers SysV memory, semaphores and message | |
4053 | queues as well as POSIX shared memory and message | |
b8bde116 JE |
4054 | queues. Traditionally, SysV and POSIX IPC had no life-cycle |
4055 | limits. With this functionality, that is corrected. This may | |
4056 | be turned off by using the RemoveIPC= switch of logind.conf. | |
51c61cda LP |
4057 | |
4058 | * The systemd-machine-id-setup and tmpfiles tools gained a | |
4059 | --root= switch to operate on a specific root directory, | |
4060 | instead of /. | |
4061 | ||
4062 | * journald can now forward logged messages to the TTYs of all | |
4063 | logged in users ("wall"). This is the default for all | |
4064 | emergency messages now. | |
4065 | ||
4066 | * A new tool systemd-journal-remote has been added to stream | |
4067 | journal log messages across the network. | |
4068 | ||
4069 | * /sys/fs/cgroup/ is now mounted read-only after all cgroup | |
4070 | controller trees are mounted into it. Note that the | |
4071 | directories mounted beneath it are not read-only. This is a | |
4072 | security measure and is particularly useful because glibc | |
4073 | actually includes a search logic to pick any tmpfs it can | |
4074 | find to implement shm_open() if /dev/shm is not available | |
4075 | (which it might very well be in namespaced setups). | |
4076 | ||
4077 | * machinectl gained a new "poweroff" command to cleanly power | |
4078 | down a local OS container. | |
4079 | ||
4080 | * The PrivateDevices= unit file setting will now also drop the | |
4081 | CAP_MKNOD capability from the capability bound set, and | |
4082 | imply DevicePolicy=closed. | |
4083 | ||
4084 | * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used | |
4085 | comprehensively on all long-running systemd services where | |
4086 | this is appropriate. | |
4087 | ||
4088 | * systemd-udevd will now run in a disassociated mount | |
b8bde116 | 4089 | namespace. To mount directories from udev rules, make sure to |
51c61cda LP |
4090 | pull in mount units via SYSTEMD_WANTS properties. |
4091 | ||
4092 | * The kdbus support gained support for uploading policy into | |
4093 | the kernel. sd-bus gained support for creating "monitoring" | |
4094 | connections that can eavesdrop into all bus communication | |
4095 | for debugging purposes. | |
4096 | ||
4097 | * Timestamps may now be specified in seconds since the UNIX | |
4098 | epoch Jan 1st, 1970 by specifying "@" followed by the value | |
4099 | in seconds. | |
4100 | ||
4101 | * Native tcpwrap support in systemd has been removed. tcpwrap | |
4102 | is old code, not really maintained anymore and has serious | |
4103 | shortcomings, and better options such as firewalls | |
4104 | exist. For setups that require tcpwrap usage, please | |
4105 | consider invoking your socket-activated service via tcpd, | |
4106 | like on traditional inetd. | |
4107 | ||
4108 | * A new system.conf configuration option | |
4109 | DefaultTimerAccuracySec= has been added that controls the | |
4110 | default AccuracySec= setting of .timer units. | |
4111 | ||
b8bde116 | 4112 | * Timer units gained a new WakeSystem= switch. If enabled, |
51c61cda LP |
4113 | timers configured this way will cause the system to resume |
4114 | from system suspend (if the system supports that, which most | |
4115 | do these days). | |
4116 | ||
b8bde116 | 4117 | * Timer units gained a new Persistent= switch. If enabled, |
51c61cda LP |
4118 | timers configured this way will save to disk when they have |
4119 | been last triggered. This information is then used on next | |
4120 | reboot to possible execute overdue timer events, that | |
d28315e4 JE |
4121 | could not take place because the system was powered off. |
4122 | This enables simple anacron-like behaviour for timer units. | |
51c61cda LP |
4123 | |
4124 | * systemctl's "list-timers" will now also list the time a | |
4125 | timer unit was last triggered in addition to the next time | |
4126 | it will be triggered. | |
4127 | ||
4128 | * systemd-networkd will now assign predictable IPv4LL | |
4129 | addresses to its local interfaces. | |
4130 | ||
4131 | Contributions from: Brandon Philips, Daniel Buch, Daniel Mack, | |
4132 | Dave Reisner, David Herrmann, Gerd Hoffmann, Greg | |
4133 | Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh | |
4134 | Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine | |
4135 | Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna, | |
4136 | Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler, | |
4137 | Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen, | |
4138 | Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew | |
4139 | Jędrzejewski-Szmek | |
4140 | ||
ccddd104 | 4141 | — Berlin, 2014-03-25 |
51c61cda | 4142 | |
699b6b34 LP |
4143 | CHANGES WITH 211: |
4144 | ||
4145 | * A new unit file setting RestrictAddressFamilies= has been | |
4146 | added to restrict which socket address families unit | |
4147 | processes gain access to. This takes address family names | |
4148 | like "AF_INET" or "AF_UNIX", and is useful to minimize the | |
4149 | attack surface of services via exotic protocol stacks. This | |
4150 | is built on seccomp system call filters. | |
4151 | ||
4152 | * Two new unit file settings RuntimeDirectory= and | |
4153 | RuntimeDirectoryMode= have been added that may be used to | |
4154 | manage a per-daemon runtime directories below /run. This is | |
4155 | an alternative for setting up directory permissions with | |
4156 | tmpfiles snippets, and has the advantage that the runtime | |
4157 | directory's lifetime is bound to the daemon runtime and that | |
4158 | the daemon starts up with an empty directory each time. This | |
4159 | is particularly useful when writing services that drop | |
f1721625 | 4160 | privileges using the User= or Group= setting. |
699b6b34 LP |
4161 | |
4162 | * The DeviceAllow= unit setting now supports globbing for | |
4163 | matching against device group names. | |
4164 | ||
4165 | * The systemd configuration file system.conf gained new | |
4166 | settings DefaultCPUAccounting=, DefaultBlockIOAccounting=, | |
4167 | DefaultMemoryAccounting= to globally turn on/off accounting | |
4168 | for specific resources (cgroups) for all units. These | |
22e7062d | 4169 | settings may still be overridden individually in each unit |
699b6b34 LP |
4170 | though. |
4171 | ||
4172 | * systemd-gpt-auto-generator is now able to discover /srv and | |
4173 | root partitions in addition to /home and swap partitions. It | |
4174 | also supports LUKS-encrypted partitions now. With this in | |
b8bde116 | 4175 | place, automatic discovery of partitions to mount following |
699b6b34 | 4176 | the Discoverable Partitions Specification |
56cadcb6 | 4177 | (https://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec) |
699b6b34 LP |
4178 | is now a lot more complete. This allows booting without |
4179 | /etc/fstab and without root= on the kernel command line on | |
b8bde116 | 4180 | systems prepared appropriately. |
699b6b34 LP |
4181 | |
4182 | * systemd-nspawn gained a new --image= switch which allows | |
4183 | booting up disk images and Linux installations on any block | |
4184 | device that follow the Discoverable Partitions Specification | |
4185 | (see above). This means that installations made with | |
4186 | appropriately updated installers may now be started and | |
4187 | deployed using container managers, completely | |
4188 | unmodified. (We hope that libvirt-lxc will add support for | |
4189 | this feature soon, too.) | |
4190 | ||
4191 | * systemd-nspawn gained a new --network-macvlan= setting to | |
4192 | set up a private macvlan interface for the | |
499b604b | 4193 | container. Similarly, systemd-networkd gained a new |
699b6b34 LP |
4194 | Kind=macvlan setting in .netdev files. |
4195 | ||
4196 | * systemd-networkd now supports configuring local addresses | |
4197 | using IPv4LL. | |
4198 | ||
4199 | * A new tool systemd-network-wait-online has been added to | |
4200 | synchronously wait for network connectivity using | |
4201 | systemd-networkd. | |
4202 | ||
4203 | * The sd-bus.h bus API gained a new sd_bus_track object for | |
4204 | tracking the life-cycle of bus peers. Note that sd-bus.h is | |
4205 | still not a public API though (unless you specify | |
4206 | --enable-kdbus on the configure command line, which however | |
4207 | voids your warranty and you get no API stability guarantee). | |
4208 | ||
4209 | * The $XDG_RUNTIME_DIR runtime directories for each user are | |
4210 | now individual tmpfs instances, which has the benefit of | |
4211 | introducing separate pools for each user, with individual | |
4ef6e535 | 4212 | size limits, and thus making sure that unprivileged clients |
699b6b34 LP |
4213 | can no longer negatively impact the system or other users by |
4214 | filling up their $XDG_RUNTIME_DIR. A new logind.conf setting | |
4215 | RuntimeDirectorySize= has been introduced that allows | |
4216 | controlling the default size limit for all users. It | |
4217 | defaults to 10% of the available physical memory. This is no | |
4218 | replacement for quotas on tmpfs though (which the kernel | |
4219 | still does not support), as /dev/shm and /tmp are still | |
4ef6e535 | 4220 | shared resources used by both the system and unprivileged |
699b6b34 LP |
4221 | users. |
4222 | ||
4223 | * logind will now automatically turn off automatic suspending | |
4224 | on laptop lid close when more than one display is | |
4225 | connected. This was previously expected to be implemented | |
4226 | individually in desktop environments (such as GNOME), | |
4227 | however has been added to logind now, in order to fix a | |
4228 | boot-time race where a desktop environment might not have | |
4229 | been started yet and thus not been able to take an inhibitor | |
4230 | lock at the time where logind already suspends the system | |
4231 | due to a closed lid. | |
4232 | ||
4233 | * logind will now wait at least 30s after each system | |
4234 | suspend/resume cycle, and 3min after system boot before | |
4235 | suspending the system due to a closed laptop lid. This | |
4236 | should give USB docking stations and similar enough time to | |
4ef6e535 | 4237 | be probed and configured after system resume and boot in |
699b6b34 LP |
4238 | order to then act as suspend blocker. |
4239 | ||
4240 | * systemd-run gained a new --property= setting which allows | |
4241 | initialization of resource control properties (and others) | |
4242 | for the created scope or service unit. Example: "systemd-run | |
4243 | --property=BlockIOWeight=10 updatedb" may be used to run | |
4244 | updatedb at a low block IO scheduling weight. | |
4245 | ||
4246 | * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches | |
4247 | now also work in --scope mode. | |
4248 | ||
4249 | * When systemd is compiled with kdbus support, basic support | |
4250 | for enforced policies is now in place. (Note that enabling | |
4251 | kdbus still voids your warranty and no API compatibility | |
4252 | promises are made.) | |
4253 | ||
4254 | Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin | |
4255 | K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni, | |
4256 | Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay | |
4257 | Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn, | |
4258 | Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt, | |
4259 | Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef | |
4260 | Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas | |
4261 | Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom | |
4262 | Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook, | |
4263 | Zbigniew Jędrzejewski-Szmek | |
4264 | ||
ccddd104 | 4265 | — Berlin, 2014-03-12 |
699b6b34 | 4266 | |
43c71255 LP |
4267 | CHANGES WITH 210: |
4268 | ||
4269 | * systemd will now relabel /dev after loading the SMACK policy | |
4270 | according to SMACK rules. | |
4271 | ||
67dd87c5 | 4272 | * A new unit file option AppArmorProfile= has been added to |
43c71255 LP |
4273 | set the AppArmor profile for the processes of a unit. |
4274 | ||
4275 | * A new condition check ConditionArchitecture= has been added | |
4276 | to conditionalize units based on the system architecture, as | |
4277 | reported by uname()'s "machine" field. | |
4278 | ||
4279 | * systemd-networkd now supports matching on the system | |
4280 | virtualization, architecture, kernel command line, host name | |
4281 | and machine ID. | |
4282 | ||
ed28905e | 4283 | * logind is now a lot more aggressive when suspending the |
43c71255 | 4284 | machine due to a closed laptop lid. Instead of acting only |
b8bde116 | 4285 | on the lid close action, it will continuously watch the lid |
43c71255 LP |
4286 | status and act on it. This is useful for laptops where the |
4287 | power button is on the outside of the chassis so that it can | |
ed28905e | 4288 | be reached without opening the lid (such as the Lenovo |
b8bde116 | 4289 | Yoga). On those machines, logind will now immediately |
ed28905e | 4290 | re-suspend the machine if the power button has been |
43c71255 LP |
4291 | accidentally pressed while the laptop was suspended and in a |
4292 | backpack or similar. | |
4293 | ||
4294 | * logind will now watch SW_DOCK switches and inhibit reaction | |
4295 | to the lid switch if it is pressed. This means that logind | |
d27893ef | 4296 | will not suspend the machine anymore if the lid is closed |
949138cc | 4297 | and the system is docked, if the laptop supports SW_DOCK |
43c71255 LP |
4298 | notifications via the input layer. Note that ACPI docking |
4299 | stations do not generate this currently. Also note that this | |
4300 | logic is usually not fully sufficient and Desktop | |
4301 | Environments should take a lid switch inhibitor lock when an | |
4302 | external display is connected, as systemd will not watch | |
4303 | this on its own. | |
4304 | ||
4305 | * nspawn will now make use of the devices cgroup controller by | |
4306 | default, and only permit creation of and access to the usual | |
4307 | API device nodes like /dev/null or /dev/random, as well as | |
4308 | access to (but not creation of) the pty devices. | |
4309 | ||
4310 | * We will now ship a default .network file for | |
4311 | systemd-networkd that automatically configures DHCP for | |
4312 | network interfaces created by nspawn's --network-veth or | |
4313 | --network-bridge= switches. | |
4314 | ||
4315 | * systemd will now understand the usual M, K, G, T suffixes | |
4316 | according to SI conventions (i.e. to the base 1000) when | |
4317 | referring to throughput and hardware metrics. It will stay | |
4318 | with IEC conventions (i.e. to the base 1024) for software | |
4319 | metrics, according to what is customary according to | |
4320 | Wikipedia. We explicitly document which base applies for | |
4321 | each configuration option. | |
4322 | ||
4323 | * The DeviceAllow= setting in unit files now supports a syntax | |
ed28905e | 4324 | to whitelist an entire group of devices node majors at once, |
43c71255 | 4325 | based on the /proc/devices listing. For example, with the |
b8bde116 | 4326 | string "char-pts", it is now possible to whitelist all |
43c71255 LP |
4327 | current and future pseudo-TTYs at once. |
4328 | ||
4329 | * sd-event learned a new "post" event source. Event sources of | |
4330 | this type are triggered by the dispatching of any event | |
4331 | source of a type that is not "post". This is useful for | |
4332 | implementing clean-up and check event sources that are | |
4333 | triggered by other work being done in the program. | |
4334 | ||
4335 | * systemd-networkd is no longer statically enabled, but uses | |
4336 | the usual [Install] sections so that it can be | |
4337 | enabled/disabled using systemctl. It still is enabled by | |
4338 | default however. | |
4339 | ||
b8bde116 | 4340 | * When creating a veth interface pair with systemd-nspawn, the |
43c71255 LP |
4341 | host side will now be prefixed with "vb-" if |
4342 | --network-bridge= is used, and with "ve-" if --network-veth | |
b8bde116 | 4343 | is used. This way, it is easy to distinguish these cases on |
43c71255 LP |
4344 | the host, for example to apply different configuration to |
4345 | them with systemd-networkd. | |
4346 | ||
d27893ef LP |
4347 | * The compatibility libraries for libsystemd-journal.so, |
4348 | libsystem-id128.so, libsystemd-login.so and | |
4349 | libsystemd-daemon.so do not make use of IFUNC | |
b8bde116 | 4350 | anymore. Instead, we now build libsystemd.so multiple times |
d27893ef LP |
4351 | under these alternative names. This means that the footprint |
4352 | is drastically increased, but given that these are | |
b8bde116 | 4353 | transitional compatibility libraries, this should not matter |
d27893ef LP |
4354 | much. This change has been made necessary to support the ARM |
4355 | platform for these compatibility libraries, as the ARM | |
d28315e4 | 4356 | toolchain is not really at the same level as the toolchain |
ed28905e | 4357 | for other architectures like x86 and does not support |
d27893ef LP |
4358 | IFUNC. Please make sure to use --enable-compat-libs only |
4359 | during a transitional period! | |
4360 | ||
13b28d82 | 4361 | Contributions from: Andreas Fuchs, Armin K., Colin Walters, |
43c71255 LP |
4362 | Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni, |
4363 | Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper | |
4364 | St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach, | |
4365 | Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike | |
4366 | Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe | |
4367 | Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog, | |
4368 | Zbigniew Jędrzejewski-Szmek | |
4369 | ||
ccddd104 | 4370 | — Berlin, 2014-02-24 |
43c71255 | 4371 | |
e49b5aad LP |
4372 | CHANGES WITH 209: |
4373 | ||
4374 | * A new component "systemd-networkd" has been added that can | |
4375 | be used to configure local network interfaces statically or | |
8b7d0494 JSJ |
4376 | via DHCP. It is capable of bringing up bridges, VLANs, and |
4377 | bonding. Currently, no hook-ups for interactive network | |
4670e9d5 | 4378 | configuration are provided. Use this for your initrd, |
8b7d0494 JSJ |
4379 | container, embedded, or server setup if you need a simple, |
4380 | yet powerful, network configuration solution. This | |
4670e9d5 | 4381 | configuration subsystem is quite nifty, as it allows wildcard |
1e190502 | 4382 | hotplug matching in interfaces. For example, with a single |
4670e9d5 | 4383 | configuration snippet, you can configure that all Ethernet |
1e190502 ZJS |
4384 | interfaces showing up are automatically added to a bridge, |
4385 | or similar. It supports link-sensing and more. | |
e49b5aad LP |
4386 | |
4387 | * A new tool "systemd-socket-proxyd" has been added which can | |
4c2413bf | 4388 | act as a bidirectional proxy for TCP sockets. This is |
e49b5aad LP |
4389 | useful for adding socket activation support to services that |
4390 | do not actually support socket activation, including virtual | |
4c2413bf | 4391 | machines and the like. |
e49b5aad LP |
4392 | |
4393 | * Add a new tool to save/restore rfkill state on | |
4394 | shutdown/boot. | |
4395 | ||
8b7d0494 JSJ |
4396 | * Save/restore state of keyboard backlights in addition to |
4397 | display backlights on shutdown/boot. | |
e49b5aad LP |
4398 | |
4399 | * udev learned a new SECLABEL{} construct to label device | |
4400 | nodes with a specific security label when they appear. For | |
4c2413bf | 4401 | now, only SECLABEL{selinux} is supported, but the syntax is |
e49b5aad LP |
4402 | prepared for additional security frameworks. |
4403 | ||
4404 | * udev gained a new scheme to configure link-level attributes | |
4405 | from files in /etc/systemd/network/*.link. These files can | |
8b7d0494 | 4406 | match against MAC address, device path, driver name and type, |
4c2413bf | 4407 | and will apply attributes like the naming policy, link speed, |
8b7d0494 | 4408 | MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC |
e49b5aad LP |
4409 | address assignment policy (randomized, ...). |
4410 | ||
dfb08b05 ZJS |
4411 | * The configuration of network interface naming rules for |
4412 | "permanent interface names" has changed: a new NamePolicy= | |
4413 | setting in the [Link] section of .link files determines the | |
a8eaaee7 | 4414 | priority of possible naming schemes (onboard, slot, MAC, |
dfb08b05 ZJS |
4415 | path). The default value of this setting is determined by |
4416 | /usr/lib/net/links/99-default.link. Old | |
4417 | 80-net-name-slot.rules udev configuration file has been | |
4418 | removed, so local configuration overriding this file should | |
ce830873 | 4419 | be adapted to override 99-default.link instead. |
dfb08b05 | 4420 | |
e49b5aad | 4421 | * When the User= switch is used in a unit file, also |
4c2413bf | 4422 | initialize $SHELL= based on the user database entry. |
e49b5aad LP |
4423 | |
4424 | * systemd no longer depends on libdbus. All communication is | |
4425 | now done with sd-bus, systemd's low-level bus library | |
4426 | implementation. | |
4427 | ||
4428 | * kdbus support has been added to PID 1 itself. When kdbus is | |
4c2413bf | 4429 | enabled, this causes PID 1 to set up the system bus and |
e49b5aad LP |
4430 | enable support for a new ".busname" unit type that |
4431 | encapsulates bus name activation on kdbus. It works a little | |
4432 | bit like ".socket" units, except for bus names. A new | |
4433 | generator has been added that converts classic dbus1 service | |
4434 | activation files automatically into native systemd .busname | |
4435 | and .service units. | |
4436 | ||
4437 | * sd-bus: add a light-weight vtable implementation that allows | |
4438 | defining objects on the bus with a simple static const | |
4439 | vtable array of its methods, signals and properties. | |
4440 | ||
8b7d0494 | 4441 | * systemd will not generate or install static dbus |
e49b5aad | 4442 | introspection data anymore to /usr/share/dbus-1/interfaces, |
1e190502 | 4443 | as the precise format of these files is unclear, and |
e49b5aad LP |
4444 | nothing makes use of it. |
4445 | ||
4446 | * A proxy daemon is now provided to proxy clients connecting | |
4447 | via classic D-Bus AF_UNIX sockets to kdbus, to provide full | |
4448 | compatibility with classic D-Bus. | |
4449 | ||
4450 | * A bus driver implementation has been added that supports the | |
4451 | classic D-Bus bus driver calls on kdbus, also for | |
4452 | compatibility purposes. | |
4453 | ||
4454 | * A new API "sd-event.h" has been added that implements a | |
4455 | minimal event loop API built around epoll. It provides a | |
4456 | couple of features that direct epoll usage is lacking: | |
b9761003 | 4457 | prioritization of events, scales to large numbers of timer |
e49b5aad LP |
4458 | events, per-event timer slack (accuracy), system-wide |
4459 | coalescing of timer events, exit handlers, watchdog | |
4460 | supervision support using systemd's sd_notify() API, child | |
4461 | process handling. | |
4462 | ||
4463 | * A new API "sd-rntl.h" has been added that provides an API | |
4464 | around the route netlink interface of the kernel, similar in | |
4465 | style to "sd-bus.h". | |
4466 | ||
7e95eda5 PF |
4467 | * A new API "sd-dhcp-client.h" has been added that provides a |
4468 | small DHCPv4 client-side implementation. This is used by | |
e49b5aad LP |
4469 | "systemd-networkd". |
4470 | ||
4c2413bf | 4471 | * There is a new kernel command line option |
8b7d0494 JSJ |
4472 | "systemd.restore_state=0|1". When set to "0", none of the |
4473 | systemd tools will restore saved runtime state to hardware | |
4474 | devices. More specifically, the rfkill and backlight states | |
4475 | are not restored. | |
e49b5aad LP |
4476 | |
4477 | * The FsckPassNo= compatibility option in mount/service units | |
4478 | has been removed. The fstab generator will now add the | |
4479 | necessary dependencies automatically, and does not require | |
4480 | PID1's support for that anymore. | |
4481 | ||
8b7d0494 | 4482 | * journalctl gained a new switch, --list-boots, that lists |
e49b5aad LP |
4483 | recent boots with their times and boot IDs. |
4484 | ||
4485 | * The various tools like systemctl, loginctl, timedatectl, | |
4486 | busctl, systemd-run, ... have gained a new switch "-M" to | |
4487 | connect to a specific, local OS container (as direct | |
4488 | connection, without requiring SSH). This works on any | |
4489 | container that is registered with machined, such as those | |
4490 | created by libvirt-lxc or nspawn. | |
4491 | ||
4492 | * systemd-run and systemd-analyze also gained support for "-H" | |
4c2413bf | 4493 | to connect to remote hosts via SSH. This is particularly |
8b7d0494 JSJ |
4494 | useful for systemd-run because it enables queuing of jobs |
4495 | onto remote systems. | |
e49b5aad LP |
4496 | |
4497 | * machinectl gained a new command "login" to open a getty | |
4498 | login in any local container. This works with any container | |
4499 | that is registered with machined (such as those created by | |
8e420494 | 4500 | libvirt-lxc or nspawn), and which runs systemd inside. |
e49b5aad LP |
4501 | |
4502 | * machinectl gained a new "reboot" command that may be used to | |
4503 | trigger a reboot on a specific container that is registered | |
4504 | with machined. This works on any container that runs an init | |
4505 | system of some kind. | |
4506 | ||
4507 | * systemctl gained a new "list-timers" command to print a nice | |
4508 | listing of installed timer units with the times they elapse | |
4509 | next. | |
4510 | ||
4511 | * Alternative reboot() parameters may now be specified on the | |
4512 | "systemctl reboot" command line and are passed to the | |
4513 | reboot() system call. | |
4514 | ||
4515 | * systemctl gained a new --job-mode= switch to configure the | |
4516 | mode to queue a job with. This is a more generic version of | |
8b7d0494 | 4517 | --fail, --irreversible, and --ignore-dependencies, which are |
e49b5aad LP |
4518 | still available but not advertised anymore. |
4519 | ||
e49b5aad LP |
4520 | * /etc/systemd/system.conf gained new settings to configure |
4521 | various default timeouts of units, as well as the default | |
b9761003 | 4522 | start limit interval and burst. These may still be overridden |
e49b5aad LP |
4523 | within each Unit. |
4524 | ||
270f1624 LP |
4525 | * PID1 will now export on the bus profile data of the security |
4526 | policy upload process (such as the SELinux policy upload to | |
8e420494 | 4527 | the kernel). |
e49b5aad | 4528 | |
4670e9d5 | 4529 | * journald: when forwarding logs to the console, include |
1e190502 ZJS |
4530 | timestamps (following the setting in |
4531 | /sys/module/printk/parameters/time). | |
e49b5aad LP |
4532 | |
4533 | * OnCalendar= in timer units now understands the special | |
4534 | strings "yearly" and "annually". (Both are equivalent) | |
4535 | ||
4536 | * The accuracy of timer units is now configurable with the new | |
4537 | AccuracySec= setting. It defaults to 1min. | |
4538 | ||
4539 | * A new dependency type JoinsNamespaceOf= has been added that | |
4540 | allows running two services within the same /tmp and network | |
4541 | namespace, if PrivateNetwork= or PrivateTmp= are used. | |
4542 | ||
4543 | * A new command "cat" has been added to systemctl. It outputs | |
4544 | the original unit file of a unit, and concatenates the | |
1e190502 ZJS |
4545 | contents of additional "drop-in" unit file snippets, so that |
4546 | the full configuration is shown. | |
e49b5aad LP |
4547 | |
4548 | * systemctl now supports globbing on the various "list-xyz" | |
4549 | commands, like "list-units" or "list-sockets", as well as on | |
1e190502 ZJS |
4550 | those commands which take multiple unit names. |
4551 | ||
4552 | * journalctl's --unit= switch gained support for globbing. | |
e49b5aad LP |
4553 | |
4554 | * All systemd daemons now make use of the watchdog logic so | |
4555 | that systemd automatically notices when they hang. | |
4556 | ||
4c2413bf | 4557 | * If the $container_ttys environment variable is set, |
e49b5aad LP |
4558 | getty-generator will automatically spawn a getty for each |
4559 | listed tty. This is useful for container managers to request | |
4560 | login gettys to be spawned on as many ttys as needed. | |
4561 | ||
4562 | * %h, %s, %U specifier support is not available anymore when | |
4563 | used in unit files for PID 1. This is because NSS calls are | |
4564 | not safe from PID 1. They stay available for --user | |
4565 | instances of systemd, and as special case for the root user. | |
4566 | ||
e49b5aad LP |
4567 | * loginctl gained a new "--no-legend" switch to turn off output |
4568 | of the legend text. | |
4569 | ||
4570 | * The "sd-login.h" API gained three new calls: | |
4571 | sd_session_is_remote(), sd_session_get_remote_user(), | |
4572 | sd_session_get_remote_host() to query information about | |
4573 | remote sessions. | |
4574 | ||
8e420494 LP |
4575 | * The udev hardware database now also carries vendor/product |
4576 | information of SDIO devices. | |
e49b5aad LP |
4577 | |
4578 | * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to | |
4579 | determine whether watchdog notifications are requested by | |
4580 | the system manager. | |
4581 | ||
1e190502 | 4582 | * Socket-activated per-connection services now include a |
e49b5aad LP |
4583 | short description of the connection parameters in the |
4584 | description. | |
4585 | ||
4c2413bf | 4586 | * tmpfiles gained a new "--boot" option. When this is not used, |
e49b5aad | 4587 | only lines where the command character is not suffixed with |
4670e9d5 | 4588 | "!" are executed. When this option is specified, those |
1e190502 ZJS |
4589 | options are executed too. This partitions tmpfiles |
4590 | directives into those that can be safely executed at any | |
4591 | time, and those which should be run only at boot (for | |
4592 | example, a line that creates /run/nologin). | |
e49b5aad | 4593 | |
c0c5af00 | 4594 | * A new API "sd-resolve.h" has been added which provides a simple |
4c2413bf | 4595 | asynchronous wrapper around glibc NSS host name resolution |
e49b5aad | 4596 | calls, such as getaddrinfo(). In contrast to glibc's |
4c2413bf JE |
4597 | getaddrinfo_a(), it does not use signals. In contrast to most |
4598 | other asynchronous name resolution libraries, this one does | |
4599 | not reimplement DNS, but reuses NSS, so that alternate | |
e49b5aad | 4600 | host name resolution systems continue to work, such as mDNS, |
8b7d0494 | 4601 | LDAP, etc. This API is based on libasyncns, but it has been |
e49b5aad LP |
4602 | cleaned up for inclusion in systemd. |
4603 | ||
6300b3ec LP |
4604 | * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h", |
4605 | "sd-daemon.h" are no longer found in individual libraries | |
4606 | libsystemd-journal.so, libsystemd-login.so, | |
4607 | libsystemd-id128.so, libsystemd-daemon.so. Instead, we have | |
8b7d0494 JSJ |
4608 | merged them into a single library, libsystemd.so, which |
4609 | provides all symbols. The reason for this is cyclic | |
e49b5aad | 4610 | dependencies, as these libraries tend to use each other's |
d28315e4 | 4611 | symbols. So far, we have managed to workaround that by linking |
6300b3ec LP |
4612 | a copy of a good part of our code into each of these |
4613 | libraries again and again, which, however, makes certain | |
4614 | things hard to do, like sharing static variables. Also, it | |
4615 | substantially increases footprint. With this change, there | |
4616 | is only one library for the basic APIs systemd | |
4617 | provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h", | |
4618 | "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this | |
4619 | library as well, however are subject to the --enable-kdbus | |
4620 | switch (see below). Note that "sd-dhcp-client.h" is not part | |
4621 | of this library (this is because it only consumes, never | |
4622 | provides, services of/to other APIs). To make the transition | |
8b7d0494 | 4623 | easy from the separate libraries to the unified one, we |
4c2413bf | 4624 | provide the --enable-compat-libs compile-time switch which |
e49b5aad LP |
4625 | will generate stub libraries that are compatible with the |
4626 | old ones but redirect all calls to the new one. | |
4627 | ||
8b7d0494 | 4628 | * All of the kdbus logic and the new APIs "sd-bus.h", |
e49b5aad | 4629 | "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h", |
8b7d0494 JSJ |
4630 | and "sd-utf8.h" are compile-time optional via the |
4631 | "--enable-kdbus" switch, and they are not compiled in by | |
4632 | default. To make use of kdbus, you have to explicitly enable | |
4c2413bf | 4633 | the switch. Note however, that neither the kernel nor the |
e49b5aad LP |
4634 | userspace API for all of this is considered stable yet. We |
4635 | want to maintain the freedom to still change the APIs for | |
4c2413bf | 4636 | now. By specifying this build-time switch, you acknowledge |
e49b5aad | 4637 | that you are aware of the instability of the current |
ad42cf73 KS |
4638 | APIs. |
4639 | ||
4640 | * Also, note that while kdbus is pretty much complete, | |
e49b5aad | 4641 | it lacks one thing: proper policy support. This means you |
8b7d0494 | 4642 | can build a fully working system with all features; however, |
4c2413bf JE |
4643 | it will be highly insecure. Policy support will be added in |
4644 | one of the next releases, at the same time that we will | |
4645 | declare the APIs stable. | |
e49b5aad | 4646 | |
81c7dd89 | 4647 | * When the kernel command line argument "kdbus" is specified, |
ad42cf73 | 4648 | systemd will automatically load the kdbus.ko kernel module. At |
8b7d0494 | 4649 | this stage of development, it is only useful for testing kdbus |
ad42cf73 | 4650 | and should not be used in production. Note: if "--enable-kdbus" |
8b7d0494 | 4651 | is specified, and the kdbus.ko kernel module is available, and |
ad42cf73 KS |
4652 | "kdbus" is added to the kernel command line, the entire system |
4653 | runs with kdbus instead of dbus-daemon, with the above mentioned | |
4654 | problem of missing the system policy enforcement. Also a future | |
4655 | version of kdbus.ko or a newer systemd will not be compatible with | |
4656 | each other, and will unlikely be able to boot the machine if only | |
4657 | one of them is updated. | |
4658 | ||
e49b5aad | 4659 | * systemctl gained a new "import-environment" command which |
4c2413bf | 4660 | uploads the caller's environment (or parts thereof) into the |
e49b5aad LP |
4661 | service manager so that it is inherited by services started |
4662 | by the manager. This is useful to upload variables like | |
4663 | $DISPLAY into the user service manager. | |
4664 | ||
4665 | * A new PrivateDevices= switch has been added to service units | |
4666 | which allows running a service with a namespaced /dev | |
4667 | directory that does not contain any device nodes for | |
4c2413bf | 4668 | physical devices. More specifically, it only includes devices |
8b7d0494 | 4669 | such as /dev/null, /dev/urandom, and /dev/zero which are API |
e49b5aad LP |
4670 | entry points. |
4671 | ||
4672 | * logind has been extended to support behaviour like VT | |
4673 | switching on seats that do not support a VT. This makes | |
4674 | multi-session available on seats that are not the first seat | |
4675 | (seat0), and on systems where kernel support for VTs has | |
8b7d0494 | 4676 | been disabled at compile-time. |
e49b5aad LP |
4677 | |
4678 | * If a process holds a delay lock for system sleep or shutdown | |
1e190502 | 4679 | and fails to release it in time, we will now log its |
e49b5aad LP |
4680 | identity. This makes it easier to identify processes that |
4681 | cause slow suspends or power-offs. | |
4682 | ||
1e190502 ZJS |
4683 | * When parsing /etc/crypttab, support for a new key-slot= |
4684 | option as supported by Debian is added. It allows indicating | |
4685 | which LUKS slot to use on disk, speeding up key loading. | |
e49b5aad | 4686 | |
1e190502 ZJS |
4687 | * The sd_journald_sendv() API call has been checked and |
4688 | officially declared to be async-signal-safe so that it may | |
4689 | be invoked from signal handlers for logging purposes. | |
e49b5aad LP |
4690 | |
4691 | * Boot-time status output is now enabled automatically after a | |
4692 | short timeout if boot does not progress, in order to give | |
8e420494 | 4693 | the user an indication what she or he is waiting for. |
1e190502 ZJS |
4694 | |
4695 | * The boot-time output has been improved to show how much time | |
4696 | remains until jobs expire. | |
e49b5aad LP |
4697 | |
4698 | * The KillMode= switch in service units gained a new possible | |
8b7d0494 | 4699 | value "mixed". If set, and the unit is shut down, then the |
e49b5aad | 4700 | initial SIGTERM signal is sent only to the main daemon |
8e420494 | 4701 | process, while the following SIGKILL signal is sent to |
e49b5aad LP |
4702 | all remaining processes of the service. |
4703 | ||
4c2413bf JE |
4704 | * When a scope unit is registered, a new property "Controller" |
4705 | may be set. If set to a valid bus name, systemd will send a | |
e49b5aad LP |
4706 | RequestStop() signal to this name when it would like to shut |
4707 | down the scope. This may be used to hook manager logic into | |
4708 | the shutdown logic of scope units. Also, scope units may now | |
8b7d0494 | 4709 | be put in a special "abandoned" state, in which case the |
e49b5aad LP |
4710 | manager process which created them takes no further |
4711 | responsibilities for it. | |
4712 | ||
1e190502 | 4713 | * When reading unit files, systemd will now verify |
e49b5aad LP |
4714 | the access mode of these files, and warn about certain |
4715 | suspicious combinations. This has been added to make it | |
4716 | easier to track down packaging bugs where unit files are | |
4717 | marked executable or world-writable. | |
4718 | ||
4719 | * systemd-nspawn gained a new "--setenv=" switch to set | |
8b7d0494 | 4720 | container-wide environment variables. The similar option in |
1e190502 ZJS |
4721 | systemd-activate was renamed from "--environment=" to |
4722 | "--setenv=" for consistency. | |
e49b5aad LP |
4723 | |
4724 | * systemd-nspawn has been updated to create a new kdbus domain | |
4725 | for each container that is invoked, thus allowing each | |
b9761003 | 4726 | container to have its own set of system and user buses, |
8b7d0494 | 4727 | independent of the host. |
e49b5aad LP |
4728 | |
4729 | * systemd-nspawn gained a new --drop-capability= switch to run | |
4730 | the container with less capabilities than the default. Both | |
b9761003 | 4731 | --drop-capability= and --capability= now take the special |
e49b5aad LP |
4732 | string "all" for dropping or keeping all capabilities. |
4733 | ||
4734 | * systemd-nspawn gained new switches for executing containers | |
4735 | with specific SELinux labels set. | |
4736 | ||
4737 | * systemd-nspawn gained a new --quiet switch to not generate | |
4738 | any additional output but the container's own console | |
4739 | output. | |
4740 | ||
4741 | * systemd-nspawn gained a new --share-system switch to run a | |
4742 | container without PID namespacing enabled. | |
4743 | ||
4744 | * systemd-nspawn gained a new --register= switch to control | |
1e190502 | 4745 | whether the container is registered with systemd-machined or |
8e420494 | 4746 | not. This is useful for containers that do not run full |
e49b5aad LP |
4747 | OS images, but only specific apps. |
4748 | ||
4749 | * systemd-nspawn gained a new --keep-unit which may be used | |
8b7d0494 | 4750 | when invoked as the only program from a service unit, and |
e49b5aad | 4751 | results in registration of the unit service itself in |
1e190502 | 4752 | systemd-machined, instead of a newly opened scope unit. |
e49b5aad LP |
4753 | |
4754 | * systemd-nspawn gained a new --network-interface= switch for | |
4755 | moving arbitrary interfaces to the container. The new | |
4c2413bf | 4756 | --network-veth switch creates a virtual Ethernet connection |
8b7d0494 JSJ |
4757 | between host and container. The new --network-bridge= |
4758 | switch then allows assigning the host side of this virtual | |
4759 | Ethernet connection to a bridge device. | |
e49b5aad | 4760 | |
6afc95b7 LP |
4761 | * systemd-nspawn gained a new --personality= switch for |
4762 | setting the kernel personality for the container. This is | |
70a44afe | 4763 | useful when running a 32-bit container on a 64-bit host. A |
b8bde116 JE |
4764 | similar option Personality= is now also available for service |
4765 | units to use. | |
6afc95b7 | 4766 | |
e49b5aad LP |
4767 | * logind will now also track a "Desktop" identifier for each |
4768 | session which encodes the desktop environment of it. This is | |
4769 | useful for desktop environments that want to identify | |
4770 | multiple running sessions of itself easily. | |
4771 | ||
4772 | * A new SELinuxContext= setting for service units has been | |
4773 | added that allows setting a specific SELinux execution | |
4774 | context for a service. | |
4775 | ||
4776 | * Most systemd client tools will now honour $SYSTEMD_LESS for | |
4777 | settings of the "less" pager. By default, these tools will | |
8b7d0494 JSJ |
4778 | override $LESS to allow certain operations to work, such as |
4779 | jump-to-the-end. With $SYSTEMD_LESS, it is possible to | |
e49b5aad LP |
4780 | influence this logic. |
4781 | ||
4782 | * systemd's "seccomp" hook-up has been changed to make use of | |
4783 | the libseccomp library instead of using its own | |
4784 | implementation. This has benefits for portability among | |
4785 | other things. | |
4786 | ||
4c2413bf | 4787 | * For usage together with SystemCallFilter=, a new |
8b7d0494 | 4788 | SystemCallErrorNumber= setting has been introduced that |
b8bde116 JE |
4789 | allows configuration of a system error number to be returned |
4790 | on filtered system calls, instead of immediately killing the | |
e49b5aad LP |
4791 | process. Also, SystemCallArchitectures= has been added to |
4792 | limit access to system calls of a particular architecture | |
4793 | (in order to turn off support for unused secondary | |
4c2413bf | 4794 | architectures). There is also a global |
8b7d0494 | 4795 | SystemCallArchitectures= setting in system.conf now to turn |
e49b5aad LP |
4796 | off support for non-native system calls system-wide. |
4797 | ||
210054d7 KS |
4798 | * systemd requires a kernel with a working name_to_handle_at(), |
4799 | please see the kernel config requirements in the README file. | |
4800 | ||
e49b5aad LP |
4801 | Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov, |
4802 | Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera, | |
4803 | Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters, | |
4804 | Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J | |
4805 | Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa, | |
4806 | David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov, | |
4807 | Elia Pinto, Florian Weimer, George McCollister, Goffredo | |
4808 | Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor | |
4809 | Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld, | |
4810 | Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose | |
4811 | Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg, | |
4812 | Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz | |
4813 | Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas, | |
4814 | Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de | |
4815 | Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael | |
4816 | Marineau, Michael Scherer, Michał Górny, Michal Sekletar, | |
4817 | Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt, | |
4818 | Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien | |
4819 | Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters, | |
4820 | Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else, | |
4821 | Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen, | |
4822 | Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav | |
4823 | Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang | |
4824 | Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek | |
4825 | ||
ccddd104 | 4826 | — Berlin, 2014-02-20 |
e49b5aad | 4827 | |
cd4010b3 LP |
4828 | CHANGES WITH 208: |
4829 | ||
4830 | * logind has gained support for facilitating privileged input | |
4831 | and drm device access for unprivileged clients. This work is | |
4832 | useful to allow Wayland display servers (and similar | |
4833 | programs, such as kmscon) to run under the user's ID and | |
4834 | access input and drm devices which are normally | |
4835 | protected. When this is used (and the kernel is new enough) | |
4836 | logind will "mute" IO on the file descriptors passed to | |
4837 | Wayland as long as it is in the background and "unmute" it | |
4838 | if it returns into the foreground. This allows secure | |
4839 | session switching without allowing background sessions to | |
4840 | eavesdrop on input and display data. This also introduces | |
4841 | session switching support if VT support is turned off in the | |
4842 | kernel, and on seats that are not seat0. | |
4843 | ||
4844 | * A new kernel command line option luks.options= is understood | |
06b643e7 | 4845 | now which allows specifying LUKS options for usage for LUKS |
cd4010b3 LP |
4846 | encrypted partitions specified with luks.uuid=. |
4847 | ||
4848 | * tmpfiles.d(5) snippets may now use specifier expansion in | |
4849 | path names. More specifically %m, %b, %H, %v, are now | |
4850 | replaced by the local machine id, boot id, hostname, and | |
4851 | kernel version number. | |
4852 | ||
4853 | * A new tmpfiles.d(5) command "m" has been introduced which | |
4854 | may be used to change the owner/group/access mode of a file | |
d28315e4 | 4855 | or directory if it exists, but do nothing if it does not. |
cd4010b3 LP |
4856 | |
4857 | * This release removes high-level support for the | |
4858 | MemorySoftLimit= cgroup setting. The underlying kernel | |
4859 | cgroup attribute memory.soft_limit= is currently badly | |
4860 | designed and likely to be removed from the kernel API in its | |
d28315e4 | 4861 | current form, hence we should not expose it for now. |
cd4010b3 LP |
4862 | |
4863 | * The memory.use_hierarchy cgroup attribute is now enabled for | |
4864 | all cgroups systemd creates in the memory cgroup | |
4865 | hierarchy. This option is likely to be come the built-in | |
cc98b302 TH |
4866 | default in the kernel anyway, and the non-hierarchical mode |
4867 | never made much sense in the intrinsically hierarchical | |
cd4010b3 LP |
4868 | cgroup system. |
4869 | ||
4870 | * A new field _SYSTEMD_SLICE= is logged along with all journal | |
4871 | messages containing the slice a message was generated | |
4872 | from. This is useful to allow easy per-customer filtering of | |
4873 | logs among other things. | |
4874 | ||
4875 | * systemd-journald will no longer adjust the group of journal | |
4876 | files it creates to the "systemd-journal" group. Instead we | |
4877 | rely on the journal directory to be owned by the | |
4878 | "systemd-journal" group, and its setgid bit set, so that the | |
4879 | kernel file system layer will automatically enforce that | |
4880 | journal files inherit this group assignment. The reason for | |
4881 | this change is that we cannot allow NSS look-ups from | |
4882 | journald which would be necessary to resolve | |
4883 | "systemd-journal" to a numeric GID, because this might | |
4884 | create deadlocks if NSS involves synchronous queries to | |
4885 | other daemons (such as nscd, or sssd) which in turn are | |
4886 | logging clients of journald and might block on it, which | |
4887 | would then dead lock. A tmpfiles.d(5) snippet included in | |
4888 | systemd will make sure the setgid bit and group are | |
4889 | properly set on the journal directory if it exists on every | |
4890 | boot. However, we recommend adjusting it manually after | |
4891 | upgrades too (or from RPM scriptlets), so that the change is | |
4892 | not delayed until next reboot. | |
4893 | ||
4894 | * Backlight and random seed files in /var/lib/ have moved into | |
4895 | the /var/lib/systemd/ directory, in order to centralize all | |
4896 | systemd generated files in one directory. | |
4897 | ||
4898 | * Boot time performance measurements (as displayed by | |
4899 | "systemd-analyze" for example) will now read ACPI 5.0 FPDT | |
4900 | performance information if that's available to determine how | |
4901 | much time BIOS and boot loader initialization required. With | |
4902 | a sufficiently new BIOS you hence no longer need to boot | |
4903 | with Gummiboot to get access to such information. | |
4904 | ||
4905 | Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters, | |
4906 | Cristian Rodríguez, Dave Reisner, David Herrmann, David | |
4907 | Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao | |
4908 | feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart | |
4909 | Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt, | |
4910 | Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty, | |
4911 | Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek | |
4912 | ||
ccddd104 | 4913 | — Berlin, 2013-10-02 |
cd4010b3 | 4914 | |
4f0be680 LP |
4915 | CHANGES WITH 207: |
4916 | ||
4917 | * The Restart= option for services now understands a new | |
f3a165b0 | 4918 | on-watchdog setting, which will restart the service |
4f0be680 LP |
4919 | automatically if the service stops sending out watchdog keep |
4920 | alive messages (as configured with WatchdogSec=). | |
4921 | ||
4922 | * The getty generator (which is responsible for bringing up a | |
4923 | getty on configured serial consoles) will no longer only | |
4924 | start a getty on the primary kernel console but on all | |
4925 | others, too. This makes the order in which console= is | |
4926 | specified on the kernel command line less important. | |
4927 | ||
4928 | * libsystemd-logind gained a new sd_session_get_vt() call to | |
4929 | retrieve the VT number of a session. | |
4930 | ||
4931 | * If the option "tries=0" is set for an entry of /etc/crypttab | |
4932 | its passphrase is queried indefinitely instead of any | |
4933 | maximum number of tries. | |
4934 | ||
4935 | * If a service with a configure PID file terminates its PID | |
4936 | file will now be removed automatically if it still exists | |
4937 | afterwards. This should put an end to stale PID files. | |
4938 | ||
4939 | * systemd-run will now also take relative binary path names | |
4940 | for execution and no longer insists on absolute paths. | |
4941 | ||
4942 | * InaccessibleDirectories= and ReadOnlyDirectories= now take | |
4943 | paths that are optionally prefixed with "-" to indicate that | |
d28315e4 | 4944 | it should not be considered a failure if they do not exist. |
4f0be680 | 4945 | |
f3a165b0 KS |
4946 | * journalctl -o (and similar commands) now understands a new |
4947 | output mode "short-precise", it is similar to "short" but | |
4f0be680 LP |
4948 | shows timestamps with usec accuracy. |
4949 | ||
4950 | * The option "discard" (as known from Debian) is now | |
4951 | synonymous to "allow-discards" in /etc/crypttab. In fact, | |
387abf80 | 4952 | "discard" is preferred now (since it is easier to remember |
4f0be680 LP |
4953 | and type). |
4954 | ||
f3a165b0 | 4955 | * Some licensing clean-ups were made, so that more code is now |
4f0be680 LP |
4956 | LGPL-2.1 licensed than before. |
4957 | ||
4958 | * A minimal tool to save/restore the display backlight | |
4959 | brightness across reboots has been added. It will store the | |
f3a165b0 | 4960 | backlight setting as late as possible at shutdown, and |
4f0be680 LP |
4961 | restore it as early as possible during reboot. |
4962 | ||
4963 | * A logic to automatically discover and enable home and swap | |
4964 | partitions on GPT disks has been added. With this in place | |
4965 | /etc/fstab becomes optional for many setups as systemd can | |
4966 | discover certain partitions located on the root disk | |
4967 | automatically. Home partitions are recognized under their | |
4968 | GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap | |
4969 | partitions are recognized under their GPT type ID | |
4970 | 0657fd6da4ab43c484e50933c84b4f4f. | |
4971 | ||
4972 | * systemd will no longer pass any environment from the kernel | |
4973 | or initrd to system services. If you want to set an | |
4974 | environment for all services, do so via the kernel command | |
4975 | line systemd.setenv= assignment. | |
4976 | ||
387abf80 LP |
4977 | * The systemd-sysctl tool no longer natively reads the file |
4978 | /etc/sysctl.conf. If desired, the file should be symlinked | |
4979 | from /etc/sysctl.d/99-sysctl.conf. Apart from providing | |
4980 | legacy support by a symlink rather than built-in code, it | |
4981 | also makes the otherwise hidden order of application of the | |
4982 | different files visible. (Note that this partly reverts to a | |
4983 | pre-198 application order of sysctl knobs!) | |
04bf3c1a | 4984 | |
4f0be680 LP |
4985 | * The "systemctl set-log-level" and "systemctl dump" commands |
4986 | have been moved to systemd-analyze. | |
4987 | ||
4988 | * systemd-run learned the new --remain-after-exit switch, | |
4989 | which causes the scope unit not to be cleaned up | |
4990 | automatically after the process terminated. | |
4991 | ||
4992 | * tmpfiles learned a new --exclude-prefix= switch to exclude | |
4993 | certain paths from operation. | |
4994 | ||
4995 | * journald will now automatically flush all messages to disk | |
f47ad593 ZJS |
4996 | as soon as a message at the log level CRIT, ALERT or EMERG |
4997 | is received. | |
4f0be680 LP |
4998 | |
4999 | Contributions from: Andrew Cook, Brandon Philips, Christian | |
5000 | Hesse, Christoph Junghans, Colin Walters, Daniel Schaal, | |
5001 | Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George | |
5002 | McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer, | |
5003 | Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt, | |
5004 | Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering, | |
5005 | Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel | |
5006 | Holtmann, Martin Pitt, Michael Biebl, Michael Marineau, | |
5007 | Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał | |
5008 | Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn | |
5009 | Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe | |
5010 | Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao, | |
5011 | William Giokas, Zbigniew Jędrzejewski-Szmek | |
5012 | ||
ccddd104 | 5013 | — Berlin, 2013-09-13 |
4f0be680 | 5014 | |
408f281b LP |
5015 | CHANGES WITH 206: |
5016 | ||
5017 | * The documentation has been updated to cover the various new | |
5018 | concepts introduced with 205. | |
5019 | ||
5020 | * Unit files now understand the new %v specifier which | |
5021 | resolves to the kernel version string as returned by "uname | |
5022 | -r". | |
5023 | ||
5024 | * systemctl now supports filtering the unit list output by | |
5025 | load state, active state and sub state, using the new | |
33b521be | 5026 | --state= parameter. |
408f281b LP |
5027 | |
5028 | * "systemctl status" will now show the results of the | |
5029 | condition checks (like ConditionPathExists= and similar) of | |
5030 | the last start attempts of the unit. They are also logged to | |
5031 | the journal. | |
5032 | ||
5033 | * "journalctl -b" may now be used to look for boot output of a | |
5034 | specific boot. Try "journalctl -b -1" for the previous boot, | |
5035 | but the syntax is substantially more powerful. | |
5036 | ||
5037 | * "journalctl --show-cursor" has been added which prints the | |
5038 | cursor string the last shown log line. This may then be used | |
5039 | with the new "journalctl --after-cursor=" switch to continue | |
5040 | browsing logs from that point on. | |
5041 | ||
5042 | * "journalctl --force" may now be used to force regeneration | |
5043 | of an FSS key. | |
5044 | ||
251cc819 LP |
5045 | * Creation of "dead" device nodes has been moved from udev |
5046 | into kmod and tmpfiles. Previously, udev would read the kmod | |
5047 | databases to pre-generate dead device nodes based on meta | |
5048 | information contained in kernel modules, so that these would | |
5049 | be auto-loaded on access rather then at boot. As this | |
d28315e4 | 5050 | does not really have much to do with the exposing actual |
251cc819 LP |
5051 | kernel devices to userspace this has always been slightly |
5052 | alien in the udev codebase. Following the new scheme kmod | |
5053 | will now generate a runtime snippet for tmpfiles from the | |
5054 | module meta information and it now is tmpfiles' job to the | |
5055 | create the nodes. This also allows overriding access and | |
5056 | other parameters for the nodes using the usual tmpfiles | |
5057 | facilities. As side effect this allows us to remove the | |
5058 | CAP_SYS_MKNOD capability bit from udevd entirely. | |
5059 | ||
5060 | * logind's device ACLs may now be applied to these "dead" | |
5061 | devices nodes too, thus finally allowing managed access to | |
ce830873 | 5062 | devices such as /dev/snd/sequencer without loading the |
251cc819 | 5063 | backing module right-away. |
408f281b LP |
5064 | |
5065 | * A new RPM macro has been added that may be used to apply | |
5066 | tmpfiles configuration during package installation. | |
5067 | ||
5068 | * systemd-detect-virt and ConditionVirtualization= now can | |
5069 | detect User-Mode-Linux machines (UML). | |
5070 | ||
251cc819 LP |
5071 | * journald will now implicitly log the effective capabilities |
5072 | set of processes in the message metadata. | |
408f281b LP |
5073 | |
5074 | * systemd-cryptsetup has gained support for TrueCrypt volumes. | |
5075 | ||
5076 | * The initrd interface has been simplified (more specifically, | |
5077 | support for passing performance data via environment | |
5078 | variables and fsck results via files in /run has been | |
5079 | removed). These features were non-essential, and are | |
5080 | nowadays available in a much nicer way by having systemd in | |
5081 | the initrd serialize its state and have the hosts systemd | |
5082 | deserialize it again. | |
5083 | ||
28f5c779 KS |
5084 | * The udev "keymap" data files and tools to apply keyboard |
5085 | specific mappings of scan to key codes, and force-release | |
5086 | scan code lists have been entirely replaced by a udev | |
5087 | "keyboard" builtin and a hwdb data file. | |
408f281b | 5088 | |
251cc819 LP |
5089 | * systemd will now honour the kernel's "quiet" command line |
5090 | argument also during late shutdown, resulting in a | |
5091 | completely silent shutdown when used. | |
5092 | ||
5093 | * There's now an option to control the SO_REUSEPORT socket | |
5094 | option in .socket units. | |
5095 | ||
5096 | * Instance units will now automatically get a per-template | |
5097 | subslice of system.slice unless something else is explicitly | |
5098 | configured. For example, instances of sshd@.service will now | |
5099 | implicitly be placed in system-sshd.slice rather than | |
5100 | system.slice as before. | |
5101 | ||
5102 | * Test coverage support may now be enabled at build time. | |
5103 | ||
5104 | Contributions from: Dave Reisner, Frederic Crozat, Harald | |
5105 | Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan | |
5106 | Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart | |
5107 | Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael | |
5108 | Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden, | |
5109 | Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William | |
5110 | Giokas, Zbigniew Jędrzejewski-Szmek | |
5111 | ||
ccddd104 | 5112 | — Berlin, 2013-07-23 |
4f0be680 | 5113 | |
00aa832b LP |
5114 | CHANGES WITH 205: |
5115 | ||
5116 | * Two new unit types have been introduced: | |
5117 | ||
5118 | Scope units are very similar to service units, however, are | |
ccddd104 | 5119 | created out of pre-existing processes — instead of PID 1 |
00aa832b LP |
5120 | forking off the processes. By using scope units it is |
5121 | possible for system services and applications to group their | |
5122 | own child processes (worker processes) in a powerful way | |
5123 | which then maybe used to organize them, or kill them | |
5124 | together, or apply resource limits on them. | |
5125 | ||
5126 | Slice units may be used to partition system resources in an | |
cc98b302 | 5127 | hierarchical fashion and then assign other units to them. By |
00aa832b LP |
5128 | default there are now three slices: system.slice (for all |
5129 | system services), user.slice (for all user sessions), | |
5130 | machine.slice (for VMs and containers). | |
5131 | ||
5132 | Slices and scopes have been introduced primarily in | |
5133 | context of the work to move cgroup handling to a | |
5134 | single-writer scheme, where only PID 1 | |
5135 | creates/removes/manages cgroups. | |
5136 | ||
5137 | * There's a new concept of "transient" units. In contrast to | |
5138 | normal units these units are created via an API at runtime, | |
5139 | not from configuration from disk. More specifically this | |
5140 | means it is now possible to run arbitrary programs as | |
5141 | independent services, with all execution parameters passed | |
5142 | in via bus APIs rather than read from disk. Transient units | |
5143 | make systemd substantially more dynamic then it ever was, | |
5144 | and useful as a general batch manager. | |
5145 | ||
5146 | * logind has been updated to make use of scope and slice units | |
5147 | for managing user sessions. As a user logs in he will get | |
5148 | his own private slice unit, to which all sessions are added | |
5149 | as scope units. We also added support for automatically | |
5150 | adding an instance of user@.service for the user into the | |
5151 | slice. Effectively logind will no longer create cgroup | |
5152 | hierarchies on its own now, it will defer entirely to PID 1 | |
5153 | for this by means of scope, service and slice units. Since | |
5154 | user sessions this way become entities managed by PID 1 | |
5155 | the output of "systemctl" is now a lot more comprehensive. | |
5156 | ||
5157 | * A new mini-daemon "systemd-machined" has been added which | |
5158 | may be used by virtualization managers to register local | |
5159 | VMs/containers. nspawn has been updated accordingly, and | |
5160 | libvirt will be updated shortly. machined will collect a bit | |
5161 | of meta information about the VMs/containers, and assign | |
5162 | them their own scope unit (see above). The collected | |
5163 | meta-data is then made available via the "machinectl" tool, | |
5164 | and exposed in "ps" and similar tools. machined/machinectl | |
5165 | is compile-time optional. | |
5166 | ||
5167 | * As discussed earlier, the low-level cgroup configuration | |
5168 | options ControlGroup=, ControlGroupModify=, | |
5169 | ControlGroupPersistent=, ControlGroupAttribute= have been | |
5170 | removed. Please use high-level attribute settings instead as | |
5171 | well as slice units. | |
5172 | ||
5173 | * A new bus call SetUnitProperties() has been added to alter | |
5174 | various runtime parameters of a unit. This is primarily | |
5175 | useful to alter cgroup parameters dynamically in a nice way, | |
5176 | but will be extended later on to make more properties | |
5177 | modifiable at runtime. systemctl gained a new set-properties | |
5178 | command that wraps this call. | |
5179 | ||
5180 | * A new tool "systemd-run" has been added which can be used to | |
5181 | run arbitrary command lines as transient services or scopes, | |
5182 | while configuring a number of settings via the command | |
5183 | line. This tool is currently very basic, however already | |
5184 | very useful. We plan to extend this tool to even allow | |
5185 | queuing of execution jobs with time triggers from the | |
5186 | command line, similar in fashion to "at". | |
5187 | ||
5188 | * nspawn will now inform the user explicitly that kernels with | |
5189 | audit enabled break containers, and suggest the user to turn | |
5190 | off audit. | |
5191 | ||
5192 | * Support for detecting the IMA and AppArmor security | |
5193 | frameworks with ConditionSecurity= has been added. | |
5194 | ||
5195 | * journalctl gained a new "-k" switch for showing only kernel | |
1fda0ab5 ZJS |
5196 | messages, mimicking dmesg output; in addition to "--user" |
5197 | and "--system" switches for showing only user's own logs | |
5198 | and system logs. | |
00aa832b LP |
5199 | |
5200 | * systemd-delta can now show information about drop-in | |
5201 | snippets extending unit files. | |
5202 | ||
5203 | * libsystemd-bus has been substantially updated but is still | |
5204 | not available as public API. | |
5205 | ||
5206 | * systemd will now look for the "debug" argument on the kernel | |
499b604b | 5207 | command line and enable debug logging, similar to what |
00aa832b LP |
5208 | "systemd.log_level=debug" already did before. |
5209 | ||
5210 | * "systemctl set-default", "systemctl get-default" has been | |
5211 | added to configure the default.target symlink, which | |
5212 | controls what to boot into by default. | |
5213 | ||
1fda0ab5 ZJS |
5214 | * "systemctl set-log-level" has been added as a convenient |
5215 | way to raise and lower systemd logging threshold. | |
5216 | ||
00aa832b LP |
5217 | * "systemd-analyze plot" will now show the time the various |
5218 | generators needed for execution, as well as information | |
5219 | about the unit file loading. | |
5220 | ||
00aa832b LP |
5221 | * libsystemd-journal gained a new sd_journal_open_files() call |
5222 | for opening specific journal files. journactl also gained a | |
5223 | new switch to expose this new functionality. Previously we | |
5224 | only supported opening all files from a directory, or all | |
5225 | files from the system, as opening individual files only is | |
5226 | racy due to journal file rotation. | |
5227 | ||
5228 | * systemd gained the new DefaultEnvironment= setting in | |
5229 | /etc/systemd/system.conf to set environment variables for | |
5230 | all services. | |
5231 | ||
5232 | * If a privileged process logs a journal message with the | |
5233 | OBJECT_PID= field set, then journald will automatically | |
5234 | augment this with additional OBJECT_UID=, OBJECT_GID=, | |
5235 | OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if | |
5236 | system services want to log events about specific client | |
5237 | processes. journactl/systemctl has been updated to make use | |
5238 | of this information if all log messages regarding a specific | |
5239 | unit is requested. | |
5240 | ||
5241 | Contributions from: Auke Kok, Chengwei Yang, Colin Walters, | |
5242 | Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave | |
5243 | Reisner, David Coppa, David King, David Strauss, Eelco | |
5244 | Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander | |
5245 | Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan | |
5246 | Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart | |
5247 | Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer, | |
5248 | Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer, | |
5249 | Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan, | |
5250 | Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern, | |
5251 | Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar, | |
5252 | Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek, | |
5253 | Łukasz Stelmach, 장동준 | |
5254 | ||
606c24e3 LP |
5255 | CHANGES WITH 204: |
5256 | ||
5257 | * The Python bindings gained some minimal support for the APIs | |
5258 | exposed by libsystemd-logind. | |
5259 | ||
5260 | * ConditionSecurity= gained support for detecting SMACK. Since | |
5261 | this condition already supports SELinux and AppArmor we only | |
5262 | miss IMA for this. Patches welcome! | |
5263 | ||
5264 | Contributions from: Karol Lewandowski, Lennart Poettering, | |
5265 | Zbigniew Jędrzejewski-Szmek | |
5266 | ||
2f3fcf85 LP |
5267 | CHANGES WITH 203: |
5268 | ||
5269 | * systemd-nspawn will now create /etc/resolv.conf if | |
5270 | necessary, before bind-mounting the host's file onto it. | |
5271 | ||
5272 | * systemd-nspawn will now store meta information about a | |
5273 | container on the container's cgroup as extended attribute | |
5274 | fields, including the root directory. | |
5275 | ||
5276 | * The cgroup hierarchy has been reworked in many ways. All | |
5277 | objects any of the components systemd creates in the cgroup | |
b82eed9a | 5278 | tree are now suffixed. More specifically, user sessions are |
2f3fcf85 LP |
5279 | now placed in cgroups suffixed with ".session", users in |
5280 | cgroups suffixed with ".user", and nspawn containers in | |
5281 | cgroups suffixed with ".nspawn". Furthermore, all cgroup | |
5282 | names are now escaped in a simple scheme to avoid collision | |
5283 | of userspace object names with kernel filenames. This work | |
5284 | is preparation for making these objects relocatable in the | |
5285 | cgroup tree, in order to allow easy resource partitioning of | |
5286 | these objects without causing naming conflicts. | |
5287 | ||
5288 | * systemctl list-dependencies gained the new switches | |
5289 | --plain, --reverse, --after and --before. | |
5290 | ||
5291 | * systemd-inhibit now shows the process name of processes that | |
5292 | have taken an inhibitor lock. | |
5293 | ||
5294 | * nss-myhostname will now also resolve "localhost" | |
5295 | implicitly. This makes /etc/hosts an optional file and | |
5296 | nicely handles that on IPv6 ::1 maps to both "localhost" and | |
5297 | the local hostname. | |
5298 | ||
5299 | * libsystemd-logind.so gained a new call | |
5300 | sd_get_machine_names() to enumerate running containers and | |
5301 | VMs (currently only supported by very new libvirt and | |
5302 | nspawn). sd_login_monitor can now be used to watch | |
5303 | VMs/containers coming and going. | |
5304 | ||
5305 | * .include is not allowed recursively anymore, and only in | |
5306 | unit files. Usually it is better to use drop-in snippets in | |
5307 | .d/*.conf anyway, as introduced with systemd 198. | |
5308 | ||
5309 | * systemd-analyze gained a new "critical-chain" command that | |
5310 | determines the slowest chain of units run during system | |
5311 | boot-up. It is very useful for tracking down where | |
5312 | optimizing boot time is the most beneficial. | |
5313 | ||
5314 | * systemd will no longer allow manipulating service paths in | |
5315 | the name=systemd:/system cgroup tree using ControlGroup= in | |
5316 | units. (But is still fine with it in all other dirs.) | |
5317 | ||
5318 | * There's a new systemd-nspawn@.service service file that may | |
5319 | be used to easily run nspawn containers as system | |
5320 | services. With the container's root directory in | |
5321 | /var/lib/container/foobar it is now sufficient to run | |
5322 | "systemctl start systemd-nspawn@foobar.service" to boot it. | |
5323 | ||
5324 | * systemd-cgls gained a new parameter "--machine" to list only | |
5325 | the processes within a certain container. | |
5326 | ||
5327 | * ConditionSecurity= now can check for "apparmor". We still | |
5328 | are lacking checks for SMACK and IMA for this condition | |
5329 | check though. Patches welcome! | |
5330 | ||
5331 | * A new configuration file /etc/systemd/sleep.conf has been | |
5332 | added that may be used to configure which kernel operation | |
5333 | systemd is supposed to execute when "suspend", "hibernate" | |
5334 | or "hybrid-sleep" is requested. This makes the new kernel | |
5335 | "freeze" state accessible to the user. | |
5336 | ||
5337 | * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape | |
5338 | the passed argument if applicable. | |
5339 | ||
5340 | Contributions from: Auke Kok, Colin Guthrie, Colin Walters, | |
5341 | Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner, | |
5342 | Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh | |
5343 | Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn, | |
5344 | MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel | |
5345 | Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom | |
5346 | Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew | |
5347 | Jędrzejewski-Szmek | |
5348 | ||
ef3b5246 LP |
5349 | CHANGES WITH 202: |
5350 | ||
5351 | * The output of 'systemctl list-jobs' got some polishing. The | |
5352 | '--type=' argument may now be passed more than once. A new | |
5353 | command 'systemctl list-sockets' has been added which shows | |
5354 | a list of kernel sockets systemd is listening on with the | |
5355 | socket units they belong to, plus the units these socket | |
5356 | units activate. | |
5357 | ||
5358 | * The experimental libsystemd-bus library got substantial | |
5359 | updates to work in conjunction with the (also experimental) | |
5360 | kdbus kernel project. It works well enough to exchange | |
5361 | messages with some sophistication. Note that kdbus is not | |
5362 | ready yet, and the library is mostly an elaborate test case | |
5363 | for now, and not installable. | |
5364 | ||
5365 | * systemd gained a new unit 'systemd-static-nodes.service' | |
5366 | that generates static device nodes earlier during boot, and | |
5367 | can run in conjunction with udev. | |
5368 | ||
5369 | * libsystemd-login gained a new call sd_pid_get_user_unit() | |
5370 | to retrieve the user systemd unit a process is running | |
5371 | in. This is useful for systems where systemd is used as | |
5372 | session manager. | |
5373 | ||
5374 | * systemd-nspawn now places all containers in the new /machine | |
5375 | top-level cgroup directory in the name=systemd | |
5376 | hierarchy. libvirt will soon do the same, so that we get a | |
5377 | uniform separation of /system, /user and /machine for system | |
5378 | services, user processes and containers/virtual | |
5379 | machines. This new cgroup hierarchy is also useful to stick | |
5380 | stable names to specific container instances, which can be | |
7c04ad2d | 5381 | recognized later this way (this name may be controlled |
ef3b5246 LP |
5382 | via systemd-nspawn's new -M switch). libsystemd-login also |
5383 | gained a new call sd_pid_get_machine_name() to retrieve the | |
5384 | name of the container/VM a specific process belongs to. | |
5385 | ||
5386 | * bootchart can now store its data in the journal. | |
5387 | ||
5388 | * libsystemd-journal gained a new call | |
5389 | sd_journal_add_conjunction() for AND expressions to the | |
5390 | matching logic. This can be used to express more complex | |
5391 | logical expressions. | |
5392 | ||
5393 | * journactl can now take multiple --unit= and --user-unit= | |
5394 | switches. | |
5395 | ||
5396 | * The cryptsetup logic now understands the "luks.key=" kernel | |
5397 | command line switch for specifying a file to read the | |
7c04ad2d | 5398 | decryption key from. Also, if a configured key file is not |
ef3b5246 LP |
5399 | found the tool will now automatically fall back to prompting |
5400 | the user. | |
5401 | ||
cbeabcfb ZJS |
5402 | * Python systemd.journal module was updated to wrap recently |
5403 | added functions from libsystemd-journal. The interface was | |
5404 | changed to bring the low level interface in s.j._Reader | |
5405 | closer to the C API, and the high level interface in | |
5406 | s.j.Reader was updated to wrap and convert all data about | |
5407 | an entry. | |
5408 | ||
ef3b5246 LP |
5409 | Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer, |
5410 | Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart | |
5411 | Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer, | |
5412 | Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt, | |
5413 | Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks, | |
5414 | Tom Gundersen, Zbigniew Jędrzejewski-Szmek | |
5415 | ||
d3a86981 LP |
5416 | CHANGES WITH 201: |
5417 | ||
5418 | * journalctl --update-catalog now understands a new --root= | |
5419 | option to operate on catalogs found in a different root | |
5420 | directory. | |
5421 | ||
5422 | * During shutdown after systemd has terminated all running | |
5423 | services a final killing loop kills all remaining left-over | |
5424 | processes. We will now print the name of these processes | |
5425 | when we send SIGKILL to them, since this usually indicates a | |
5426 | problem. | |
5427 | ||
5428 | * If /etc/crypttab refers to password files stored on | |
5429 | configured mount points automatic dependencies will now be | |
5430 | generated to ensure the specific mount is established first | |
5431 | before the key file is attempted to be read. | |
5432 | ||
5433 | * 'systemctl status' will now show information about the | |
5434 | network sockets a socket unit is listening on. | |
5435 | ||
5436 | * 'systemctl status' will also shown information about any | |
5437 | drop-in configuration file for units. (Drop-In configuration | |
5438 | files in this context are files such as | |
5439 | /etc/systemd/systemd/foobar.service.d/*.conf) | |
5440 | ||
5441 | * systemd-cgtop now optionally shows summed up CPU times of | |
5442 | cgroups. Press '%' while running cgtop to switch between | |
5443 | percentage and absolute mode. This is useful to determine | |
5444 | which cgroups use up the most CPU time over the entire | |
5445 | runtime of the system. systemd-cgtop has also been updated | |
5446 | to be 'pipeable' for processing with further shell tools. | |
5447 | ||
5448 | * 'hostnamectl set-hostname' will now allow setting of FQDN | |
5449 | hostnames. | |
5450 | ||
5451 | * The formatting and parsing of time span values has been | |
5452 | changed. The parser now understands fractional expressions | |
5453 | such as "5.5h". The formatter will now output fractional | |
5454 | expressions for all time spans under 1min, i.e. "5.123456s" | |
5455 | rather than "5s 123ms 456us". For time spans under 1s | |
5456 | millisecond values are shown, for those under 1ms | |
5457 | microsecond values are shown. This should greatly improve | |
5458 | all time-related output of systemd. | |
5459 | ||
5460 | * libsystemd-login and libsystemd-journal gained new | |
5461 | functions for querying the poll() events mask and poll() | |
5462 | timeout value for integration into arbitrary event | |
5463 | loops. | |
5464 | ||
5465 | * localectl gained the ability to list available X11 keymaps | |
5466 | (models, layouts, variants, options). | |
5467 | ||
5468 | * 'systemd-analyze dot' gained the ability to filter for | |
5469 | specific units via shell-style globs, to create smaller, | |
d28315e4 | 5470 | more useful graphs. I.e. it is now possible to create simple |
d3a86981 LP |
5471 | graphs of all the dependencies between only target units, or |
5472 | of all units that Avahi has dependencies with. | |
5473 | ||
5474 | Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck, | |
5475 | Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly | |
5476 | Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau, | |
5477 | Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal | |
5478 | Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie, | |
5479 | Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav | |
5480 | Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach | |
5481 | ||
9ca3c17f LP |
5482 | CHANGES WITH 200: |
5483 | ||
5484 | * The boot-time readahead implementation for rotating media | |
5485 | will now read the read-ahead data in multiple passes which | |
5486 | consist of all read requests made in equidistant time | |
5487 | intervals. This means instead of strictly reading read-ahead | |
5488 | data in its physical order on disk we now try to find a | |
5489 | middle ground between physical and access time order. | |
5490 | ||
5491 | * /etc/os-release files gained a new BUILD_ID= field for usage | |
5492 | on operating systems that provide continuous builds of OS | |
5493 | images. | |
5494 | ||
5495 | Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers, | |
5496 | Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín | |
5497 | William Douglas, Zbigniew Jędrzejewski-Szmek | |
5498 | ||
35911459 LP |
5499 | CHANGES WITH 199: |
5500 | ||
5501 | * systemd-python gained an API exposing libsystemd-daemon. | |
5502 | ||
5503 | * The SMACK setup logic gained support for uploading CIPSO | |
5504 | security policy. | |
5505 | ||
5506 | * Behaviour of PrivateTmp=, ReadWriteDirectories=, | |
5507 | ReadOnlyDirectories= and InaccessibleDirectories= has | |
5508 | changed. The private /tmp and /var/tmp directories are now | |
5509 | shared by all processes of a service (which means | |
5510 | ExecStartPre= may now leave data in /tmp that ExecStart= of | |
5511 | the same service can still access). When a service is | |
5512 | stopped its temporary directories are immediately deleted | |
a87197f5 | 5513 | (normal clean-up with tmpfiles is still done in addition to |
35911459 LP |
5514 | this though). |
5515 | ||
5516 | * By default, systemd will now set a couple of sysctl | |
5517 | variables in the kernel: the safe sysrq options are turned | |
5518 | on, IP route verification is turned on, and source routing | |
5519 | disabled. The recently added hardlink and softlink | |
5520 | protection of the kernel is turned on. These settings should | |
5521 | be reasonably safe, and good defaults for all new systems. | |
5522 | ||
5523 | * The predictable network naming logic may now be turned off | |
a87197f5 | 5524 | with a new kernel command line switch: net.ifnames=0. |
35911459 LP |
5525 | |
5526 | * A new libsystemd-bus module has been added that implements a | |
5527 | pretty complete D-Bus client library. For details see: | |
5528 | ||
56cadcb6 | 5529 | https://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html |
35911459 | 5530 | |
c20d8298 | 5531 | * journald will now explicitly flush the journal files to disk |
a87197f5 ZJS |
5532 | at the latest 5min after each write. The file will then also |
5533 | be marked offline until the next write. This should increase | |
5534 | reliability in case of a crash. The synchronization delay | |
5535 | can be configured via SyncIntervalSec= in journald.conf. | |
35911459 LP |
5536 | |
5537 | * There's a new remote-fs-setup.target unit that can be used | |
5538 | to pull in specific services when at least one remote file | |
5539 | system is to be mounted. | |
5540 | ||
5541 | * There are new targets timers.target and paths.target as | |
5542 | canonical targets to pull user timer and path units in | |
5543 | from. This complements sockets.target with a similar | |
5544 | purpose for socket units. | |
5545 | ||
6a7d3d68 LP |
5546 | * libudev gained a new call udev_device_set_attribute_value() |
5547 | to set sysfs attributes of a device. | |
5548 | ||
a87197f5 ZJS |
5549 | * The udev daemon now sets the default number of worker |
5550 | processes executed in parallel based on the number of available | |
c20d8298 | 5551 | CPUs instead of the amount of available RAM. This is supposed |
ab06eef8 | 5552 | to provide a more reliable default and limit a too aggressive |
ce830873 | 5553 | parallelism for setups with 1000s of devices connected. |
c20d8298 | 5554 | |
35911459 LP |
5555 | Contributions from: Auke Kok, Colin Walters, Cristian |
5556 | Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes | |
5557 | Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan | |
5558 | Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering, | |
5559 | Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl, | |
5560 | Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen, | |
5561 | Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel | |
5562 | Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, | |
5563 | Zbigniew Jędrzejewski-Szmek | |
5564 | ||
85d68397 LP |
5565 | CHANGES WITH 198: |
5566 | ||
5567 | * Configuration of unit files may now be extended via drop-in | |
5568 | files without having to edit/override the unit files | |
5569 | themselves. More specifically, if the administrator wants to | |
5570 | change one value for a service file foobar.service he can | |
5571 | now do so by dropping in a configuration snippet into | |
ad88e758 | 5572 | /etc/systemd/system/foobar.service.d/*.conf. The unit logic |
85d68397 LP |
5573 | will load all these snippets and apply them on top of the |
5574 | main unit configuration file, possibly extending or | |
5575 | overriding its settings. Using these drop-in snippets is | |
40e21da8 KS |
5576 | generally nicer than the two earlier options for changing |
5577 | unit files locally: copying the files from | |
85d68397 LP |
5578 | /usr/lib/systemd/system/ to /etc/systemd/system/ and editing |
5579 | them there; or creating a new file in /etc/systemd/system/ | |
5580 | that incorporates the original one via ".include". Drop-in | |
5581 | snippets into these .d/ directories can be placed in any | |
fd868975 | 5582 | directory systemd looks for units in, and the usual |
85d68397 LP |
5583 | overriding semantics between /usr/lib, /etc and /run apply |
5584 | for them too. | |
5585 | ||
5586 | * Most unit file settings which take lists of items can now be | |
6aa8d43a | 5587 | reset by assigning the empty string to them. For example, |
85d68397 LP |
5588 | normally, settings such as Environment=FOO=BAR append a new |
5589 | environment variable assignment to the environment block, | |
5590 | each time they are used. By assigning Environment= the empty | |
5591 | string the environment block can be reset to empty. This is | |
5592 | particularly useful with the .d/*.conf drop-in snippets | |
156f7d09 KS |
5593 | mentioned above, since this adds the ability to reset list |
5594 | settings from vendor unit files via these drop-ins. | |
85d68397 LP |
5595 | |
5596 | * systemctl gained a new "list-dependencies" command for | |
5597 | listing the dependencies of a unit recursively. | |
5598 | ||
40e21da8 | 5599 | * Inhibitors are now honored and listed by "systemctl |
85d68397 LP |
5600 | suspend", "systemctl poweroff" (and similar) too, not only |
5601 | GNOME. These commands will also list active sessions by | |
5602 | other users. | |
5603 | ||
5604 | * Resource limits (as exposed by the various control group | |
5605 | controllers) can now be controlled dynamically at runtime | |
5606 | for all units. More specifically, you can now use a command | |
5607 | like "systemctl set-cgroup-attr foobar.service cpu.shares | |
5608 | 2000" to alter the CPU shares a specific service gets. These | |
6aa8d43a | 5609 | settings are stored persistently on disk, and thus allow the |
85d68397 LP |
5610 | administrator to easily adjust the resource usage of |
5611 | services with a few simple commands. This dynamic resource | |
6aa8d43a | 5612 | management logic is also available to other programs via the |
85d68397 LP |
5613 | bus. Almost any kernel cgroup attribute and controller is |
5614 | supported. | |
5615 | ||
5616 | * systemd-vconsole-setup will now copy all font settings to | |
6aa8d43a LP |
5617 | all allocated VTs, where it previously applied them only to |
5618 | the foreground VT. | |
85d68397 LP |
5619 | |
5620 | * libsystemd-login gained the new sd_session_get_tty() API | |
5621 | call. | |
5622 | ||
6aa8d43a LP |
5623 | * This release drops support for a few legacy or |
5624 | distribution-specific LSB facility names when parsing init | |
5625 | scripts: $x-display-manager, $mail-transfer-agent, | |
85d68397 LP |
5626 | $mail-transport-agent, $mail-transfer-agent, $smtp, |
5627 | $null. Also, the mail-transfer-agent.target unit backing | |
5628 | this has been removed. Distributions which want to retain | |
6aa8d43a LP |
5629 | compatibility with this should carry the burden for |
5630 | supporting this themselves and patch support for these back | |
5631 | in, if they really need to. Also, the facilities $syslog and | |
5632 | $local_fs are now ignored, since systemd does not support | |
5633 | early-boot LSB init scripts anymore, and these facilities | |
5634 | are implied anyway for normal services. syslog.target has | |
5635 | also been removed. | |
85d68397 | 5636 | |
40e21da8 | 5637 | * There are new bus calls on PID1's Manager object for |
6aa8d43a | 5638 | cancelling jobs, and removing snapshot units. Previously, |
85d68397 LP |
5639 | both calls were only available on the Job and Snapshot |
5640 | objects themselves. | |
5641 | ||
5642 | * systemd-journal-gatewayd gained SSL support. | |
5643 | ||
5644 | * The various "environment" files, such as /etc/locale.conf | |
5645 | now support continuation lines with a backslash ("\") as | |
499b604b | 5646 | last character in the line, similarly in style (but different) |
85d68397 LP |
5647 | to how this is supported in shells. |
5648 | ||
5649 | * For normal user processes the _SYSTEMD_USER_UNIT= field is | |
5650 | now implicitly appended to every log entry logged. systemctl | |
5651 | has been updated to filter by this field when operating on a | |
5652 | user systemd instance. | |
5653 | ||
5654 | * nspawn will now implicitly add the CAP_AUDIT_WRITE and | |
5655 | CAP_AUDIT_CONTROL capabilities to the capabilities set for | |
5656 | the container. This makes it easier to boot unmodified | |
5657 | Fedora systems in a container, which however still requires | |
5658 | audit=0 to be passed on the kernel command line. Auditing in | |
5659 | kernel and userspace is unfortunately still too broken in | |
5660 | context of containers, hence we recommend compiling it out | |
5661 | of the kernel or using audit=0. Hopefully this will be fixed | |
5662 | one day for good in the kernel. | |
5663 | ||
5664 | * nspawn gained the new --bind= and --bind-ro= parameters to | |
5665 | bind mount specific directories from the host into the | |
5666 | container. | |
5667 | ||
40e21da8 | 5668 | * nspawn will now mount its own devpts file system instance |
6aa8d43a | 5669 | into the container, in order not to leak pty devices from |
85d68397 LP |
5670 | the host into the container. |
5671 | ||
5672 | * systemd will now read the firmware boot time performance | |
6aa8d43a LP |
5673 | information from the EFI variables, if the used boot loader |
5674 | supports this, and takes it into account for boot performance | |
5675 | analysis via "systemd-analyze". This is currently supported | |
5676 | only in conjunction with Gummiboot, but could be supported | |
5677 | by other boot loaders too. For details see: | |
85d68397 | 5678 | |
56cadcb6 | 5679 | https://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface |
85d68397 LP |
5680 | |
5681 | * A new generator has been added that automatically mounts the | |
5682 | EFI System Partition (ESP) to /boot, if that directory | |
6aa8d43a LP |
5683 | exists, is empty, and no other file system has been |
5684 | configured to be mounted there. | |
85d68397 LP |
5685 | |
5686 | * logind will now send out PrepareForSleep(false) out | |
5687 | unconditionally, after coming back from suspend. This may be | |
5688 | used by applications as asynchronous notification for | |
5689 | system resume events. | |
5690 | ||
5691 | * "systemctl unlock-sessions" has been added, that allows | |
5692 | unlocking the screens of all user sessions at once, similar | |
499b604b | 5693 | to how "systemctl lock-sessions" already locked all users |
40e21da8 | 5694 | sessions. This is backed by a new D-Bus call UnlockSessions(). |
85d68397 LP |
5695 | |
5696 | * "loginctl seat-status" will now show the master device of a | |
5697 | seat. (i.e. the device of a seat that needs to be around for | |
5698 | the seat to be considered available, usually the graphics | |
5699 | card). | |
5700 | ||
5701 | * tmpfiles gained a new "X" line type, that allows | |
5702 | configuration of files and directories (with wildcards) that | |
5703 | shall be excluded from automatic cleanup ("aging"). | |
5704 | ||
bf933560 KS |
5705 | * udev default rules set the device node permissions now only |
5706 | at "add" events, and do not change them any longer with a | |
5707 | later "change" event. | |
85d68397 LP |
5708 | |
5709 | * The log messages for lid events and power/sleep keypresses | |
5710 | now carry a message ID. | |
5711 | ||
5712 | * We now have a substantially larger unit test suite, but this | |
5713 | continues to be work in progress. | |
5714 | ||
5715 | * udevadm hwdb gained a new --root= parameter to change the | |
5716 | root directory to operate relative to. | |
5717 | ||
40e21da8 KS |
5718 | * logind will now issue a background sync() request to the kernel |
5719 | early at shutdown, so that dirty buffers are flushed to disk early | |
85d68397 LP |
5720 | instead of at the last moment, in order to optimize shutdown |
5721 | times a little. | |
5722 | ||
5723 | * A new bootctl tool has been added that is an interface for | |
5724 | certain boot loader operations. This is currently a preview | |
5725 | and is likely to be extended into a small mechanism daemon | |
5726 | like timedated, localed, hostnamed, and can be used by | |
5727 | graphical UIs to enumerate available boot options, and | |
5728 | request boot into firmware operations. | |
5729 | ||
5730 | * systemd-bootchart has been relicensed to LGPLv2.1+ to match | |
5731 | the rest of the package. It also has been updated to work | |
5732 | correctly in initrds. | |
5733 | ||
5734 | * Policykit previously has been runtime optional, and is now | |
5735 | also compile time optional via a configure switch. | |
5736 | ||
5737 | * systemd-analyze has been reimplemented in C. Also "systemctl | |
5738 | dot" has moved into systemd-analyze. | |
5739 | ||
5740 | * "systemctl status" with no further parameters will now print | |
5741 | the status of all active or failed units. | |
5742 | ||
5743 | * Operations such as "systemctl start" can now be executed | |
5744 | with a new mode "--irreversible" which may be used to queue | |
5745 | operations that cannot accidentally be reversed by a later | |
6aa8d43a | 5746 | job queuing. This is by default used to make shutdown |
85d68397 LP |
5747 | requests more robust. |
5748 | ||
5749 | * The Python API of systemd now gained a new module for | |
5750 | reading journal files. | |
5751 | ||
5752 | * A new tool kernel-install has been added that can install | |
5753 | kernel images according to the Boot Loader Specification: | |
5754 | ||
56cadcb6 | 5755 | https://www.freedesktop.org/wiki/Specifications/BootLoaderSpec |
85d68397 LP |
5756 | |
5757 | * Boot time console output has been improved to provide | |
6aa8d43a | 5758 | animated boot time output for hanging jobs. |
85d68397 LP |
5759 | |
5760 | * A new tool systemd-activate has been added which can be used | |
5761 | to test socket activation with, directly from the command | |
5762 | line. This should make it much easier to test and debug | |
5763 | socket activation in daemons. | |
5764 | ||
5765 | * journalctl gained a new "--reverse" (or -r) option to show | |
5766 | journal output in reverse order (i.e. newest line first). | |
5767 | ||
43447fb7 LP |
5768 | * journalctl gained a new "--pager-end" (or -e) option to jump |
5769 | to immediately jump to the end of the journal in the | |
5770 | pager. This is only supported in conjunction with "less". | |
5771 | ||
85d68397 | 5772 | * journalctl gained a new "--user-unit=" option, that works |
499b604b | 5773 | similarly to "--unit=" but filters for user units rather than |
85d68397 LP |
5774 | system units. |
5775 | ||
5776 | * A number of unit files to ease adoption of systemd in | |
5777 | initrds has been added. This moves some minimal logic from | |
5778 | the various initrd implementations into systemd proper. | |
5779 | ||
5780 | * The journal files are now owned by a new group | |
5781 | "systemd-journal", which exists specifically to allow access | |
5782 | to the journal, and nothing else. Previously, we used the | |
6aa8d43a | 5783 | "adm" group for that, which however possibly covers more |
85d68397 LP |
5784 | than just journal/log file access. This new group is now |
5785 | already used by systemd-journal-gatewayd to ensure this | |
5786 | daemon gets access to the journal files and as little else | |
5787 | as possible. Note that "make install" will also set FS ACLs | |
5788 | up for /var/log/journal to give "adm" and "wheel" read | |
5789 | access to it, in addition to "systemd-journal" which owns | |
5790 | the journal files. We recommend that packaging scripts also | |
6aa8d43a | 5791 | add read access to "adm" + "wheel" to /var/log/journal, and |
85d68397 LP |
5792 | all existing/future journal files. To normal users and |
5793 | administrators little changes, however packagers need to | |
5794 | ensure to create the "systemd-journal" system group at | |
5795 | package installation time. | |
5796 | ||
5797 | * The systemd-journal-gatewayd now runs as unprivileged user | |
5798 | systemd-journal-gateway:systemd-journal-gateway. Packaging | |
5799 | scripts need to create these system user/group at | |
5800 | installation time. | |
5801 | ||
5802 | * timedated now exposes a new boolean property CanNTP that | |
5803 | indicates whether a local NTP service is available or not. | |
5804 | ||
5805 | * systemd-detect-virt will now also detect xen PVs | |
5806 | ||
40e21da8 KS |
5807 | * The pstore file system is now mounted by default, if it is |
5808 | available. | |
85d68397 | 5809 | |
1aed4590 LP |
5810 | * In addition to the SELinux and IMA policies we will now also |
5811 | load SMACK policies at early boot. | |
5812 | ||
85d68397 LP |
5813 | Contributions from: Adel Gadllah, Aleksander Morgado, Auke |
5814 | Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch, | |
5815 | Daniel Wallace, Dave Reisner, David Herrmann, David Strauss, | |
5816 | Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer, | |
5817 | Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering, | |
5818 | Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin | |
5819 | Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael | |
5820 | Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil, | |
5821 | Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor | |
5822 | Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob | |
5823 | Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven | |
5824 | Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom | |
5825 | Gundersen, Umut Tezduyar, William Giokas, Zbigniew | |
5826 | Jędrzejewski-Szmek, Zeeshan Ali (Khattak) | |
5827 | ||
8ad26859 LP |
5828 | CHANGES WITH 197: |
5829 | ||
5830 | * Timer units now support calendar time events in addition to | |
5831 | monotonic time events. That means you can now trigger a unit | |
5832 | based on a calendar time specification such as "Thu,Fri | |
5833 | 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first | |
5834 | or fifth day of any month of the year 2013, given that it is | |
5835 | a thursday or friday. This brings timer event support | |
5836 | considerably closer to cron's capabilities. For details on | |
5837 | the supported calendar time specification language see | |
5838 | systemd.time(7). | |
5839 | ||
5840 | * udev now supports a number of different naming policies for | |
5841 | network interfaces for predictable names, and a combination | |
5842 | of these policies is now the default. Please see this wiki | |
5843 | document for details: | |
5844 | ||
56cadcb6 | 5845 | https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames |
8ad26859 LP |
5846 | |
5847 | * Auke Kok's bootchart implementation has been added to the | |
d28315e4 JE |
5848 | systemd tree. It is an optional component that can graph the |
5849 | boot in quite some detail. It is one of the best bootchart | |
8ad26859 LP |
5850 | implementations around and minimal in its code and |
5851 | dependencies. | |
5852 | ||
5853 | * nss-myhostname has been integrated into the systemd source | |
5854 | tree. nss-myhostname guarantees that the local hostname | |
5855 | always stays resolvable via NSS. It has been a weak | |
5856 | requirement of systemd-hostnamed since a long time, and | |
5857 | since its code is actually trivial we decided to just | |
5858 | include it in systemd's source tree. It can be turned off | |
5859 | with a configure switch. | |
5860 | ||
5861 | * The read-ahead logic is now capable of properly detecting | |
5862 | whether a btrfs file system is on SSD or rotating media, in | |
5863 | order to optimize the read-ahead scheme. Previously, it was | |
5864 | only capable of detecting this on traditional file systems | |
5865 | such as ext4. | |
5866 | ||
5867 | * In udev, additional device properties are now read from the | |
5868 | IAB in addition to the OUI database. Also, Bluetooth company | |
5869 | identities are attached to the devices as well. | |
5870 | ||
5871 | * In service files %U may be used as specifier that is | |
5872 | replaced by the configured user name of the service. | |
5873 | ||
5874 | * nspawn may now be invoked without a controlling TTY. This | |
5875 | makes it suitable for invocation as its own service. This | |
5876 | may be used to set up a simple containerized server system | |
5877 | using only core OS tools. | |
5878 | ||
5879 | * systemd and nspawn can now accept socket file descriptors | |
5880 | when they are started for socket activation. This enables | |
5881 | implementation of socket activated nspawn | |
5882 | containers. i.e. think about autospawning an entire OS image | |
5883 | when the first SSH or HTTP connection is received. We expect | |
5884 | that similar functionality will also be added to libvirt-lxc | |
5885 | eventually. | |
5886 | ||
5887 | * journalctl will now suppress ANSI color codes when | |
5888 | presenting log data. | |
5889 | ||
5890 | * systemctl will no longer show control group information for | |
ce830873 | 5891 | a unit if the control group is empty anyway. |
8ad26859 LP |
5892 | |
5893 | * logind can now automatically suspend/hibernate/shutdown the | |
5894 | system on idle. | |
5895 | ||
5896 | * /etc/machine-info and hostnamed now also expose the chassis | |
5897 | type of the system. This can be used to determine whether | |
5898 | the local system is a laptop, desktop, handset or | |
5899 | tablet. This information may either be configured by the | |
5900 | user/vendor or is automatically determined from ACPI and DMI | |
5901 | information if possible. | |
5902 | ||
5903 | * A number of PolicyKit actions are now bound together with | |
5904 | "imply" rules. This should simplify creating UIs because | |
5905 | many actions will now authenticate similar ones as well. | |
5906 | ||
5907 | * Unit files learnt a new condition ConditionACPower= which | |
5908 | may be used to conditionalize a unit depending on whether an | |
5909 | AC power source is connected or not, of whether the system | |
5910 | is running on battery power. | |
5911 | ||
5912 | * systemctl gained a new "is-failed" verb that may be used in | |
5913 | shell scripts and suchlike to check whether a specific unit | |
5914 | is in the "failed" state. | |
5915 | ||
5916 | * The EnvironmentFile= setting in unit files now supports file | |
5917 | globbing, and can hence be used to easily read a number of | |
5918 | environment files at once. | |
5919 | ||
5920 | * systemd will no longer detect and recognize specific | |
5921 | distributions. All distribution-specific #ifdeffery has been | |
5922 | removed, systemd is now fully generic and | |
5923 | distribution-agnostic. Effectively, not too much is lost as | |
5924 | a lot of the code is still accessible via explicit configure | |
5925 | switches. However, support for some distribution specific | |
5926 | legacy configuration file formats has been dropped. We | |
5927 | recommend distributions to simply adopt the configuration | |
5928 | files everybody else uses now and convert the old | |
5929 | configuration from packaging scripts. Most distributions | |
5930 | already did that. If that's not possible or desirable, | |
5931 | distributions are welcome to forward port the specific | |
5932 | pieces of code locally from the git history. | |
5933 | ||
5934 | * When logging a message about a unit systemd will now always | |
5935 | log the unit name in the message meta data. | |
5936 | ||
5937 | * localectl will now also discover system locale data that is | |
5938 | not stored in locale archives, but directly unpacked. | |
5939 | ||
5940 | * logind will no longer unconditionally use framebuffer | |
5941 | devices as seat masters, i.e. as devices that are required | |
5942 | to be existing before a seat is considered preset. Instead, | |
5943 | it will now look for all devices that are tagged as | |
b938cb90 JE |
5944 | "seat-master" in udev. By default, framebuffer devices will |
5945 | be marked as such, but depending on local systems, other | |
8ad26859 LP |
5946 | devices might be marked as well. This may be used to |
5947 | integrate graphics cards using closed source drivers (such | |
5948 | as NVidia ones) more nicely into logind. Note however, that | |
5949 | we recommend using the open source NVidia drivers instead, | |
5950 | and no udev rules for the closed-source drivers will be | |
5951 | shipped from us upstream. | |
5952 | ||
5953 | Contributions from: Adam Williamson, Alessandro Crismani, Auke | |
5954 | Kok, Colin Walters, Daniel Wallace, Dave Reisner, David | |
5955 | Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra, | |
5956 | Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik | |
5957 | Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart | |
5958 | Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann, | |
5959 | Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry, | |
5960 | Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg | |
5961 | Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar | |
5962 | Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn | |
5963 | Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch, | |
5964 | Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew | |
5965 | Jędrzejewski-Szmek | |
5966 | ||
0428ddb7 LP |
5967 | CHANGES WITH 196: |
5968 | ||
5969 | * udev gained support for loading additional device properties | |
5970 | from an indexed database that is keyed by vendor/product IDs | |
5971 | and similar device identifiers. For the beginning this | |
5972 | "hwdb" is populated with data from the well-known PCI and | |
5973 | USB database, but also includes PNP, ACPI and OID data. In | |
5974 | the longer run this indexed database shall grow into | |
5975 | becoming the one central database for non-essential | |
5976 | userspace device metadata. Previously, data from the PCI/USB | |
96ec33c0 | 5977 | database was only attached to select devices, since the |
0428ddb7 | 5978 | lookup was a relatively expensive operation due to O(n) time |
96ec33c0 LP |
5979 | complexity (with n being the number of entries in the |
5980 | database). Since this is now O(1), we decided to add in this | |
5981 | data for all devices where this is available, by | |
0428ddb7 LP |
5982 | default. Note that the indexed database needs to be rebuilt |
5983 | when new data files are installed. To achieve this you need | |
5984 | to update your packaging scripts to invoke "udevadm hwdb | |
5985 | --update" after installation of hwdb data files. For | |
5986 | RPM-based distributions we introduced the new | |
5987 | %udev_hwdb_update macro for this purpose. | |
5988 | ||
5989 | * The Journal gained support for the "Message Catalog", an | |
5990 | indexed database to link up additional information with | |
5991 | journal entries. For further details please check: | |
5992 | ||
56cadcb6 | 5993 | https://www.freedesktop.org/wiki/Software/systemd/catalog |
0428ddb7 LP |
5994 | |
5995 | The indexed message catalog database also needs to be | |
5996 | rebuilt after installation of message catalog files. Use | |
5997 | "journalctl --update-catalog" for this. For RPM-based | |
5998 | distributions we introduced the %journal_catalog_update | |
5999 | macro for this purpose. | |
6000 | ||
6001 | * The Python Journal bindings gained support for the standard | |
6002 | Python logging framework. | |
6003 | ||
6004 | * The Journal API gained new functions for checking whether | |
6005 | the underlying file system of a journal file is capable of | |
6006 | properly reporting file change notifications, or whether | |
6007 | applications that want to reflect journal changes "live" | |
ab06eef8 | 6008 | need to recheck journal files continuously in appropriate |
0428ddb7 LP |
6009 | time intervals. |
6010 | ||
6011 | * It is now possible to set the "age" field for tmpfiles | |
6012 | entries to 0, indicating that files matching this entry | |
6013 | shall always be removed when the directories are cleaned up. | |
6014 | ||
6015 | * coredumpctl gained a new "gdb" verb which invokes gdb | |
6016 | right-away on the selected coredump. | |
6017 | ||
6018 | * There's now support for "hybrid sleep" on kernels that | |
6019 | support this, in addition to "suspend" and "hibernate". Use | |
6020 | "systemctl hybrid-sleep" to make use of this. | |
6021 | ||
6022 | * logind's HandleSuspendKey= setting (and related settings) | |
6023 | now gained support for a new "lock" setting to simply | |
6024 | request the screen lock on all local sessions, instead of | |
6025 | actually executing a suspend or hibernation. | |
6026 | ||
6027 | * systemd will now mount the EFI variables file system by | |
6028 | default. | |
6029 | ||
6030 | * Socket units now gained support for configuration of the | |
6031 | SMACK security label. | |
6032 | ||
6033 | * timedatectl will now output the time of the last and next | |
6034 | daylight saving change. | |
6035 | ||
6036 | * We dropped support for various legacy and distro-specific | |
6037 | concepts, such as insserv, early-boot SysV services | |
6038 | (i.e. those for non-standard runlevels such as 'b' or 'S') | |
6039 | or ArchLinux /etc/rc.conf support. We recommend the | |
6040 | distributions who still need support this to either continue | |
6041 | to maintain the necessary patches downstream, or find a | |
6042 | different solution. (Talk to us if you have questions!) | |
6043 | ||
6044 | * Various systemd components will now bypass PolicyKit checks | |
6045 | for root and otherwise handle properly if PolicyKit is not | |
6046 | found to be around. This should fix most issues for | |
6047 | PolicyKit-less systems. Quite frankly this should have been | |
6048 | this way since day one. It is absolutely our intention to | |
6049 | make systemd work fine on PolicyKit-less systems, and we | |
d28315e4 | 6050 | consider it a bug if something does not work as it should if |
0428ddb7 LP |
6051 | PolicyKit is not around. |
6052 | ||
6053 | * For embedded systems it is now possible to build udev and | |
6054 | systemd without blkid and/or kmod support. | |
6055 | ||
6056 | * "systemctl switch-root" is now capable of switching root | |
6057 | more than once. I.e. in addition to transitions from the | |
6058 | initrd to the host OS it is now possible to transition to | |
6059 | further OS images from the host. This is useful to implement | |
6060 | offline updating tools. | |
6061 | ||
6062 | * Various other additions have been made to the RPM macros | |
6063 | shipped with systemd. Use %udev_rules_update() after | |
6064 | installing new udev rules files. %_udevhwdbdir, | |
6065 | %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir, | |
6066 | %_sysctldir are now available which resolve to the right | |
6067 | directories for packages to place various data files in. | |
6068 | ||
6069 | * journalctl gained the new --full switch (in addition to | |
6070 | --all, to disable ellipsation for long messages. | |
6071 | ||
6072 | Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel, | |
6073 | Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner, | |
6074 | Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers, | |
6075 | Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas, | |
6076 | Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl, | |
6077 | Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen, | |
6078 | Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas | |
6079 | Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony | |
6080 | Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek | |
6081 | ||
139ee8cc LP |
6082 | CHANGES WITH 195: |
6083 | ||
6827101a | 6084 | * journalctl gained new --since= and --until= switches to |
139ee8cc LP |
6085 | filter by time. It also now supports nice filtering for |
6086 | units via --unit=/-u. | |
6087 | ||
6827101a | 6088 | * Type=oneshot services may use ExecReload= and do the |
139ee8cc LP |
6089 | right thing. |
6090 | ||
6091 | * The journal daemon now supports time-based rotation and | |
6092 | vacuuming, in addition to the usual disk-space based | |
6093 | rotation. | |
6094 | ||
6095 | * The journal will now index the available field values for | |
6096 | each field name. This enables clients to show pretty drop | |
6097 | downs of available match values when filtering. The bash | |
6098 | completion of journalctl has been updated | |
6099 | accordingly. journalctl gained a new switch -F to list all | |
6100 | values a certain field takes in the journal database. | |
6101 | ||
6102 | * More service events are now written as structured messages | |
6103 | to the journal, and made recognizable via message IDs. | |
6104 | ||
6105 | * The timedated, localed and hostnamed mini-services which | |
6106 | previously only provided support for changing time, locale | |
6107 | and hostname settings from graphical DEs such as GNOME now | |
6108 | also have a minimal (but very useful) text-based client | |
6109 | utility each. This is probably the nicest way to changing | |
6110 | these settings from the command line now, especially since | |
6111 | it lists available options and is fully integrated with bash | |
6112 | completion. | |
6113 | ||
6114 | * There's now a new tool "systemd-coredumpctl" to list and | |
6115 | extract coredumps from the journal. | |
6116 | ||
6117 | * We now install a README each in /var/log/ and | |
6118 | /etc/rc.d/init.d explaining where the system logs and init | |
6119 | scripts went. This hopefully should help folks who go to | |
6120 | that dirs and look into the otherwise now empty void and | |
6121 | scratch their heads. | |
6122 | ||
6123 | * When user-services are invoked (by systemd --user) the | |
6124 | $MANAGERPID env var is set to the PID of systemd. | |
6125 | ||
6126 | * SIGRTMIN+24 when sent to a --user instance will now result | |
6127 | in immediate termination of systemd. | |
6128 | ||
6129 | * gatewayd received numerous feature additions such as a | |
6130 | "follow" mode, for live syncing and filtering. | |
6131 | ||
6132 | * browse.html now allows filtering and showing detailed | |
6133 | information on specific entries. Keyboard navigation and | |
6134 | mouse screen support has been added. | |
6135 | ||
6136 | * gatewayd/journalctl now supports HTML5/JSON | |
6137 | Server-Sent-Events as output. | |
6138 | ||
1cb88f2c | 6139 | * The SysV init script compatibility logic will now |
139ee8cc LP |
6140 | heuristically determine whether a script supports the |
6141 | "reload" verb, and only then make this available as | |
6142 | "systemctl reload". | |
6143 | ||
15f47220 | 6144 | * "systemctl status --follow" has been removed, use "journalctl |
139ee8cc LP |
6145 | -u" instead. |
6146 | ||
6147 | * journald.conf's RuntimeMinSize=, PersistentMinSize= settings | |
6148 | have been removed since they are hardly useful to be | |
6149 | configured. | |
6150 | ||
6151 | * And I'd like to take the opportunity to specifically mention | |
6152 | Zbigniew for his great contributions. Zbigniew, you rock! | |
6153 | ||
6154 | Contributions from: Andrew Eikum, Christian Hesse, Colin | |
6155 | Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc | |
4d92e078 LP |
6156 | Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas |
6157 | Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich, | |
6158 | Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas | |
6159 | Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew | |
6160 | Jędrzejewski-Szmek, Сковорода Никита Андреевич | |
139ee8cc | 6161 | |
f9b55720 LP |
6162 | CHANGES WITH 194: |
6163 | ||
6164 | * If /etc/vconsole.conf is non-existent or empty we will no | |
6165 | longer load any console font or key map at boot by | |
6166 | default. Instead the kernel defaults will be left | |
6167 | intact. This is definitely the right thing to do, as no | |
6168 | configuration should mean no configuration, and hard-coding | |
6169 | font names that are different on all archs is probably a bad | |
6170 | idea. Also, the kernel default key map and font should be | |
6171 | good enough for most cases anyway, and mostly identical to | |
6172 | the userspace fonts/key maps we previously overloaded them | |
6173 | with. If distributions want to continue to default to a | |
6174 | non-kernel font or key map they should ship a default | |
6175 | /etc/vconsole.conf with the appropriate contents. | |
6176 | ||
6177 | Contributions from: Colin Walters, Daniel J Walsh, Dave | |
6178 | Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef | |
6179 | Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek | |
6180 | ||
597c52cf LP |
6181 | CHANGES WITH 193: |
6182 | ||
6183 | * journalctl gained a new --cursor= switch to show entries | |
6184 | starting from the specified location in the journal. | |
6185 | ||
6186 | * We now enforce a size limit on journal entry fields exported | |
6187 | with "-o json" in journalctl. Fields larger than 4K will be | |
6188 | assigned null. This can be turned off with --all. | |
6189 | ||
6190 | * An (optional) journal gateway daemon is now available as | |
6191 | "systemd-journal-gatewayd.service". This service provides | |
6192 | access to the journal via HTTP and JSON. This functionality | |
6193 | will be used to implement live log synchronization in both | |
6194 | pull and push modes, but has various other users too, such | |
6195 | as easy log access for debugging of embedded devices. Right | |
6196 | now it is already useful to retrieve the journal via HTTP: | |
6197 | ||
6198 | # systemctl start systemd-journal-gatewayd.service | |
6199 | # wget http://localhost:19531/entries | |
6200 | ||
6201 | This will download the journal contents in a | |
6202 | /var/log/messages compatible format. The same as JSON: | |
6203 | ||
6204 | # curl -H"Accept: application/json" http://localhost:19531/entries | |
6205 | ||
6206 | This service is also accessible via a web browser where a | |
6207 | single static HTML5 app is served that uses the JSON logic | |
6208 | to enable the user to do some basic browsing of the | |
6209 | journal. This will be extended later on. Here's an example | |
6210 | screenshot of this app in its current state: | |
6211 | ||
6212 | http://0pointer.de/public/journal-gatewayd | |
6213 | ||
6214 | Contributions from: Kay Sievers, Lennart Poettering, Robert | |
6215 | Milasan, Tom Gundersen | |
6216 | ||
075d4ecb LP |
6217 | CHANGES WITH 192: |
6218 | ||
6219 | * The bash completion logic is now available for journalctl | |
6220 | too. | |
6221 | ||
d28315e4 | 6222 | * We do not mount the "cpuset" controller anymore together with |
075d4ecb LP |
6223 | "cpu" and "cpuacct", as "cpuset" groups generally cannot be |
6224 | started if no parameters are assigned to it. "cpuset" hence | |
61233823 | 6225 | broke code that assumed it could create "cpu" groups and |
075d4ecb LP |
6226 | just start them. |
6227 | ||
6228 | * journalctl -f will now subscribe to terminal size changes, | |
6229 | and line break accordingly. | |
6230 | ||
597c52cf LP |
6231 | Contributions from: Dave Reisner, Kay Sievers, Lennart |
6232 | Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín | |
075d4ecb | 6233 | |
b6a86739 LP |
6234 | CHANGES WITH 191: |
6235 | ||
6236 | * nspawn will now create a symlink /etc/localtime in the | |
6237 | container environment, copying the host's timezone | |
6238 | setting. Previously this has been done via a bind mount, but | |
6239 | since symlinks cannot be bind mounted this has now been | |
6240 | changed to create/update the appropriate symlink. | |
6241 | ||
6242 | * journalctl -n's line number argument is now optional, and | |
6243 | will default to 10 if omitted. | |
6244 | ||
6245 | * journald will now log the maximum size the journal files may | |
6246 | take up on disk. This is particularly useful if the default | |
6247 | built-in logic of determining this parameter from the file | |
6248 | system size is used. Use "systemctl status | |
6563b535 | 6249 | systemd-journald.service" to see this information. |
b6a86739 LP |
6250 | |
6251 | * The multi-seat X wrapper tool has been stripped down. As X | |
6252 | is now capable of enumerating graphics devices via udev in a | |
6253 | seat-aware way the wrapper is not strictly necessary | |
6254 | anymore. A stripped down temporary stop-gap is still shipped | |
6255 | until the upstream display managers have been updated to | |
6256 | fully support the new X logic. Expect this wrapper to be | |
6563b535 | 6257 | removed entirely in one of the next releases. |
b6a86739 LP |
6258 | |
6259 | * HandleSleepKey= in logind.conf has been split up into | |
6260 | HandleSuspendKey= and HandleHibernateKey=. The old setting | |
6563b535 | 6261 | is not available anymore. X11 and the kernel are |
45afd519 | 6262 | distinguishing between these keys and we should too. This |
b6a86739 LP |
6263 | also means the inhibition lock for these keys has been split |
6264 | into two. | |
6265 | ||
597c52cf LP |
6266 | Contributions from: Dave Airlie, Eelco Dolstra, Lennart |
6267 | Poettering, Lukas Nykryn, Václav Pavlín | |
b6a86739 | 6268 | |
0c11f949 LP |
6269 | CHANGES WITH 190: |
6270 | ||
d28315e4 | 6271 | * Whenever a unit changes state we will now log this to the |
0c11f949 LP |
6272 | journal and show along the unit's own log output in |
6273 | "systemctl status". | |
6274 | ||
6275 | * ConditionPathIsMountPoint= can now properly detect bind | |
6276 | mount points too. (Previously, a bind mount of one file | |
8d0256b7 | 6277 | system to another place in the same file system could not be |
0c11f949 LP |
6278 | detected as mount, since they shared struct stat's st_dev |
6279 | field.) | |
6280 | ||
6281 | * We will now mount the cgroup controllers cpu, cpuacct, | |
6282 | cpuset and the controllers net_cls, net_prio together by | |
6283 | default. | |
6284 | ||
6285 | * nspawn containers will now have a virtualized boot | |
6286 | ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted | |
6287 | over with a randomized ID at container initialization). This | |
6288 | has the effect of making "journalctl -b" do the right thing | |
6289 | in a container. | |
6290 | ||
6291 | * The JSON output journal serialization has been updated not | |
6292 | to generate "endless" list objects anymore, but rather one | |
6293 | JSON object per line. This is more in line how most JSON | |
6294 | parsers expect JSON objects. The new output mode | |
6295 | "json-pretty" has been added to provide similar output, but | |
6296 | neatly aligned for readability by humans. | |
6297 | ||
6298 | * We dropped all explicit sync() invocations in the shutdown | |
6299 | code. The kernel does this implicitly anyway in the kernel | |
6300 | reboot() syscall. halt(8)'s -n option is now a compatibility | |
6301 | no-op. | |
6302 | ||
6303 | * We now support virtualized reboot() in containers, as | |
6304 | supported by newer kernels. We will fall back to exit() if | |
6305 | CAP_SYS_REBOOT is not available to the container. Also, | |
6306 | nspawn makes use of this now and will actually reboot the | |
6307 | container if the containerized OS asks for that. | |
6308 | ||
6309 | * journalctl will only show local log output by default | |
6310 | now. Use --merge (-m) to show remote log output, too. | |
6311 | ||
6312 | * libsystemd-journal gained the new sd_journal_get_usage() | |
6313 | call to determine the current disk usage of all journal | |
6314 | files. This is exposed in the new "journalctl --disk-usage" | |
6315 | command. | |
6316 | ||
6317 | * journald gained a new configuration setting SplitMode= in | |
6318 | journald.conf which may be used to control how user journals | |
6319 | are split off. See journald.conf(5) for details. | |
6320 | ||
6321 | * A new condition type ConditionFileNotEmpty= has been added. | |
6322 | ||
6323 | * tmpfiles' "w" lines now support file globbing, to write | |
6324 | multiple files at once. | |
6325 | ||
6326 | * We added Python bindings for the journal submission | |
6327 | APIs. More Python APIs for a number of selected APIs will | |
6328 | likely follow. Note that we intend to add native bindings | |
6329 | only for the Python language, as we consider it common | |
6330 | enough to deserve bindings shipped within systemd. There are | |
6331 | various projects outside of systemd that provide bindings | |
6332 | for languages such as PHP or Lua. | |
6333 | ||
a98d5d64 LP |
6334 | * Many conditions will now resolve specifiers such as %i. In |
6335 | addition, PathChanged= and related directives of .path units | |
6336 | now support specifiers as well. | |
0c11f949 LP |
6337 | |
6338 | * There's now a new RPM macro definition for the system preset | |
6339 | dir: %_presetdir. | |
6340 | ||
d28315e4 | 6341 | * journald will now warn if it ca not forward a message to the |
dca348bc | 6342 | syslog daemon because its socket is full. |
0c11f949 LP |
6343 | |
6344 | * timedated will no longer write or process /etc/timezone, | |
6345 | except on Debian. As we do not support late mounted /usr | |
6346 | anymore /etc/localtime always being a symlink is now safe, | |
6347 | and hence the information in /etc/timezone is not necessary | |
6348 | anymore. | |
6349 | ||
aaccc32c | 6350 | * logind will now always reserve one VT for a text getty (VT6 |
0c11f949 LP |
6351 | by default). Previously if more than 6 X sessions where |
6352 | started they took up all the VTs with auto-spawned gettys, | |
6353 | so that no text gettys were available anymore. | |
6354 | ||
6355 | * udev will now automatically inform the btrfs kernel logic | |
6356 | about btrfs RAID components showing up. This should make | |
6357 | simple hotplug based btrfs RAID assembly work. | |
6358 | ||
6359 | * PID 1 will now increase its RLIMIT_NOFILE to 64K by default | |
6360 | (but not for its children which will stay at the kernel | |
6361 | default). This should allow setups with a lot more listening | |
6362 | sockets. | |
6363 | ||
6364 | * systemd will now always pass the configured timezone to the | |
6365 | kernel at boot. timedated will do the same when the timezone | |
6366 | is changed. | |
6367 | ||
6368 | * logind's inhibition logic has been updated. By default, | |
6369 | logind will now handle the lid switch, the power and sleep | |
6370 | keys all the time, even in graphical sessions. If DEs want | |
6371 | to handle these events on their own they should take the new | |
6372 | handle-power-key, handle-sleep-key and handle-lid-switch | |
f131770b | 6373 | inhibitors during their runtime. A simple way to achieve |
0c11f949 LP |
6374 | that is to invoke the DE wrapped in an invocation of: |
6375 | ||
6376 | systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ... | |
6377 | ||
6378 | * Access to unit operations is now checked via SELinux taking | |
6379 | the unit file label and client process label into account. | |
6380 | ||
aad803af LP |
6381 | * systemd will now notify the administrator in the journal |
6382 | when he over-mounts a non-empty directory. | |
6383 | ||
6384 | * There are new specifiers that are resolved in unit files, | |
6385 | for the host name (%H), the machine ID (%m) and the boot ID | |
6386 | (%b). | |
6387 | ||
b6a86739 | 6388 | Contributions from: Allin Cottrell, Auke Kok, Brandon Philips, |
0c11f949 LP |
6389 | Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner, |
6390 | Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart | |
6391 | Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas, | |
6392 | Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz, | |
6393 | Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen, | |
6394 | Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek | |
6395 | ||
38a60d71 LP |
6396 | CHANGES WITH 189: |
6397 | ||
6398 | * Support for reading structured kernel messages from | |
6399 | /dev/kmsg has now been added and is enabled by default. | |
6400 | ||
6401 | * Support for reading kernel messages from /proc/kmsg has now | |
6402 | been removed. If you want kernel messages in the journal | |
6403 | make sure to run a recent kernel (>= 3.5) that supports | |
6404 | reading structured messages from /dev/kmsg (see | |
6405 | above). /proc/kmsg is now exclusive property of classic | |
6406 | syslog daemons again. | |
6407 | ||
6408 | * The libudev API gained the new | |
6409 | udev_device_new_from_device_id() call. | |
6410 | ||
6411 | * The logic for file system namespace (ReadOnlyDirectory=, | |
6412 | ReadWriteDirectoy=, PrivateTmp=) has been reworked not to | |
6413 | require pivot_root() anymore. This means fewer temporary | |
6414 | directories are created below /tmp for this feature. | |
6415 | ||
6416 | * nspawn containers will now see and receive all submounts | |
6417 | made on the host OS below the root file system of the | |
6418 | container. | |
6419 | ||
6420 | * Forward Secure Sealing is now supported for Journal files, | |
6421 | which provide cryptographical sealing of journal files so | |
6422 | that attackers cannot alter log history anymore without this | |
6423 | being detectable. Lennart will soon post a blog story about | |
6424 | this explaining it in more detail. | |
6425 | ||
6426 | * There are two new service settings RestartPreventExitStatus= | |
6427 | and SuccessExitStatus= which allow configuration of exit | |
6428 | status (exit code or signal) which will be excepted from the | |
6429 | restart logic, resp. consider successful. | |
6430 | ||
6431 | * journalctl gained the new --verify switch that can be used | |
6432 | to check the integrity of the structure of journal files and | |
6433 | (if Forward Secure Sealing is enabled) the contents of | |
6434 | journal files. | |
6435 | ||
6436 | * nspawn containers will now be run with /dev/stdin, /dev/fd/ | |
6437 | and similar symlinks pre-created. This makes running shells | |
6438 | as container init process a lot more fun. | |
6439 | ||
6440 | * The fstab support can now handle PARTUUID= and PARTLABEL= | |
6441 | entries. | |
6442 | ||
6443 | * A new ConditionHost= condition has been added to match | |
6444 | against the hostname (with globs) and machine ID. This is | |
6445 | useful for clusters where a single OS image is used to | |
6446 | provision a large number of hosts which shall run slightly | |
6447 | different sets of services. | |
6448 | ||
6449 | * Services which hit the restart limit will now be placed in a | |
6450 | failure state. | |
6451 | ||
b6a86739 | 6452 | Contributions from: Bertram Poettering, Dave Reisner, Huang |
38a60d71 LP |
6453 | Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin |
6454 | Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek | |
6455 | ||
c269cec3 LP |
6456 | CHANGES WITH 188: |
6457 | ||
6458 | * When running in --user mode systemd will now become a | |
6459 | subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps | |
6460 | tree a lot more organized. | |
6461 | ||
6462 | * A new PartOf= unit dependency type has been introduced that | |
6463 | may be used to group services in a natural way. | |
6464 | ||
6465 | * "systemctl enable" may now be used to enable instances of | |
6466 | services. | |
6467 | ||
6468 | * journalctl now prints error log levels in red, and | |
6469 | warning/notice log levels in bright white. It also supports | |
6470 | filtering by log level now. | |
6471 | ||
6472 | * cgtop gained a new -n switch (similar to top), to configure | |
6473 | the maximum number of iterations to run for. It also gained | |
6474 | -b, to run in batch mode (accepting no input). | |
6475 | ||
ab06eef8 | 6476 | * The suffix ".service" may now be omitted on most systemctl |
c269cec3 LP |
6477 | command lines involving service unit names. |
6478 | ||
6479 | * There's a new bus call in logind to lock all sessions, as | |
6480 | well as a loginctl verb for it "lock-sessions". | |
6481 | ||
6482 | * libsystemd-logind.so gained a new call sd_journal_perror() | |
6483 | that works similar to libc perror() but logs to the journal | |
6484 | and encodes structured information about the error number. | |
6485 | ||
6486 | * /etc/crypttab entries now understand the new keyfile-size= | |
6487 | option. | |
6488 | ||
6489 | * shutdown(8) now can send a (configurable) wall message when | |
6490 | a shutdown is cancelled. | |
6491 | ||
6492 | * The mount propagation mode for the root file system will now | |
6493 | default to "shared", which is useful to make containers work | |
6494 | nicely out-of-the-box so that they receive new mounts from | |
6495 | the host. This can be undone locally by running "mount | |
6496 | --make-rprivate /" if needed. | |
6497 | ||
6498 | * The prefdm.service file has been removed. Distributions | |
6499 | should maintain this unit downstream if they intend to keep | |
6500 | it around. However, we recommend writing normal unit files | |
6501 | for display managers instead. | |
6502 | ||
6503 | * Since systemd is a crucial part of the OS we will now | |
6504 | default to a number of compiler switches that improve | |
6505 | security (hardening) such as read-only relocations, stack | |
6506 | protection, and suchlike. | |
6507 | ||
6508 | * The TimeoutSec= setting for services is now split into | |
6509 | TimeoutStartSec= and TimeoutStopSec= to allow configuration | |
6510 | of individual time outs for the start and the stop phase of | |
6511 | the service. | |
6512 | ||
6513 | Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke | |
6514 | Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer, | |
6515 | Jim Meyering, Kay Sievers, Lennart Poettering, Mantas | |
6516 | Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter | |
6517 | Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom | |
6518 | Gundersen, Zbigniew Jędrzejewski-Szmek | |
6519 | ||
c4f1b862 LP |
6520 | CHANGES WITH 187: |
6521 | ||
6522 | * The journal and id128 C APIs are now fully documented as man | |
6523 | pages. | |
6524 | ||
6525 | * Extra safety checks have been added when transitioning from | |
6526 | the initial RAM disk to the main system to avoid accidental | |
6527 | data loss. | |
6528 | ||
c269cec3 | 6529 | * /etc/crypttab entries now understand the new keyfile-offset= |
c4f1b862 LP |
6530 | option. |
6531 | ||
6532 | * systemctl -t can now be used to filter by unit load state. | |
6533 | ||
6534 | * The journal C API gained the new sd_journal_wait() call to | |
6535 | make writing synchronous journal clients easier. | |
6536 | ||
6537 | * journalctl gained the new -D switch to show journals from a | |
6538 | specific directory. | |
6539 | ||
6540 | * journalctl now displays a special marker between log | |
6541 | messages of two different boots. | |
6542 | ||
6543 | * The journal is now explicitly flushed to /var via a service | |
6544 | systemd-journal-flush.service, rather than implicitly simply | |
6545 | by seeing /var/log/journal to be writable. | |
6546 | ||
6547 | * journalctl (and the journal C APIs) can now match for much | |
6548 | more complex expressions, with alternatives and | |
6549 | disjunctions. | |
6550 | ||
6551 | * When transitioning from the initial RAM disk to the main | |
6552 | system we will now kill all processes in a killing spree to | |
6553 | ensure no processes stay around by accident. | |
6554 | ||
6555 | * Three new specifiers may be used in unit files: %u, %h, %s | |
6556 | resolve to the user name, user home directory resp. user | |
6557 | shell. This is useful for running systemd user instances. | |
6558 | ||
6559 | * We now automatically rotate journal files if their data | |
6560 | object hash table gets a fill level > 75%. We also size the | |
6561 | hash table based on the configured maximum file size. This | |
6562 | together should lower hash collisions drastically and thus | |
6563 | speed things up a bit. | |
6564 | ||
6565 | * journalctl gained the new "--header" switch to introspect | |
6566 | header data of journal files. | |
6567 | ||
6568 | * A new setting SystemCallFilters= has been added to services | |
6569 | which may be used to apply blacklists or whitelists to | |
6570 | system calls. This is based on SECCOMP Mode 2 of Linux 3.5. | |
6571 | ||
6572 | * nspawn gained a new --link-journal= switch (and quicker: -j) | |
6573 | to link the container journal with the host. This makes it | |
6574 | very easy to centralize log viewing on the host for all | |
6575 | guests while still keeping the journal files separated. | |
6576 | ||
6577 | * Many bugfixes and optimizations | |
6578 | ||
6579 | Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay | |
6580 | Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex | |
6581 | Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew | |
6582 | Jędrzejewski-Szmek | |
6583 | ||
b5b4c94a LP |
6584 | CHANGES WITH 186: |
6585 | ||
6586 | * Several tools now understand kernel command line arguments, | |
6587 | which are only read when run in an initial RAM disk. They | |
6588 | usually follow closely their normal counterparts, but are | |
6589 | prefixed with rd. | |
6590 | ||
6591 | * There's a new tool to analyze the readahead files that are | |
6592 | automatically generated at boot. Use: | |
6593 | ||
6594 | /usr/lib/systemd/systemd-readahead analyze /.readahead | |
6595 | ||
6596 | * We now provide an early debug shell on tty9 if this enabled. Use: | |
6597 | ||
d1f9edaf | 6598 | systemctl enable debug-shell.service |
b5b4c94a LP |
6599 | |
6600 | * All plymouth related units have been moved into the Plymouth | |
6601 | package. Please make sure to upgrade your Plymouth version | |
6602 | as well. | |
6603 | ||
6604 | * systemd-tmpfiles now supports getting passed the basename of | |
6605 | a configuration file only, in which case it will look for it | |
6606 | in all appropriate directories automatically. | |
6607 | ||
6608 | * udevadm info now takes a /dev or /sys path as argument, and | |
6609 | does the right thing. Example: | |
6610 | ||
6611 | udevadm info /dev/sda | |
6612 | udevadm info /sys/class/block/sda | |
6613 | ||
6614 | * systemctl now prints a warning if a unit is stopped but a | |
6615 | unit that might trigger it continues to run. Example: a | |
6616 | service is stopped but the socket that activates it is left | |
6617 | running. | |
6618 | ||
6619 | * "systemctl status" will now mention if the log output was | |
6620 | shortened due to rotation since a service has been started. | |
6621 | ||
6622 | * The journal API now exposes functions to determine the | |
6623 | "cutoff" times due to rotation. | |
6624 | ||
6625 | * journald now understands SIGUSR1 and SIGUSR2 for triggering | |
6626 | immediately flushing of runtime logs to /var if possible, | |
6627 | resp. for triggering immediate rotation of the journal | |
6628 | files. | |
6629 | ||
6630 | * It is now considered an error if a service is attempted to | |
6631 | be stopped that is not loaded. | |
6632 | ||
6633 | * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames. | |
6634 | ||
6635 | * systemd-analyze now supports Python 3 | |
6636 | ||
6637 | * tmpfiles now supports cleaning up directories via aging | |
6638 | where the first level dirs are always kept around but | |
6639 | directories beneath it automatically aged. This is enabled | |
6640 | by prefixing the age field with '~'. | |
6641 | ||
6642 | * Seat objects now expose CanGraphical, CanTTY properties | |
6643 | which is required to deal with very fast bootups where the | |
6644 | display manager might be running before the graphics drivers | |
6645 | completed initialization. | |
6646 | ||
6647 | * Seat objects now expose a State property. | |
6648 | ||
6649 | * We now include RPM macros for service enabling/disabling | |
6650 | based on the preset logic. We recommend RPM based | |
6651 | distributions to make use of these macros if possible. This | |
6652 | makes it simpler to reuse RPM spec files across | |
6653 | distributions. | |
6654 | ||
6655 | * We now make sure that the collected systemd unit name is | |
6656 | always valid when services log to the journal via | |
6657 | STDOUT/STDERR. | |
6658 | ||
6659 | * There's a new man page kernel-command-line(7) detailing all | |
6660 | command line options we understand. | |
6661 | ||
6662 | * The fstab generator may now be disabled at boot by passing | |
6663 | fstab=0 on the kernel command line. | |
6664 | ||
91ac7425 | 6665 | * A new kernel command line option modules-load= is now understood |
b5b4c94a LP |
6666 | to load a specific kernel module statically, early at boot. |
6667 | ||
6668 | * Unit names specified on the systemctl command line are now | |
6669 | automatically escaped as needed. Also, if file system or | |
6670 | device paths are specified they are automatically turned | |
6671 | into the appropriate mount or device unit names. Example: | |
6672 | ||
6673 | systemctl status /home | |
6674 | systemctl status /dev/sda | |
6675 | ||
6676 | * The SysVConsole= configuration option has been removed from | |
6677 | system.conf parsing. | |
6678 | ||
6679 | * The SysV search path is no longer exported on the D-Bus | |
6680 | Manager object. | |
6681 | ||
ce830873 | 6682 | * The Names= option has been removed from unit file parsing. |
b5b4c94a LP |
6683 | |
6684 | * There's a new man page bootup(7) detailing the boot process. | |
6685 | ||
6686 | * Every unit and every generator we ship with systemd now | |
6687 | comes with full documentation. The self-explanatory boot is | |
6688 | complete. | |
6689 | ||
6690 | * A couple of services gained "systemd-" prefixes in their | |
6691 | name if they wrap systemd code, rather than only external | |
6692 | code. Among them fsck@.service which is now | |
6693 | systemd-fsck@.service. | |
6694 | ||
6695 | * The HaveWatchdog property has been removed from the D-Bus | |
6696 | Manager object. | |
6697 | ||
6698 | * systemd.confirm_spawn= on the kernel command line should now | |
6699 | work sensibly. | |
6700 | ||
6701 | * There's a new man page crypttab(5) which details all options | |
6702 | we actually understand. | |
6703 | ||
6704 | * systemd-nspawn gained a new --capability= switch to pass | |
6705 | additional capabilities to the container. | |
6706 | ||
6707 | * timedated will now read known NTP implementation unit names | |
5b00c016 | 6708 | from /usr/lib/systemd/ntp-units.d/*.list, |
b5b4c94a LP |
6709 | systemd-timedated-ntp.target has been removed. |
6710 | ||
6711 | * journalctl gained a new switch "-b" that lists log data of | |
6712 | the current boot only. | |
6713 | ||
6714 | * The notify socket is in the abstract namespace again, in | |
6715 | order to support daemons which chroot() at start-up. | |
6716 | ||
6717 | * There is a new Storage= configuration option for journald | |
6718 | which allows configuration of where log data should go. This | |
6719 | also provides a way to disable journal logging entirely, so | |
6720 | that data collected is only forwarded to the console, the | |
6721 | kernel log buffer or another syslog implementation. | |
6722 | ||
c4f1b862 | 6723 | * Many bugfixes and optimizations |
b5b4c94a | 6724 | |
2d938ac7 LP |
6725 | Contributions from: Auke Kok, Colin Guthrie, Dave Reisner, |
6726 | David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering, | |
6727 | Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel, | |
6728 | Shawn Landden, Tom Gundersen | |
b5b4c94a | 6729 | |
2d197285 | 6730 | CHANGES WITH 185: |
b6a86739 | 6731 | |
2d197285 KS |
6732 | * "systemctl help <unit>" now shows the man page if one is |
6733 | available. | |
6734 | ||
6735 | * Several new man pages have been added. | |
6736 | ||
b5b4c94a LP |
6737 | * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=, |
6738 | MaxLevelConsole= can now be specified in | |
6739 | journald.conf. These options allow reducing the amount of | |
6740 | data stored on disk or forwarded by the log level. | |
2d197285 | 6741 | |
b5b4c94a LP |
6742 | * TimerSlackNSec= can now be specified in system.conf for |
6743 | PID1. This allows system-wide power savings. | |
2d197285 KS |
6744 | |
6745 | Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen, | |
6746 | Lennart Poettering, Malte Starostik, Marc-Antoine Perennou, | |
6747 | Matthias Clasen | |
6748 | ||
4c8cd173 | 6749 | CHANGES WITH 184: |
b6a86739 | 6750 | |
4c8cd173 LP |
6751 | * logind is now capable of (optionally) handling power and |
6752 | sleep keys as well as the lid switch. | |
6753 | ||
6754 | * journalctl now understands the syntax "journalctl | |
6755 | /usr/bin/avahi-daemon" to get all log output of a specific | |
6756 | daemon. | |
6757 | ||
6758 | * CapabilityBoundingSet= in system.conf now also influences | |
6759 | the capability bound set of usermode helpers of the kernel. | |
6760 | ||
6761 | Contributions from: Daniel Drake, Daniel J. Walsh, Gert | |
6762 | Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers, | |
6763 | Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul | |
6764 | Menzel, Shawn Landden, Tero Roponen, Tom Gundersen | |
6765 | ||
ea5943d3 | 6766 | CHANGES WITH 183: |
b6a86739 | 6767 | |
187076d4 LP |
6768 | * Note that we skipped 139 releases here in order to set the |
6769 | new version to something that is greater than both udev's | |
6770 | and systemd's most recent version number. | |
6771 | ||
194bbe33 KS |
6772 | * udev: all udev sources are merged into the systemd source tree now. |
6773 | All future udev development will happen in the systemd tree. It | |
6774 | is still fully supported to use the udev daemon and tools without | |
6775 | systemd running, like in initramfs or other init systems. Building | |
6776 | udev though, will require the *build* of the systemd tree, but | |
ea5943d3 | 6777 | udev can be properly *run* without systemd. |
07cd4fc1 | 6778 | |
91cf7e5c | 6779 | * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles |
f13b388f KS |
6780 | should be used to create dead device nodes as workarounds for broken |
6781 | subsystems. | |
64661ee7 | 6782 | |
2d13da88 KS |
6783 | * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is |
6784 | no longer supported. udev_monitor_new_from_netlink() needs to be | |
6785 | used to subscribe to events. | |
6786 | ||
194bbe33 KS |
6787 | * udev: when udevd is started by systemd, processes which are left |
6788 | behind by forking them off of udev rules, are unconditionally cleaned | |
6789 | up and killed now after the event handling has finished. Services or | |
6790 | daemons must be started as systemd services. Services can be | |
ea5943d3 | 6791 | pulled-in by udev to get started, but they can no longer be directly |
194bbe33 KS |
6792 | forked by udev rules. |
6793 | ||
f13b388f KS |
6794 | * udev: the daemon binary is called systemd-udevd now and installed |
6795 | in /usr/lib/systemd/. Standalone builds or non-systemd systems need | |
6796 | to adapt to that, create symlink, or rename the binary after building | |
6797 | it. | |
6798 | ||
ea5943d3 | 6799 | * libudev no longer provides these symbols: |
c1959569 KS |
6800 | udev_monitor_from_socket() |
6801 | udev_queue_get_failed_list_entry() | |
6802 | udev_get_{dev,sys,run}_path() | |
ea5943d3 | 6803 | The versions number was bumped and symbol versioning introduced. |
c1959569 | 6804 | |
ea5943d3 | 6805 | * systemd-loginctl and systemd-journalctl have been renamed |
9ae9afce | 6806 | to loginctl and journalctl to match systemctl. |
18b754d3 KS |
6807 | |
6808 | * The config files: /etc/systemd/systemd-logind.conf and | |
6809 | /etc/systemd/systemd-journald.conf have been renamed to | |
6810 | logind.conf and journald.conf. Package updates should rename | |
6811 | the files to the new names on upgrade. | |
6812 | ||
ea5943d3 LP |
6813 | * For almost all files the license is now LGPL2.1+, changed |
6814 | from the previous GPL2.0+. Exceptions are some minor stuff | |
6815 | of udev (which will be changed to LGPL2.1 eventually, too), | |
6816 | and the MIT licensed sd-daemon.[ch] library that is suitable | |
6817 | to be used as drop-in files. | |
6818 | ||
6819 | * systemd and logind now handle system sleep states, in | |
49f43d5f | 6820 | particular suspending and hibernating. |
ea5943d3 LP |
6821 | |
6822 | * logind now implements a sleep/shutdown/idle inhibiting logic | |
6823 | suitable for a variety of uses. Soonishly Lennart will blog | |
6824 | about this in more detail. | |
6825 | ||
6826 | * var-run.mount and var-lock.mount are no longer provided | |
ce830873 | 6827 | (which previously bind mounted these directories to their new |
ea5943d3 LP |
6828 | places). Distributions which have not converted these |
6829 | directories to symlinks should consider stealing these files | |
6830 | from git history and add them downstream. | |
6831 | ||
6832 | * We introduced the Documentation= field for units and added | |
6833 | this to all our shipped units. This is useful to make it | |
3943231c | 6834 | easier to explore the boot and the purpose of the various |
ea5943d3 LP |
6835 | units. |
6836 | ||
6837 | * All smaller setup units (such as | |
6838 | systemd-vconsole-setup.service) now detect properly if they | |
6839 | are run in a container and are skipped when | |
6840 | appropriate. This guarantees an entirely noise-free boot in | |
6841 | Linux container environments such as systemd-nspawn. | |
6842 | ||
6843 | * A framework for implementing offline system updates is now | |
6844 | integrated, for details see: | |
c6749ba5 | 6845 | https://www.freedesktop.org/wiki/Software/systemd/SystemUpdates |
ea5943d3 LP |
6846 | |
6847 | * A new service type Type=idle is available now which helps us | |
6848 | avoiding ugly interleaving of getty output and boot status | |
6849 | messages. | |
6850 | ||
439d6dfd LP |
6851 | * There's now a system-wide CapabilityBoundingSet= option to |
6852 | globally reduce the set of capabilities for the | |
ea5943d3 LP |
6853 | system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO, |
6854 | CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or | |
6855 | even CAP_NET_ADMIN system-wide for secure systems. | |
6856 | ||
6857 | * There are now system-wide DefaultLimitXXX= options to | |
6858 | globally change the defaults of the various resource limits | |
6859 | for all units started by PID 1. | |
6860 | ||
6861 | * Harald Hoyer's systemd test suite has been integrated into | |
6862 | systemd which allows easy testing of systemd builds in qemu | |
6863 | and nspawn. (This is really awesome! Ask us for details!) | |
6864 | ||
3943231c LP |
6865 | * The fstab parser is now implemented as generator, not inside |
6866 | of PID 1 anymore. | |
ea5943d3 LP |
6867 | |
6868 | * systemctl will now warn you if .mount units generated from | |
6869 | /etc/fstab are out of date due to changes in fstab that | |
d28315e4 | 6870 | have not been read by systemd yet. |
ea5943d3 LP |
6871 | |
6872 | * systemd is now suitable for usage in initrds. Dracut has | |
6873 | already been updated to make use of this. With this in place | |
6874 | initrds get a slight bit faster but primarily are much | |
6875 | easier to introspect and debug since "systemctl status" in | |
6876 | the host system can be used to introspect initrd services, | |
6877 | and the journal from the initrd is kept around too. | |
6878 | ||
6879 | * systemd-delta has been added, a tool to explore differences | |
6880 | between user/admin configuration and vendor defaults. | |
6881 | ||
6882 | * PrivateTmp= now affects both /tmp and /var/tmp. | |
6883 | ||
6884 | * Boot time status messages are now much prettier and feature | |
6885 | proper english language. Booting up systemd has never been | |
6886 | so sexy. | |
6887 | ||
6888 | * Read-ahead pack files now include the inode number of all | |
6889 | files to pre-cache. When the inode changes the pre-caching | |
6890 | is not attempted. This should be nicer to deal with updated | |
6891 | packages which might result in changes of read-ahead | |
6892 | patterns. | |
6893 | ||
6894 | * We now temporaritly lower the kernel's read_ahead_kb variable | |
6895 | when collecting read-ahead data to ensure the kernel's | |
6896 | built-in read-ahead does not add noise to our measurements | |
6897 | of necessary blocks to pre-cache. | |
6898 | ||
6899 | * There's now RequiresMountsFor= to add automatic dependencies | |
6900 | for all mounts necessary for a specific file system path. | |
6901 | ||
6902 | * MountAuto= and SwapAuto= have been removed from | |
6903 | system.conf. Mounting file systems at boot has to take place | |
6904 | in systemd now. | |
6905 | ||
6906 | * nspawn now learned a new switch --uuid= to set the machine | |
6907 | ID on the command line. | |
6908 | ||
f8c0a2cb | 6909 | * nspawn now learned the -b switch to automatically search |
ea5943d3 LP |
6910 | for an init system. |
6911 | ||
6912 | * vt102 is now the default TERM for serial TTYs, upgraded from | |
6913 | vt100. | |
6914 | ||
6915 | * systemd-logind now works on VT-less systems. | |
6916 | ||
6917 | * The build tree has been reorganized. The individual | |
3943231c | 6918 | components now have directories of their own. |
ea5943d3 LP |
6919 | |
6920 | * A new condition type ConditionPathIsReadWrite= is now available. | |
6921 | ||
6922 | * nspawn learned the new -C switch to create cgroups for the | |
6923 | container in other hierarchies. | |
6924 | ||
6925 | * We now have support for hardware watchdogs, configurable in | |
6926 | system.conf. | |
6927 | ||
6928 | * The scheduled shutdown logic now has a public API. | |
6929 | ||
6930 | * We now mount /tmp as tmpfs by default, but this can be | |
6931 | masked and /etc/fstab can override it. | |
6932 | ||
d28315e4 | 6933 | * Since udisks does not make use of /media anymore we are not |
ea5943d3 LP |
6934 | mounting a tmpfs on it anymore. |
6935 | ||
6936 | * journalctl gained a new --local switch to only interleave | |
6937 | locally generated journal files. | |
6938 | ||
6939 | * We can now load the IMA policy at boot automatically. | |
6940 | ||
6941 | * The GTK tools have been split off into a systemd-ui. | |
6942 | ||
79849bf9 LP |
6943 | Contributions from: Andreas Schwab, Auke Kok, Ayan George, |
6944 | Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan | |
6945 | Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal, | |
6946 | Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers, | |
6947 | Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure, | |
6948 | Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim | |
6949 | A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal | |
6950 | Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn | |
6951 | Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom | |
6952 | Gundersen | |
6953 | ||
16f1239e | 6954 | CHANGES WITH 44: |
b6a86739 | 6955 | |
16f1239e LP |
6956 | * This is mostly a bugfix release |
6957 | ||
6958 | * Support optional initialization of the machine ID from the | |
6959 | KVM or container configured UUID. | |
6960 | ||
6961 | * Support immediate reboots with "systemctl reboot -ff" | |
6962 | ||
6963 | * Show /etc/os-release data in systemd-analyze output | |
6964 | ||
ab06eef8 | 6965 | * Many bugfixes for the journal, including endianness fixes and |
16f1239e LP |
6966 | ensuring that disk space enforcement works |
6967 | ||
ce830873 | 6968 | * sd-login.h is C++ compatible again |
16f1239e LP |
6969 | |
6970 | * Extend the /etc/os-release format on request of the Debian | |
6971 | folks | |
6972 | ||
6973 | * We now refuse non-UTF8 strings used in various configuration | |
d28315e4 | 6974 | and unit files. This is done to ensure we do not pass invalid |
16f1239e LP |
6975 | data over D-Bus or expose it elsewhere. |
6976 | ||
6977 | * Register Mimo USB Screens as suitable for automatic seat | |
6978 | configuration | |
6979 | ||
6980 | * Read SELinux client context from journal clients in a race | |
6981 | free fashion | |
6982 | ||
6983 | * Reorder configuration file lookup order. /etc now always | |
6984 | overrides /run in order to allow the administrator to always | |
b938cb90 | 6985 | and unconditionally override vendor-supplied or |
16f1239e LP |
6986 | automatically generated data. |
6987 | ||
6988 | * The various user visible bits of the journal now have man | |
6989 | pages. We still lack man pages for the journal API calls | |
6990 | however. | |
6991 | ||
6992 | * We now ship all man pages in HTML format again in the | |
6993 | tarball. | |
6994 | ||
6995 | Contributions from: Dave Reisner, Dirk Eibach, Frederic | |
6996 | Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti | |
6997 | Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry | |
6998 | Reding | |
6999 | ||
437b7dee | 7000 | CHANGES WITH 43: |
b6a86739 | 7001 | |
437b7dee LP |
7002 | * This is mostly a bugfix release |
7003 | ||
7004 | * systems lacking /etc/os-release are no longer supported. | |
7005 | ||
7006 | * Various functionality updates to libsystemd-login.so | |
7007 | ||
45afd519 | 7008 | * Track class of PAM logins to distinguish greeters from |
437b7dee LP |
7009 | normal user logins. |
7010 | ||
7011 | Contributions from: Kay Sievers, Lennart Poettering, Michael | |
7012 | Biebl | |
7013 | ||
204fa33c | 7014 | CHANGES WITH 42: |
b6a86739 | 7015 | |
204fa33c LP |
7016 | * This is an important bugfix release for v41. |
7017 | ||
7018 | * Building man pages is now optional which should be useful | |
7019 | for those building systemd from git but unwilling to install | |
7020 | xsltproc. | |
7021 | ||
7022 | * Watchdog support for supervising services is now usable. In | |
7023 | a future release support for hardware watchdogs | |
7024 | (i.e. /dev/watchdog) will be added building on this. | |
7025 | ||
7026 | * Service start rate limiting is now configurable and can be | |
7027 | turned off per service. When a start rate limit is hit a | |
7028 | reboot can automatically be triggered. | |
7029 | ||
7030 | * New CanReboot(), CanPowerOff() bus calls in systemd-logind. | |
7031 | ||
7032 | Contributions from: Benjamin Franzke, Bill Nottingham, | |
7033 | Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal | |
7034 | Schmidt, Michał Górny, Piotr Drąg | |
7035 | ||
e0d25329 | 7036 | CHANGES WITH 41: |
b6a86739 | 7037 | |
e0d25329 KS |
7038 | * The systemd binary is installed /usr/lib/systemd/systemd now; |
7039 | An existing /sbin/init symlink needs to be adapted with the | |
7040 | package update. | |
7041 | ||
b13df964 LP |
7042 | * The code that loads kernel modules has been ported to invoke |
7043 | libkmod directly, instead of modprobe. This means we do not | |
7044 | support systems with module-init-tools anymore. | |
7045 | ||
7046 | * Watchdog support is now already useful, but still not | |
7047 | complete. | |
7048 | ||
7049 | * A new kernel command line option systemd.setenv= is | |
7050 | understood to set system wide environment variables | |
7051 | dynamically at boot. | |
7052 | ||
e9c1ea9d | 7053 | * We now limit the set of capabilities of systemd-journald. |
ccd07a08 | 7054 | |
353e12c2 LP |
7055 | * We now set SIGPIPE to ignore by default, since it only is |
7056 | useful in shell pipelines, and has little use in general | |
7057 | code. This can be disabled with IgnoreSIPIPE=no in unit | |
7058 | files. | |
7059 | ||
b13df964 LP |
7060 | Contributions from: Benjamin Franzke, Kay Sievers, Lennart |
7061 | Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen, | |
7062 | William Douglas | |
7063 | ||
d26e4270 | 7064 | CHANGES WITH 40: |
b6a86739 | 7065 | |
d26e4270 LP |
7066 | * This is mostly a bugfix release |
7067 | ||
7068 | * We now expose the reason why a service failed in the | |
7069 | "Result" D-Bus property. | |
7070 | ||
7071 | * Rudimentary service watchdog support (will be completed over | |
7072 | the next few releases.) | |
7073 | ||
7074 | * When systemd forks off in order execute some service we will | |
7075 | now immediately changes its argv[0] to reflect which process | |
7076 | it will execute. This is useful to minimize the time window | |
7077 | with a generic argv[0], which makes bootcharts more useful | |
7078 | ||
b13df964 LP |
7079 | Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay |
7080 | Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt, | |
7081 | Mike Kazantsev, Ray Strode | |
7082 | ||
220a21d3 | 7083 | CHANGES WITH 39: |
b6a86739 | 7084 | |
220a21d3 LP |
7085 | * This is mostly a test release, but incorporates many |
7086 | bugfixes. | |
7087 | ||
7088 | * New systemd-cgtop tool to show control groups by their | |
7089 | resource usage. | |
7090 | ||
7091 | * Linking against libacl for ACLs is optional again. If | |
7092 | disabled, support tracking device access for active logins | |
7093 | goes becomes unavailable, and so does access to the user | |
7094 | journals by the respective users. | |
7095 | ||
7096 | * If a group "adm" exists, journal files are automatically | |
7097 | owned by them, thus allow members of this group full access | |
7098 | to the system journal as well as all user journals. | |
7099 | ||
7100 | * The journal now stores the SELinux context of the logging | |
7101 | client for all entries. | |
7102 | ||
7103 | * Add C++ inclusion guards to all public headers | |
7104 | ||
7105 | * New output mode "cat" in the journal to print only text | |
7106 | messages, without any meta data like date or time. | |
7107 | ||
7108 | * Include tiny X server wrapper as a temporary stop-gap to | |
7109 | teach XOrg udev display enumeration. This is used by display | |
7110 | managers such as gdm, and will go away as soon as XOrg | |
7111 | learned native udev hotplugging for display devices. | |
7112 | ||
7113 | * Add new systemd-cat tool for executing arbitrary programs | |
7114 | with STDERR/STDOUT connected to the journal. Can also act as | |
7115 | BSD logger replacement, and does so by default. | |
7116 | ||
7117 | * Optionally store all locally generated coredumps in the | |
7118 | journal along with meta data. | |
7119 | ||
7120 | * systemd-tmpfiles learnt four new commands: n, L, c, b, for | |
7121 | writing short strings to files (for usage for /sys), and for | |
7122 | creating symlinks, character and block device nodes. | |
7123 | ||
7124 | * New unit file option ControlGroupPersistent= to make cgroups | |
7125 | persistent, following the mechanisms outlined in | |
56cadcb6 | 7126 | https://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups |
220a21d3 LP |
7127 | |
7128 | * Support multiple local RTCs in a sane way | |
7129 | ||
7130 | * No longer monopolize IO when replaying readahead data on | |
7131 | rotating disks, since we might starve non-file-system IO to | |
7132 | death, since fanotify() will not see accesses done by blkid, | |
7133 | or fsck. | |
7134 | ||
d28315e4 | 7135 | * Do not show kernel threads in systemd-cgls anymore, unless |
220a21d3 LP |
7136 | requested with new -k switch. |
7137 | ||
7138 | Contributions from: Dan Horák, Kay Sievers, Lennart | |
7139 | Poettering, Michal Schmidt | |
7140 | ||
7141 | CHANGES WITH 38: | |
b6a86739 | 7142 | |
220a21d3 LP |
7143 | * This is mostly a test release, but incorporates many |
7144 | bugfixes. | |
7145 | ||
7146 | * The git repository moved to: | |
7147 | git://anongit.freedesktop.org/systemd/systemd | |
7148 | ssh://git.freedesktop.org/git/systemd/systemd | |
7149 | ||
7150 | * First release with the journal | |
7151 | http://0pointer.de/blog/projects/the-journal.html | |
7152 | ||
7153 | * The journal replaces both systemd-kmsg-syslogd and | |
7154 | systemd-stdout-bridge. | |
7155 | ||
7156 | * New sd_pid_get_unit() API call in libsystemd-logind | |
7157 | ||
7158 | * Many systemadm clean-ups | |
7159 | ||
7160 | * Introduce remote-fs-pre.target which is ordered before all | |
7161 | remote mounts and may be used to start services before all | |
7162 | remote mounts. | |
7163 | ||
7164 | * Added Mageia support | |
7165 | ||
7166 | * Add bash completion for systemd-loginctl | |
7167 | ||
7168 | * Actively monitor PID file creation for daemons which exit in | |
7169 | the parent process before having finished writing the PID | |
7170 | file in the daemon process. Daemons which do this need to be | |
7171 | fixed (i.e. PID file creation must have finished before the | |
7172 | parent exits), but we now react a bit more gracefully to them. | |
7173 | ||
7174 | * Add colourful boot output, mimicking the well-known output | |
7175 | of existing distributions. | |
7176 | ||
7177 | * New option PassCredentials= for socket units, for | |
7178 | compatibility with a recent kernel ABI breakage. | |
7179 | ||
7180 | * /etc/rc.local is now hooked in via a generator binary, and | |
7181 | thus will no longer act as synchronization point during | |
7182 | boot. | |
7183 | ||
7184 | * systemctl list-unit-files now supports --root=. | |
7185 | ||
7186 | * systemd-tmpfiles now understands two new commands: z, Z for | |
7187 | relabelling files according to the SELinux database. This is | |
7188 | useful to apply SELinux labels to specific files in /sys, | |
7189 | among other things. | |
7190 | ||
7191 | * Output of SysV services is now forwarded to both the console | |
7192 | and the journal by default, not only just the console. | |
7193 | ||
7194 | * New man pages for all APIs from libsystemd-login. | |
7195 | ||
ce830873 | 7196 | * The build tree got reorganized and the build system is a |
220a21d3 LP |
7197 | lot more modular allowing embedded setups to specifically |
7198 | select the components of systemd they are interested in. | |
7199 | ||
7200 | * Support for Linux systems lacking the kernel VT subsystem is | |
7201 | restored. | |
7202 | ||
7203 | * configure's --with-rootdir= got renamed to | |
7204 | --with-rootprefix= to follow the naming used by udev and | |
7205 | kmod | |
7206 | ||
d28315e4 | 7207 | * Unless specified otherwise we will now install to /usr instead |
220a21d3 LP |
7208 | of /usr/local by default. |
7209 | ||
7210 | * Processes with '@' in argv[0][0] are now excluded from the | |
7211 | final shut-down killing spree, following the logic explained | |
7212 | in: | |
56cadcb6 | 7213 | https://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons |
220a21d3 LP |
7214 | |
7215 | * All processes remaining in a service cgroup when we enter | |
7216 | the START or START_PRE states are now killed with | |
7217 | SIGKILL. That means it is no longer possible to spawn | |
7218 | background processes from ExecStart= lines (which was never | |
7219 | supported anyway, and bad style). | |
7220 | ||
7221 | * New PropagateReloadTo=/PropagateReloadFrom= options to bind | |
7222 | reloading of units together. | |
7223 | ||
4c8cd173 | 7224 | Contributions from: Bill Nottingham, Daniel J. Walsh, Dave |
220a21d3 LP |
7225 | Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay |
7226 | Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt, | |
7227 | Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef | |
7228 | Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek |