You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

7859 lines
398 KiB

  1. systemd System and Service Manager
  2. CHANGES WITH 236:
  3. * The modprobe.d/ drop-in for the bonding.ko kernel module introduced
  4. in v235 has been extended to also set the dummy.ko module option
  5. numdummies=0, preventing the kernel from automatically creating
  6. dummy0. All dummy interfaces must now be explicitly created.
  7. * Unknown '%' specifiers in configuration files are now rejected. This
  8. applies to units and tmpfiles.d configuration. Any percent characters
  9. that are followed by a letter or digit that are not supposed to be
  10. interpreted as the beginning of a specifier should be escaped by
  11. doubling ("%%"). (So "size=5%" is still accepted, as well as
  12. "size=5%,foo=bar", but not "LABEL=x%y%z" since %y and %z are not
  13. valid specifiers today.)
  14. * systemd-resolved now maintains a new dynamic
  15. /run/systemd/resolve/stub-resolv.conf compatibility file. It is
  16. recommended to make /etc/resolv.conf a symlink to it. This file
  17. points at the systemd-resolved stub DNS resolver and
  18. includes dynamically acquired search domains, achieving more correct
  19. DNS resolution by software that bypasses local DNS APIs such as NSS.
  20. * The "uaccess" udev tag has been dropped from /dev/kvm and
  21. /dev/dri/renderD*. These devices now have the 0666 permissions by
  22. default (but this may be changed at build-time). /dev/dri/renderD*
  23. will now be owned by the "render" group along with /dev/kfd.
  24. * "DynamicUser=yes" has been enabled for systemd-timesyncd.service,
  25. systemd-journal-gatewayd.service and
  26. systemd-journal-upload.service. This means "nss-systemd" must be
  27. enabled in /etc/nsswitch.conf to ensure the UIDs assigned to these
  28. services are resolved properly.
  29. * In /etc/fstab two new mount options are now understood:
  30. x-systemd.makefs and x-systemd.growfs. The former has the effect that
  31. the configured file system is formatted before it is mounted, the
  32. latter that the file system is resized to the full block device size
  33. after it is mounted (i.e. if the file system is smaller than the
  34. partition it resides on, it's grown). This is similar to the fsck
  35. logic in /etc/fstab, and pulls in systemd-makefs@.service and
  36. systemd-growfs@.service as necessary, similar to
  37. systemd-fsck@.service. Resizing is currently only supported on ext4
  38. and btrfs.
  39. * In systemd-networkd, the IPv6 RA logic now optionally may announce
  40. DNS server and domain information.
  41. * Support for the LUKS2 on-disk format for encrypted partitions has
  42. been added. This requires libcryptsetup2 during compilation and
  43. runtime.
  44. * The systemd --user instance will now signal "readiness" when its
  45. unit has been reached, instead of when the run queue ran
  46. empty for the first time.
  47. * Tmpfiles.d with user configuration are now also supported.
  48. systemd-tmpfiles gained a new --user switch, and snippets placed in
  49. ~/.config/user-tmpfiles.d/ and corresponding directories will be
  50. executed by systemd-tmpfiles --user running in the new
  51. systemd-tmpfiles-setup.service and systemd-tmpfiles-clean.service
  52. running in the user session.
  53. * Unit files and tmpfiles.d snippets learnt three new % specifiers:
  54. %S resolves to the top-level state directory (/var/lib for the system
  55. instance, $XDG_CONFIG_HOME for the user instance), %C resolves to the
  56. top-level cache directory (/var/cache for the system instance,
  57. $XDG_CACHE_HOME for the user instance), %L resolves to the top-level
  58. logs directory (/var/log for the system instance,
  59. $XDG_CONFIG_HOME/log/ for the user instance). This matches the
  60. existing %t specifier, that resolves to the top-level runtime
  61. directory (/run for the system instance, and $XDG_RUNTIME_DIR for the
  62. user instance).
  63. * journalctl learnt a new parameter --output-fields= for limiting the
  64. set of journal fields to output in verbose and JSON output modes.
  65. * systemd-timesyncd's configuration file gained a new option
  66. RootDistanceMaxSec= for setting the maximum root distance of servers
  67. it'll use, as well as the new options PollIntervalMinSec= and
  68. PollIntervalMaxSec= to tweak the minimum and maximum poll interval.
  69. * bootctl gained a new command "list" for listing all available boot
  70. menu items on systems that follow the boot loader specification.
  71. * systemctl gained a new --dry-run switch that shows what would be done
  72. instead of doing it, and is currently supported by the shutdown and
  73. sleep verbs.
  74. * ConditionSecurity= can now detect the TOMOYO security module.
  75. * Unit file [Install] sections are now also respected in unit drop-in
  76. files. This is intended to be used by drop-ins under /usr/lib/.
  77. * systemd-firstboot may now also set the initial keyboard mapping.
  78. * Udev "changed" events for devices which are exposed as systemd
  79. .device units are now propagated to units specified in
  80. ReloadPropagatedFrom= as reload requests.
  81. * If a udev device has a SYSTEMD_WANTS= property containing a systemd
  82. unit template name (i.e. a name in the form of 'foobar@.service',
  83. without the instance component between the '@' and - the '.'), then
  84. the escaped sysfs path of the device is automatically used as the
  85. instance.
  86. * SystemCallFilter= in unit files has been extended so that an "errno"
  87. can be specified individually for each system call. Example:
  88. SystemCallFilter=~uname:EILSEQ.
  89. * The cgroup delegation logic has been substantially updated. Delegate=
  90. now optionally takes a list of controllers (instead of a boolean, as
  91. before), which lists the controllers to delegate at least.
  92. * The networkd DHCPv6 client now implements the FQDN option (RFC 4704).
  93. * A new LogLevelMax= setting configures the maximum log level any
  94. process of the service may log at (i.e. anything with a lesser
  95. priority than what is specified is automatically dropped). A new
  96. LogExtraFields= setting allows configuration of additional journal
  97. fields to attach to all log records generated by any of the unit's
  98. processes.
  99. * New StandardInputData= and StandardInputText= settings along with the
  100. new option StandardInput=data may be used to configure textual or
  101. binary data that shall be passed to the executed service process via
  102. standard input, encoded in-line in the unit file.
  103. * StandardInput=, StandardOutput= and StandardError= may now be used to
  104. connect stdin/stdout/stderr of executed processes directly with a
  105. file or AF_UNIX socket in the file system, using the new "file:" option.
  106. * A new unit file option CollectMode= has been added, that allows
  107. tweaking the garbage collection logic for units. It may be used to
  108. tell systemd to garbage collect units that have failed automatically
  109. (normally it only GCs units that exited successfully). systemd-run
  110. and systemd-mount expose this new functionality with a new -G option.
  111. * "machinectl bind" may now be used to bind mount non-directories
  112. (i.e. regularfiles, devices, fifos, sockets).
  113. * systemd-analyze gained a new verb "calendar" for validating and
  114. testing calendar time specifications to use for OnCalendar= in timer
  115. units. Besides validating the expression it will calculate the next
  116. time the specified expression would elapse.
  117. * In addition to the pre-existing FailureAction= unit file setting
  118. there's now SuccessAction=, for configuring a shutdown action to
  119. execute when a unit completes successfully. This is useful in
  120. particular inside containers that shall terminate after some workload
  121. has been completed. Also, both options are now supported for all unit
  122. types, not just services.
  123. * networkds's IP rule support gained two new options
  124. IncomingInterface= and OutgoingInterface= for configuring the incoming
  125. and outgoing interfaces of configured rules. systemd-networkd also
  126. gained support for "vxcan" network devices.
  127. * networkd gained a new setting RequiredForOnline=, taking a
  128. boolean. If set, systemd-wait-online will take it into consideration
  129. when determining that the system is up, otherwise it will ignore the
  130. interface for this purpose.
  131. * The sd_notify() protocol gained support for a new operation: with
  132. FDSTOREREMOVE=1 file descriptors may be removed from the per-service
  133. store again, ahead of POLLHUP or POLLERR when they are removed
  134. anyway.
  135. * A new document has been added to the source tree, that
  136. documents the UID/GID range and assignment assumptions and
  137. requirements of systemd.
  138. * The watchdog device PID 1 will ping may now be configured through the
  139. WatchdogDevice= configuration file setting, or by setting the
  140. systemd.watchdog_service= kernel commandline option.
  141. * systemd-resolved's gained support for registering DNS-SD services on
  142. the local network using MulticastDNS. Services may either be
  143. registered by dropping in a .dnssd file in /etc/systemd/dnssd/ (or
  144. the same dir below /run, /usr/lib), or through its D-Bus API.
  145. * The sd_notify() protocol can now with EXTEND_TIMEOUT_USEC=microsecond
  146. extend the effective start, runtime, and stop time. The service must
  147. continue to send EXTEND_TIMEOUT_USEC within the period specified to
  148. prevent the service manager from making the service as timedout.
  149. * systemd-resolved's DNSSEC support gained support for RFC 8080
  150. (Ed25519 keys and signatures).
  151. * The systemd-resolve command line tool gained a new set of options
  152. --set-dns=, --set-domain=, --set-llmnr=, --set-mdns=, --set-dnssec=,
  153. --set-nta= and --revert to configure per-interface DNS configuration
  154. dynamically during runtime. It's useful for pushing DNS information
  155. into systemd-resolved from DNS hook scripts that various interface
  156. managing software supports (such as pppd).
  157. * systemd-nspawn gained a new --network-namespace-path= command line
  158. option, which may be used to make a container join an existing
  159. network namespace, by specifying a path to a "netns" file.
  160. Contributions from: Alan Jenkins, Alan Robertson, Alessandro Ghedini,
  161. Andrew Jeddeloh, Antonio Rojas, Ari, asavah, bleep_blop, Carsten
  162. Strotmann, Christian Brauner, Christian Hesse, Clinton Roy, Collin
  163. Eggert, Cong Wang, Daniel Black, Daniel Lockyer, Daniel Rusek, Dimitri
  164. John Ledkov, Dmitry Rozhkov, Dongsu Park, Edward A. James, Evgeny
  165. Vereshchagin, Florian Klink, Franck Bui, Gwendal Grignou, Hans de
  166. Goede, Harald Hoyer, Hristo Venev, Iago López Galeiras, Ikey Doherty,
  167. Jakub Wilk, Jérémy Rosen, Jiahui Xie, John Lin, José Bollo, Josef
  168. Andersson, juga0, Krzysztof Nowicki, Kyle Walker, Lars Karlitski, Lars
  169. Kellogg-Stedman, Lauri Tirkkonen, Lennart Poettering, Lubomir Rintel,
  170. Luca Bruno, Lucas Werkmeister, Lukáš Nykrýn, Lukáš Říha, Lukasz
  171. Rubaszewski, Maciej S. Szmigiero, Mantas Mikulėnas, Marcus Folkesson,
  172. Martin Steuer, Mathieu Trudel-Lapierre, Matija Skala,
  173. Matthias-Christian Ott, Max Resch, Michael Biebl, Michael Vogt, Michal
  174. Koutný, Michal Sekletar, Mike Gilbert, Muhammet Kara, Neil Brown, Olaf
  175. Hering, Ondrej Kozina, Patrik Flykt, Patryk Kocielnik, Peter Hutterer,
  176. Piotr Drąg, Razvan Cojocaru, Robin McCorkell, Roland Hieber, Saran
  177. Tunyasuvunakool, Sergey Ptashnick, Shawn Landden, Shuang Liu, Simon
  178. Arlott, Simon Peeters, Stanislav Angelovič, Stefan Agner, Susant
  179. Sahani, Sylvain Plantefève, Thomas Blume, Thomas Haller, Tiago Salem
  180. Herrmann, Tinu Weber, Tom Stellard, Topi Miettinen, Torsten Hilbrich,
  181. Vito Caputo, Vladislav Vishnyakov, WaLyong Cho, Yu Watanabe, Zbigniew
  182. Jędrzejewski-Szmek, Zeal Jagannatha
  183. — Berlin, 2017-12-14
  184. CHANGES WITH 235:
  185. * INCOMPATIBILITY: systemd-logind.service and other long-running
  186. services now run inside an IPv4/IPv6 sandbox, prohibiting them any IP
  187. communication with the outside. This generally improves security of
  188. the system, and is in almost all cases a safe and good choice, as
  189. these services do not and should not provide any network-facing
  190. functionality. However, systemd-logind uses the glibc NSS API to
  191. query the user database. This creates problems on systems where NSS
  192. is set up to directly consult network services for user database
  193. lookups. In particular, this creates incompatibilities with the
  194. "nss-nis" module, which attempts to directly contact the NIS/YP
  195. network servers it is configured for, and will now consistently
  196. fail. In such cases, it is possible to turn off IP sandboxing for
  197. systemd-logind.service (set IPAddressDeny= in its [Service] section
  198. to the empty string, via a .d/ unit file drop-in). Downstream
  199. distributions might want to update their nss-nis packaging to include
  200. such a drop-in snippet, accordingly, to hide this incompatibility
  201. from the user. Another option is to make use of glibc's nscd service
  202. to proxy such network requests through a privilege-separated, minimal
  203. local caching daemon, or to switch to more modern technologies such
  204. sssd, whose NSS hook-ups generally do not involve direct network
  205. access. In general, we think it's definitely time to question the
  206. implementation choices of nss-nis, i.e. whether it's a good idea
  207. today to embed a network-facing loadable module into all local
  208. processes that need to query the user database, including the most
  209. trivial and benign ones, such as "ls". For more details about
  210. IPAddressDeny= see below.
  211. * A new modprobe.d drop-in is now shipped by default that sets the
  212. bonding module option max_bonds=0. This overrides the kernel default,
  213. to avoid conflicts and ambiguity as to whether or not bond0 should be
  214. managed by systemd-networkd or not. This resolves multiple issues
  215. with bond0 properties not being applied, when bond0 is configured
  216. with systemd-networkd. Distributors may choose to not package this,
  217. however in that case users will be prevented from correctly managing
  218. bond0 interface using systemd-networkd.
  219. * systemd-analyze gained new verbs "get-log-level" and "get-log-target"
  220. which print the logging level and target of the system manager. They
  221. complement the existing "set-log-level" and "set-log-target" verbs
  222. used to change those values.
  223. * journald.conf gained a new boolean setting ReadKMsg= which defaults
  224. to on. If turned off kernel log messages will not be read by
  225. systemd-journald or included in the logs. It also gained a new
  226. setting LineMax= for configuring the maximum line length in
  227. STDOUT/STDERR log streams. The new default for this value is 48K, up
  228. from the previous hardcoded 2048.
  229. * A new unit setting RuntimeDirectoryPreserve= has been added, which
  230. allows more detailed control of what to do with a runtime directory
  231. configured with RuntimeDirectory= (i.e. a directory below /run or
  232. $XDG_RUNTIME_DIR) after a unit is stopped.
  233. * The RuntimeDirectory= setting for units gained support for creating
  234. deeper subdirectories below /run or $XDG_RUNTIME_DIR, instead of just
  235. one top-level directory.
  236. * Units gained new options StateDirectory=, CacheDirectory=,
  237. LogsDirectory= and ConfigurationDirectory= which are closely related
  238. to RuntimeDirectory= but manage per-service directories below
  239. /var/lib, /var/cache, /var/log and /etc. By making use of them it is
  240. possible to write unit files which when activated automatically gain
  241. properly owned service specific directories in these locations, thus
  242. making unit files self-contained and increasing compatibility with
  243. stateless systems and factory reset where /etc or /var are
  244. unpopulated at boot. Matching these new settings there's also
  245. StateDirectoryMode=, CacheDirectoryMode=, LogsDirectoryMode=,
  246. ConfigurationDirectoryMode= for configuring the access mode of these
  247. directories. These settings are particularly useful in combination
  248. with DynamicUser=yes as they provide secure, properly-owned,
  249. writable, and stateful locations for storage, excluded from the
  250. sandbox that such services live in otherwise.
  251. * Automake support has been removed from this release. systemd is now
  252. Meson-only.
  253. * systemd-journald will now aggressively cache client metadata during
  254. runtime, speeding up log write performance under pressure. This comes
  255. at a small price though: as much of the metadata is read
  256. asynchronously from /proc/ (and isn't implicitly attached to log
  257. datagrams by the kernel, like UID/GID/PID/SELinux are) this means the
  258. metadata stored alongside a log entry might be slightly
  259. out-of-date. Previously it could only be slightly newer than the log
  260. message. The time window is small however, and given that the kernel
  261. is unlikely to be improved anytime soon in this regard, this appears
  262. acceptable to us.
  263. * nss-myhostname/systemd-resolved will now by default synthesize an
  264. A/AAAA resource record for the "_gateway" hostname, pointing to the
  265. current default IP gateway. Previously it did that for the "gateway"
  266. name, hampering adoption, as some distributions wanted to leave that
  267. host name open for local use. The old behaviour may still be
  268. requested at build time.
  269. * systemd-networkd's [Address] section in .network files gained a new
  270. Scope= setting for configuring the IP address scope. The [Network]
  271. section gained a new boolean setting ConfigureWithoutCarrier= that
  272. tells systemd-networkd to ignore link sensing when configuring the
  273. device. The [DHCP] section gained a new Anonymize= boolean option for
  274. turning on a number of options suggested in RFC 7844. A new
  275. [RoutingPolicyRule] section has been added for configuring the IP
  276. routing policy. The [Route] section has gained support for a new
  277. Type= setting which permits configuring
  278. blackhole/unreachable/prohibit routes.
  279. * The [VRF] section in .netdev files gained a new Table= setting for
  280. configuring the routing table to use. The [Tunnel] section gained a
  281. new Independent= boolean field for configuring tunnels independent of
  282. an underlying network interface. The [Bridge] section gained a new
  283. GroupForwardMask= option for configuration of propagation of link
  284. local frames between bridge ports.
  285. * The WakeOnLan= setting in .link files gained support for a number of
  286. new modes. A new TCP6SegmentationOffload= setting has been added for
  287. configuring TCP/IPv6 hardware segmentation offload.
  288. * The IPv6 RA sender implementation may now optionally send out RDNSS
  289. and RDNSSL records to supply DNS configuration to peers.
  290. * systemd-nspawn gained support for a new --system-call-filter= command
  291. line option for adding and removing entries in the default system
  292. call filter it applies. Moreover systemd-nspawn has been changed to
  293. implement a system call whitelist instead of a blacklist.
  294. * systemd-run gained support for a new --pipe command line option. If
  295. used the STDIN/STDOUT/STDERR file descriptors passed to systemd-run
  296. are directly passed on to the activated transient service
  297. executable. This allows invoking arbitrary processes as systemd
  298. services (for example to take benefit of dependency management,
  299. accounting management, resource management or log management that is
  300. done automatically for services) — while still allowing them to be
  301. integrated in a classic UNIX shell pipeline.
  302. * When a service sends RELOAD=1 via sd_notify() and reload propagation
  303. using ReloadPropagationTo= is configured, a reload is now propagated
  304. to configured units. (Previously this was only done on explicitly
  305. requested reloads, using "systemctl reload" or an equivalent
  306. command.)
  307. * For each service unit a restart counter is now kept: it is increased
  308. each time the service is restarted due to Restart=, and may be
  309. queried using "systemctl show -p NRestarts …".
  310. * New system call filter groups @aio, @sync, @chown, @setuid, @memlock,
  311. @signal and @timer have been added, for usage with SystemCallFilter=
  312. in unit files and the new --system-call-filter= command line option
  313. of systemd-nspawn (see above).
  314. * ExecStart= lines in unit files gained two new modifiers: when a
  315. command line is prefixed with "!" the command will be executed as
  316. configured, except for the credentials applied by
  317. setuid()/setgid()/setgroups(). It is very similar to the pre-existing
  318. "+", but does still apply namespacing options unlike "+". There's
  319. also "!!" now, which is mostly identical, but becomes a NOP on
  320. systems that support ambient capabilities. This is useful to write
  321. unit files that work with ambient capabilities where possible but
  322. automatically fall back to traditional privilege dropping mechanisms
  323. on systems where this is not supported.
  324. * ListenNetlink= settings in socket units now support RDMA netlink
  325. sockets.
  326. * A new unit file setting LockPersonality= has been added which permits
  327. locking down the chosen execution domain ("personality") of a service
  328. during runtime.
  329. * A new special target "" has been added, which is
  330. ordered before all text logins, and may be used to order services
  331. before textual logins acquire access to the console.
  332. * systemd will now attempt to load the virtio-rng.ko kernel module very
  333. early on if a VM environment supporting this is detected. This should
  334. improve entropy during early boot in virtualized environments.
  335. * A _netdev option is now supported in /etc/crypttab that operates in a
  336. similar way as the same option in /etc/fstab: it permits configuring
  337. encrypted devices that need to be ordered after the network is up.
  338. Following this logic, two new special targets
  339. and have been
  340. added that are to what and
  341. are to
  342. * Service units gained a new UnsetEnvironment= setting which permits
  343. unsetting specific environment variables for services that are
  344. normally passed to it (for example in order to mask out locale
  345. settings for specific services that can't deal with it).
  346. * Units acquired a new boolean option IPAccounting=. When turned on, IP
  347. traffic accounting (packet count as well as byte count) is done for
  348. the service, and shown as part of "systemctl status" or "systemd-run
  349. --wait".
  350. * Service units acquired two new options IPAddressAllow= and
  351. IPAddressDeny=, taking a list of IPv4 or IPv6 addresses and masks,
  352. for configuring a simple IP access control list for all sockets of
  353. the unit. These options are available also on .slice and .socket
  354. units, permitting flexible access list configuration for individual
  355. services as well as groups of services (as defined by a slice unit),
  356. including system-wide. Note that IP ACLs configured this way are
  357. enforced on every single IPv4 and IPv6 socket created by any process
  358. of the service unit, and apply to ingress as well as egress traffic.
  359. * If CPUAccounting= or IPAccounting= is turned on for a unit a new
  360. structured log message is generated each time the unit is stopped,
  361. containing information about the consumed resources of this
  362. invocation.
  363. * A new setting KeyringMode= has been added to unit files, which may be
  364. used to control how the kernel keyring is set up for executed
  365. processes.
  366. * "systemctl poweroff", "systemctl reboot", "systemctl halt",
  367. "systemctl kexec" and "systemctl exit" are now always asynchronous in
  368. behaviour (that is: these commands return immediately after the
  369. operation was enqueued instead of waiting for the operation to
  370. complete). Previously, "systemctl poweroff" and "systemctl reboot"
  371. were asynchronous on systems using systemd-logind (i.e. almost
  372. always, and like they were on sysvinit), and the other three commands
  373. were unconditionally synchronous. With this release this is cleaned
  374. up, and callers will see the same asynchronous behaviour on all
  375. systems for all five operations.
  376. * systemd-logind gained new Halt() and CanHalt() bus calls for halting
  377. the system.
  378. * .timer units now accept calendar specifications in other timezones
  379. than UTC or the local timezone.
  380. * The tmpfiles snippet var.conf has been changed to create
  381. /var/log/btmp with access mode 0660 instead of 0600. It was owned by
  382. the "utmp" group already, and it appears to be generally understood
  383. that members of "utmp" can modify/flush the utmp/wtmp/lastlog/btmp
  384. databases. Previously this was implemented correctly for all these
  385. databases excepts btmp, which has been opened up like this now
  386. too. Note that while the other databases are world-readable
  387. (i.e. 0644), btmp is not and remains more restrictive.
  388. * The systemd-resolve tool gained a new --reset-server-features
  389. switch. When invoked like this systemd-resolved will forget
  390. everything it learnt about the features supported by the configured
  391. upstream DNS servers, and restarts the feature probing logic on the
  392. next resolver look-up for them at the highest feature level
  393. again.
  394. * The status dump systemd-resolved sends to the logs upon receiving
  395. SIGUSR1 now also includes information about all DNS servers it is
  396. configured to use, and the features levels it probed for them.
  397. Contributions from: Abdó Roig-Maranges, Alan Jenkins, Alexander
  398. Kuleshov, Andreas Rammhold, Andrew Jeddeloh, Andrew Soutar, Ansgar
  399. Burchardt, Beniamino Galvani, Benjamin Berg, Benjamin Robin, Charles
  400. Huber, Christian Hesse, Daniel Berrange, Daniel Kahn Gillmor, Daniel
  401. Mack, Daniel Rusek, Daniel Șerbănescu, Davide Cavalca, Dimitri John
  402. Ledkov, Diogo Pereira, Djalal Harouni, Dmitriy Geels, Dmitry Torokhov,
  403. ettavolt, Evgeny Vereshchagin, Fabio Kung, Felipe Sateler, Franck Bui,
  404. Hans de Goede, Harald Hoyer, Insun Pyo, Ivan Kurnosov, Ivan Shapovalov,
  405. Jakub Wilk, Jan Synacek, Jason Gunthorpe, Jeremy Bicha, Jérémy Rosen,
  406. John Lin, jonasBoss, Jonathan Lebon, Jonathan Teh, Jon Ringle, Jörg
  407. Thalheim, Jouke Witteveen, juga0, Justin Capella, Justin Michaud,
  408. Kai-Heng Feng, Lennart Poettering, Lion Yang, Luca Bruno, Lucas
  409. Werkmeister, Lukáš Nykrýn, Marcel Hollerbach, Marcus Lundblad, Martin
  410. Pitt, Michael Biebl, Michael Grzeschik, Michal Sekletar, Mike Gilbert,
  411. Neil Brown, Nicolas Iooss, Patrik Flykt, pEJipE, Piotr Drąg, Russell
  412. Stuart, S. Fan, Shengyao Xue, Stefan Pietsch, Susant Sahani, Tejun Heo,
  413. Thomas Miller, Thomas Sailer, Tobias Hunger, Tomasz Pala, Tom
  414. Gundersen, Tommi Rantala, Topi Miettinen, Torstein Husebø, userwithuid,
  415. Vasilis Liaskovitis, Vito Caputo, WaLyong Cho, William Douglas, Xiang
  416. Fan, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
  417. — Berlin, 2017-10-06
  418. CHANGES WITH 234:
  419. * Meson is now supported as build system in addition to Automake. It is
  420. our plan to remove Automake in one of our next releases, so that
  421. Meson becomes our exclusive build system. Hence, please start using
  422. the Meson build system in your downstream packaging. There's plenty
  423. of documentation around how to use Meson, the extremely brief
  424. summary:
  425. ./ && ./configure && make && sudo make install
  426. becomes:
  427. meson build && ninja -C build && sudo ninja -C build install
  428. * Unit files gained support for a new JobRunningTimeoutUSec= setting,
  429. which permits configuring a timeout on the time a job is
  430. running. This is particularly useful for setting timeouts on jobs for
  431. .device units.
  432. * Unit files gained two new options ConditionUser= and ConditionGroup=
  433. for conditionalizing units based on the identity of the user/group
  434. running a systemd user instance.
  435. * systemd-networkd now understands a new FlowLabel= setting in the
  436. [VXLAN] section of .network files, as well as a Priority= in
  437. [Bridge], GVRP= + MVRP= + LooseBinding= + ReorderHeader= in [VLAN]
  438. and GatewayOnlink= + IPv6Preference= + Protocol= in [Route]. It also
  439. gained support for configuration of GENEVE links, and IPv6 address
  440. labels. The [Network] section gained the new IPv6ProxyNDP= setting.
  441. * .link files now understand a new Port= setting.
  442. * systemd-networkd's DHCP support gained support for DHCP option 119
  443. (domain search list).
  444. * systemd-networkd gained support for serving IPv6 address ranges using
  445. the Router Advertisment protocol. The new .network configuration
  446. section [IPv6Prefix] may be used to configure the ranges to
  447. serve. This is implemented based on a new, minimal, native server
  448. implementation of RA.
  449. * journalctl's --output= switch gained support for a new parameter
  450. "short-iso-precise" for a mode where timestamps are shown as precise
  451. ISO date values.
  452. * systemd-udevd's "net_id" builtin may now generate stable network
  453. interface names from IBM PowerVM VIO devices as well as ACPI platform
  454. devices.
  455. * MulticastDNS support in systemd-resolved may now be explicitly
  456. enabled/disabled using the new MulticastDNS= configuration file
  457. option.
  458. * systemd-resolved may now optionally use libidn2 instead of the libidn
  459. for processing internationalized domain names. Support for libidn2
  460. should be considered experimental and should not be enabled by
  461. default yet.
  462. * "machinectl pull-tar" and related call may now do verification of
  463. downloaded images using SUSE-style .sha256 checksum files in addition
  464. to the already existing support for validating using Ubuntu-style
  465. SHA256SUMS files.
  466. * sd-bus gained support for a new sd_bus_message_appendv() call which
  467. is va_list equivalent of sd_bus_message_append().
  468. * sd-boot gained support for validating images using SHIM/MOK.
  469. * The SMACK code learnt support for "onlycap".
  470. * systemd-mount --umount is now much smarter in figuring out how to
  471. properly unmount a device given its mount or device path.
  472. * The code to call libnss_dns as a fallback from libnss_resolve when
  473. the communication with systemd-resolved fails was removed. This
  474. fallback was redundant and interfered with the [!UNAVAIL=return]
  475. suffix. See nss-resolve(8) for the recommended configuration.
  476. * systemd-logind may now be restarted without losing state. It stores
  477. the file descriptors for devices it manages in the system manager
  478. using the FDSTORE= mechanism. Please note that further changes in
  479. other components may be required to make use of this (for example
  480. Xorg has code to listen for stops of systemd-logind and terminate
  481. itself when logind is stopped or restarted, in order to avoid using
  482. stale file descriptors for graphical devices, which is now
  483. counterproductive and must be reverted in order for restarts of
  484. systemd-logind to be safe. See
  486. * All kernel install plugins are called with the environment variable
  487. KERNEL_INSTALL_MACHINE_ID which is set to the machine ID given by
  488. /etc/machine-id. If the file is missing or empty, the variable is
  489. empty and BOOT_DIR_ABS is the path of a temporary directory which is
  490. removed after all the plugins exit. So, if KERNEL_INSTALL_MACHINE_ID
  491. is empty, all plugins should not put anything in BOOT_DIR_ABS.
  492. Contributions from: Adrian Heine né Lang, Aggelos Avgerinos, Alexander
  493. Kurtz, Alexandros Frantzis, Alexey Brodkin, Alex Lu, Amir Pakdel, Amir
  494. Yalon, Anchor Cat, Anthony Parsons, Bastien Nocera, Benjamin Gilbert,
  495. Benjamin Robin, Boucman, Charles Plessy, Chris Chiu, Chris Lamb,
  496. Christian Brauner, Christian Hesse, Colin Walters, Daniel Drake,
  497. Danielle Church, Daniel Molkentin, Daniel Rusek, Daniel Wang, Davide
  498. Cavalca, David Herrmann, David Michael, Dax Kelson, Dimitri John
  499. Ledkov, Djalal Harouni, Dušan Kazik, Elias Probst, Evgeny Vereshchagin,
  500. Federico Di Pierro, Felipe Sateler, Felix Zhang, Franck Bui, Gary
  501. Tierney, George McCollister, Giedrius Statkevičius, Hans de Goede,
  502. hecke, Hendrik Westerberg, Hristo Venev, Ian Wienand, Insun Pyo, Ivan
  503. Shapovalov, James Cowgill, James Hemsing, Janne Heß, Jan Synacek, Jason
  504. Reeder, João Paulo Rechi Vita, John Paul Adrian Glaubitz, Jörg
  505. Thalheim, Josef Andersson, Josef Gajdusek, Julian Mehne, Kai Krakow,
  506. Krzysztof Jackiewicz, Lars Karlitski, Lennart Poettering, Lluís Gili,
  507. Lucas Werkmeister, Lukáš Nykrýn, Łukasz Stelmach, Mantas Mikulėnas,
  508. Marcin Bachry, Marcus Cooper, Mark Stosberg, Martin Pitt, Matija Skala,
  509. Matt Clarkson, Matthew Garrett, Matthias Greiner, Matthijs van Duin,
  510. Max Resch, Michael Biebl, Michal Koutný, Michal Sekletar, Michal
  511. Soltys, Michal Suchanek, Mike Gilbert, Nate Clark, Nathaniel R. Lewis,
  512. Neil Brown, Nikolai Kondrashov, Pascal S. de Kloe, Pat Riehecky, Patrik
  513. Flykt, Paul Kocialkowski, Peter Hutterer, Philip Withnall, Piotr
  514. Szydełko, Rafael Fontenelle, Ray Strode, Richard Maw, Roelf Wichertjes,
  515. Ronny Chevalier, Sarang S. Dalal, Sjoerd Simons, slodki, Stefan
  516. Schweter, Susant Sahani, Ted Wood, Thomas Blume, Thomas Haller, Thomas
  517. H. P. Andersen, Timothée Ravier, Tobias Jungel, Tobias Stoeckmann, Tom
  518. Gundersen, Tom Yan, Torstein Husebø, Umut Tezduyar Lindskog,
  519. userwithuid, Vito Caputo, Waldemar Brodkorb, WaLyong Cho, Yu, Li-Yu,
  520. Yusuke Nojima, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Дамјан
  521. Георгиевски
  522. — Berlin, 2017-07-12
  523. CHANGES WITH 233:
  524. * The "hybrid" control group mode has been modified to improve
  525. compatibility with "legacy" cgroups-v1 setups. Specifically, the
  526. "hybrid" setup of /sys/fs/cgroup is now pretty much identical to
  527. "legacy" (including /sys/fs/cgroup/systemd as "name=systemd" named
  528. cgroups-v1 hierarchy), the only externally visible change being that
  529. the cgroups-v2 hierarchy is also mounted, to
  530. /sys/fs/cgroup/unified. This should provide a large degree of
  531. compatibility with "legacy" cgroups-v1, while taking benefit of the
  532. better management capabilities of cgroups-v2.
  533. * The default control group setup mode may be selected both a boot-time
  534. via a set of kernel command line parameters (specifically:
  535. systemd.unified_cgroup_hierarchy= and
  536. systemd.legacy_systemd_cgroup_controller=), as well as a compile-time
  537. default selected on the configure command line
  538. (--with-default-hierarchy=). The upstream default is "hybrid"
  539. (i.e. the cgroups-v1 + cgroups-v2 mixture discussed above) now, but
  540. this will change in a future systemd version to be "unified" (pure
  541. cgroups-v2 mode). The third option for the compile time option is
  542. "legacy", to enter pure cgroups-v1 mode. We recommend downstream
  543. distributions to default to "hybrid" mode for release distributions,
  544. starting with v233. We recommend "unified" for development
  545. distributions (specifically: distributions such as Fedora's rawhide)
  546. as that's where things are headed in the long run. Use "legacy" for
  547. greatest stability and compatibility only.
  548. * Note one current limitation of "unified" and "hybrid" control group
  549. setup modes: the kernel currently does not permit the systemd --user
  550. instance (i.e. unprivileged code) to migrate processes between two
  551. disconnected cgroup subtrees, even if both are managed and owned by
  552. the user. This effectively means "systemd-run --user --scope" doesn't
  553. work when invoked from outside of any "systemd --user" service or
  554. scope. Specifically, it is not supported from session scopes. We are
  555. working on fixing this in a future systemd version. (See #3388 for
  556. further details about this.)
  557. * DBus policy files are now installed into /usr rather than /etc. Make
  558. sure your system has dbus >= 1.9.18 running before upgrading to this
  559. version, or override the install path with --with-dbuspolicydir= .
  560. * All python scripts shipped with systemd (specifically: the various
  561. tests written in Python) now require Python 3.
  562. * systemd unit tests can now run standalone (without the source or
  563. build directories), and can be installed into /usr/lib/systemd/tests/
  564. with 'make install-tests'.
  565. * Note that from this version on, CONFIG_CRYPTO_USER_API_HASH,
  566. CONFIG_CRYPTO_HMAC and CONFIG_CRYPTO_SHA256 need to be enabled in the
  567. kernel.
  568. * Support for the %c, %r, %R specifiers in unit files has been
  569. removed. Specifiers are not supposed to be dependent on configuration
  570. in the unit file itself (so that they resolve the same regardless
  571. where used in the unit files), but these specifiers were influenced
  572. by the Slice= option.
  573. * The shell invoked by debug-shell.service now defaults to /bin/sh in
  574. all cases. If distributions want to use a different shell for this
  575. purpose (for example Fedora's /sbin/sushell) they need to specify
  576. this explicitly at configure time using --with-debug-shell=.
  577. * The confirmation spawn prompt has been reworked to offer the
  578. following choices:
  579. (c)ontinue, proceed without asking anymore
  580. (D)ump, show the state of the unit
  581. (f)ail, don't execute the command and pretend it failed
  582. (h)elp
  583. (i)nfo, show a short summary of the unit
  584. (j)obs, show jobs that are in progress
  585. (s)kip, don't execute the command and pretend it succeeded
  586. (y)es, execute the command
  587. The 'n' choice for the confirmation spawn prompt has been removed,
  588. because its meaning was confusing.
  589. The prompt may now also be redirected to an alternative console by
  590. specifying the console as parameter to systemd.confirm_spawn=.
  591. * Services of Type=notify require a READY=1 notification to be sent
  592. during startup. If no such message is sent, the service now fails,
  593. even if the main process exited with a successful exit code.
  594. * Services that fail to start up correctly now always have their
  595. ExecStopPost= commands executed. Previously, they'd enter "failed"
  596. state directly, without executing these commands.
  597. * The option MulticastDNS= of network configuration files has acquired
  598. an actual implementation. With MulticastDNS=yes a host can resolve
  599. names of remote hosts and reply to mDNS A and AAAA requests.
  600. * When units are about to be started an additional check is now done to
  601. ensure that all dependencies of type BindsTo= (when used in
  602. combination with After=) have been started.
  603. * systemd-analyze gained a new verb "syscall-filter" which shows which
  604. system call groups are defined for the SystemCallFilter= unit file
  605. setting, and which system calls they contain.
  606. * A new system call filter group "@filesystem" has been added,
  607. consisting of various file system related system calls. Group
  608. "@reboot" has been added, covering reboot, kexec and shutdown related
  609. calls. Finally, group "@swap" has been added covering swap
  610. configuration related calls.
  611. * A new unit file option RestrictNamespaces= has been added that may be
  612. used to restrict access to the various process namespace types the
  613. Linux kernel provides. Specifically, it may be used to take away the
  614. right for a service unit to create additional file system, network,
  615. user, and other namespaces. This sandboxing option is particularly
  616. relevant due to the high amount of recently discovered namespacing
  617. related vulnerabilities in the kernel.
  618. * systemd-udev's .link files gained support for a new AutoNegotiation=
  619. setting for configuring Ethernet auto-negotiation.
  620. * systemd-networkd's .network files gained support for a new
  621. ListenPort= setting in the [DHCP] section to explicitly configure the
  622. UDP client port the DHCP client shall listen on.
  623. * .network files gained a new Unmanaged= boolean setting for explicitly
  624. excluding one or more interfaces from management by systemd-networkd.
  625. * The systemd-networkd ProxyARP= option has been renamed to
  626. IPV4ProxyARP=. Similarly, VXLAN-specific option ARPProxy= has been
  627. renamed to ReduceARPProxy=. The old names continue to be available
  628. for compatibility.
  629. * systemd-networkd gained support for configuring IPv6 Proxy NDP
  630. addresses via the new IPv6ProxyNDPAddress= .network file setting.
  631. * systemd-networkd's bonding device support gained support for two new
  632. configuration options ActiveSlave= and PrimarySlave=.
  633. * The various options in the [Match] section of .network files gained
  634. support for negative matching.
  635. * New systemd-specific mount options are now understood in /etc/fstab:
  636. x-systemd.mount-timeout= may be used to configure the maximum
  637. permitted runtime of the mount command.
  638. x-systemd.device-bound may be set to bind a mount point to its
  639. backing device unit, in order to automatically remove a mount point
  640. if its backing device is unplugged. This option may also be
  641. configured through the new SYSTEMD_MOUNT_DEVICE_BOUND udev property
  642. on the block device, which is now automatically set for all CDROM
  643. drives, so that mounted CDs are automatically unmounted when they are
  644. removed from the drive.
  645. x-systemd.after= and x-systemd.before= may be used to explicitly
  646. order a mount after or before another unit or mount point.
  647. * Enqueued start jobs for device units are now automatically garbage
  648. collected if there are no jobs waiting for them anymore.
  649. * systemctl list-jobs gained two new switches: with --after, for every
  650. queued job the jobs it's waiting for are shown; with --before the
  651. jobs which it's blocking are shown.
  652. * systemd-nspawn gained support for ephemeral boots from disk images
  653. (or in other words: --ephemeral and --image= may now be
  654. combined). Moreover, ephemeral boots are now supported for normal
  655. directories, even if the backing file system is not btrfs. Of course,
  656. if the file system does not support file system snapshots or
  657. reflinks, the initial copy operation will be relatively expensive, but
  658. this should still be suitable for many use cases.
  659. * Calendar time specifications in .timer units now support
  660. specifications relative to the end of a month by using "~" instead of
  661. "-" as separator between month and day. For example, "*-02~03" means
  662. "the third last day in February". In addition a new syntax for
  663. repeated events has been added using the "/" character. For example,
  664. "9..17/2:00" means "every two hours from 9am to 5pm".
  665. * systemd-socket-proxyd gained a new parameter --connections-max= for
  666. configuring the maximum number of concurrent connections.
  667. * sd-id128 gained a new API for generating unique IDs for the host in a
  668. way that does not leak the machine ID. Specifically,
  669. sd_id128_get_machine_app_specific() derives an ID based on the
  670. machine ID a in well-defined, non-reversible, stable way. This is
  671. useful whenever an identifier for the host is needed but where the
  672. identifier shall not be useful to identify the system beyond the
  673. scope of the application itself. (Internally this uses HMAC-SHA256 as
  674. keyed hash function using the machine ID as input.)
  675. * NotifyAccess= gained a new supported value "exec". When set
  676. notifications are accepted from all processes systemd itself invoked,
  677. including all control processes.
  678. * .nspawn files gained support for defining overlay mounts using the
  679. Overlay= and OverlayReadOnly= options. Previously this functionality
  680. was only available on the systemd-nspawn command line.
  681. * systemd-nspawn's --bind= and --overlay= options gained support for
  682. bind/overlay mounts whose source lies within the container tree by
  683. prefixing the source path with "+".
  684. * systemd-nspawn's --bind= and --overlay= options gained support for
  685. automatically allocating a temporary source directory in /var/tmp
  686. that is removed when the container dies. Specifically, if the source
  687. directory is specified as empty string this mechanism is selected. An
  688. example usage is --overlay=+/var::/var, which creates an overlay
  689. mount based on the original /var contained in the image, overlayed
  690. with a temporary directory in the host's /var/tmp. This way changes
  691. to /var are automatically flushed when the container shuts down.
  692. * systemd-nspawn --image= option does now permit raw file system block
  693. devices (in addition to images containing partition tables, as
  694. before).
  695. * The disk image dissection logic in systemd-nspawn gained support for
  696. automatically setting up LUKS encrypted as well as Verity protected
  697. partitions. When a container is booted from an encrypted image the
  698. passphrase is queried at start-up time. When a container with Verity
  699. data is started, the root hash is search in a ".roothash" file
  700. accompanying the disk image (alternatively, pass the root hash via
  701. the new --root-hash= command line option).
  702. * A new tool /usr/lib/systemd/systemd-dissect has been added that may
  703. be used to dissect disk images the same way as systemd-nspawn does
  704. it, following the Bootable Partition Specification. It may even be
  705. used to mount disk images with complex partition setups (including
  706. LUKS and Verity partitions) to a local host directory, in order to
  707. inspect them. This tool is not considered public API (yet), and is
  708. thus not installed into /usr/bin. Please do not rely on its
  709. existence, since it might go away or be changed in later systemd
  710. versions.
  711. * A new generator "systemd-verity-generator" has been added, similar in
  712. style to "systemd-cryptsetup-generator", permitting automatic setup of
  713. Verity root partitions when systemd boots up. In order to make use of
  714. this your partition setup should follow the Discoverable Partitions
  715. Specification, and the GPT partition ID of the root file system
  716. partition should be identical to the upper 128bit of the Verity root
  717. hash. The GPT partition ID of the Verity partition protecting it
  718. should be the lower 128bit of the Verity root hash. If the partition
  719. image follows this model it is sufficient to specify a single
  720. "roothash=" kernel command line argument to both configure which root
  721. image and verity partition to use as well as the root hash for
  722. it. Note that systemd-nspawn's Verity support follows the same
  723. semantics, meaning that disk images with proper Verity data in place
  724. may be booted in containers with systemd-nspawn as well as on
  725. physical systems via the verity generator. Also note that the "mkosi"
  726. tool available at has been updated
  727. to generate Verity protected disk images following this scheme. In
  728. fact, it has been updated to generate disk images that optionally
  729. implement a complete UEFI SecureBoot trust chain, involving a signed
  730. kernel and initrd image that incorporates such a root hash as well as
  731. a Verity-enabled root partition.
  732. * The hardware database (hwdb) udev supports has been updated to carry
  733. accelerometer quirks.
  734. * All system services are now run with a fresh kernel keyring set up
  735. for them. The invocation ID is stored by default in it, thus
  736. providing a safe, non-overridable way to determine the invocation
  737. ID of each service.
  738. * Service unit files gained new BindPaths= and BindReadOnlyPaths=
  739. options for bind mounting arbitrary paths in a service-specific
  740. way. When these options are used, arbitrary host or service files and
  741. directories may be mounted to arbitrary locations in the service's
  742. view.
  743. * Documentation has been added that lists all of systemd's low-level
  744. environment variables:
  746. * sd-daemon gained a new API sd_is_socket_sockaddr() for determining
  747. whether a specific socket file descriptor matches a specified socket
  748. address.
  749. * systemd-firstboot has been updated to check for the
  750. systemd.firstboot= kernel command line option. It accepts a boolean
  751. and when set to false the first boot questions are skipped.
  752. * systemd-fstab-generator has been updated to check for the
  753. systemd.volatile= kernel command line option, which either takes an
  754. optional boolean parameter or the special value "state". If used the
  755. system may be booted in a "volatile" boot mode. Specifically,
  756. "systemd.volatile" is used, the root directory will be mounted as
  757. tmpfs, and only /usr is mounted from the actual root file system. If
  758. "systemd.volatile=state" is used, the root directory will be mounted
  759. as usual, but /var is mounted as tmpfs. This concept provides similar
  760. functionality as systemd-nspawn's --volatile= option, but provides it
  761. on physical boots. Use this option for implementing stateless
  762. systems, or testing systems with all state and/or configuration reset
  763. to the defaults. (Note though that many distributions are not
  764. prepared to boot up without a populated /etc or /var, though.)
  765. * systemd-gpt-auto-generator gained support for LUKS encrypted root
  766. partitions. Previously it only supported LUKS encrypted partitions
  767. for all other uses, except for the root partition itself.
  768. * Socket units gained support for listening on AF_VSOCK sockets for
  769. communication in virtualized QEMU environments.
  770. * The "configure" script gained a new option --with-fallback-hostname=
  771. for specifying the fallback hostname to use if none is configured in
  772. /etc/hostname. For example, by specifying
  773. --with-fallback-hostname=fedora it is possible to default to a
  774. hostname of "fedora" on pristine installations.
  775. * systemd-cgls gained support for a new --unit= switch for listing only
  776. the control groups of a specific unit. Similar --user-unit= has been
  777. added for listing only the control groups of a specific user unit.
  778. * systemd-mount gained a new --umount switch for unmounting a mount or
  779. automount point (and all mount/automount points below it).
  780. * systemd will now refuse full configuration reloads (via systemctl
  781. daemon-reload and related calls) unless at least 16MiB of free space
  782. are available in /run. This is a safety precaution in order to ensure
  783. that generators can safely operate after the reload completed.
  784. * A new unit file option RootImage= has been added, which has a similar
  785. effect as RootDirectory= but mounts the service's root directory from
  786. a disk image instead of plain directory. This logic reuses the same
  787. image dissection and mount logic that systemd-nspawn already uses,
  788. and hence supports any disk images systemd-nspawn supports, including
  789. those following the Discoverable Partition Specification, as well as
  790. Verity enabled images. This option enables systemd to run system
  791. services directly off disk images acting as resource bundles,
  792. possibly even including full integrity data.
  793. * A new MountAPIVFS= unit file option has been added, taking a boolean
  794. argument. If enabled /proc, /sys and /dev (collectively called the
  795. "API VFS") will be mounted for the service. This is only relevant if
  796. RootDirectory= or RootImage= is used for the service, as these mounts
  797. are of course in place in the host mount namespace anyway.
  798. * systemd-nspawn gained support for a new --pivot-root= switch. If
  799. specified the root directory within the container image is pivoted to
  800. the specified mount point, while the original root disk is moved to a
  801. different place. This option enables booting of ostree images
  802. directly with systemd-nspawn.
  803. * The systemd build scripts will no longer complain if the NTP server
  804. addresses are not changed from the defaults. Google now supports
  805. these NTP servers officially. We still recommend downstreams to
  806. properly register an NTP pool with the NTP pool project though.
  807. * coredumpctl gained a new "--reverse" option for printing the list
  808. of coredumps in reverse order.
  809. * coredumpctl will now show additional information about truncated and
  810. inaccessible coredumps, as well as coredumps that are still being
  811. processed. It also gained a new --quiet switch for suppressing
  812. additional informational message in its output.
  813. * coredumpctl gained support for only showing coredumps newer and/or
  814. older than specific timestamps, using the new --since= and --until=
  815. options, reminiscent of journalctl's options by the same name.
  816. * The systemd-coredump logic has been improved so that it may be reused
  817. to collect backtraces in non-compiled languages, for example in
  818. scripting languages such as Python.
  819. * machinectl will now show the UID shift of local containers, if user
  820. namespacing is enabled for them.
  821. * systemd will now optionally run "environment generator" binaries at
  822. configuration load time. They may be used to add environment
  823. variables to the environment block passed to services invoked. One
  824. user environment generator is shipped by default that sets up
  825. environment variables based on files dropped into /etc/environment.d
  826. and ~/.config/environment.d/.
  827. * systemd-resolved now includes the new, recently published 2017 DNSSEC
  828. root key (KSK).
  829. * hostnamed has been updated to report a new chassis type of
  830. "convertible" to cover "foldable" laptops that can both act as a
  831. tablet and as a laptop, such as various Lenovo Yoga devices.
  832. Contributions from: Adrián López, Alexander Galanin, Alexander
  833. Kochetkov, Alexandros Frantzis, Andrey Ulanov, Antoine Eiche, Baruch
  834. Siach, Bastien Nocera, Benjamin Robin, Björn, Brandon Philips, Cédric
  835. Schieli, Charles (Chas) Williams, Christian Hesse, Daniele Medri,
  836. Daniel Drake, Daniel Rusek, Daniel Wagner, Dan Streetman, Dave Reisner,
  837. David Glasser, David Herrmann, David Michael, Djalal Harouni, Dmitry
  838. Khlebnikov, Dmitry Rozhkov, Dongsu Park, Douglas Christman, Earnestly,
  839. Emil Soleyman, Eric Cook, Evgeny Vereshchagin, Felipe Sateler, Fionn
  840. Cleary, Florian Klink, Francesco Brozzu, Franck Bui, Gabriel Rauter,
  841. Gianluca Boiano, Giedrius Statkevičius, Graeme Lawes, Hans de Goede,
  842. Harald Hoyer, Ian Kelling, Ivan Shapovalov, Jakub Wilk, Janne Heß, Jan
  843. Synacek, Jason Reeder, Jonathan Boulle, Jörg Thalheim, Jouke Witteveen,
  844. Karl Kraus, Kees Cook, Keith Busch, Kieran Colford, kilian-k, Lennart
  845. Poettering, Lubomir Rintel, Lucas Werkmeister, Lukas Rusak, Maarten de
  846. Vries, Maks Naumov, Mantas Mikulėnas, Marc-Andre Lureau, Marcin Bachry,
  847. Mark Stosberg, Martin Ejdestig, Martin Pitt, Mauricio Faria de
  848. Oliveira, micah, Michael Biebl, Michael Shields, Michal Schmidt, Michal
  849. Sekletar, Michel Kraus, Mike Gilbert, Mikko Ylinen, Mirza Krak,
  850. Namhyung Kim, nikolaof, peoronoob, Peter Hutterer, Peter Körner, Philip
  851. Withnall, Piotr Drąg, Ray Strode, Reverend Homer, Rike-Benjamin
  852. Schuppner, Robert Kreuzer, Ronny Chevalier, Ruslan Bilovol, sammynx,
  853. Sergey Ptashnick, Sergiusz Urbaniak, Stefan Berger, Stefan Hajnoczi,
  854. Stefan Schweter, Stuart McLaren, Susant Sahani, Sylvain Plantefève,
  855. Taylor Smock, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tibor
  856. Nagy, Tobias Stoeckmann, Tom Gundersen, Torstein Husebø, Viktar
  857. Vaŭčkievič, Viktor Mihajlovski, Vitaly Sulimov, Waldemar Brodkorb,
  858. Walter Garcia-Fontes, Wim de With, Yassine Imounachen, Yi EungJun,
  859. YunQiang Su, Yu Watanabe, Zbigniew Jędrzejewski-Szmek, Александр
  860. Тихонов
  861. — Berlin, 2017-03-01
  862. CHANGES WITH 232:
  863. * udev now runs with MemoryDenyWriteExecute=, RestrictRealtime= and
  864. RestrictAddressFamilies= enabled. These sandboxing options should
  865. generally be compatible with the various external udev call-out
  866. binaries we are aware of, however there may be exceptions, in
  867. particular when exotic languages for these call-outs are used. In
  868. this case, consider turning off these settings locally.
  869. * The new RemoveIPC= option can be used to remove IPC objects owned by
  870. the user or group of a service when that service exits.
  871. * The new ProtectKernelModules= option can be used to disable explicit
  872. load and unload operations of kernel modules by a service. In
  873. addition access to /usr/lib/modules is removed if this option is set.
  874. * ProtectSystem= option gained a new value "strict", which causes the
  875. whole file system tree with the exception of /dev, /proc, and /sys,
  876. to be remounted read-only for a service.
  877. * The new ProtectKernelTunables= option can be used to disable
  878. modification of configuration files in /sys and /proc by a service.
  879. Various directories and files are remounted read-only, so access is
  880. restricted even if the file permissions would allow it.
  881. * The new ProtectControlGroups= option can be used to disable write
  882. access by a service to /sys/fs/cgroup.
  883. * Various systemd services have been hardened with
  884. ProtectKernelTunables=yes, ProtectControlGroups=yes,
  885. RestrictAddressFamilies=.
  886. * Support for dynamically creating users for the lifetime of a service
  887. has been added. If DynamicUser=yes is specified, user and group IDs
  888. will be allocated from the range 61184..65519 for the lifetime of the
  889. service. They can be resolved using the new NSS
  890. module. The module must be enabled in /etc/nsswitch.conf. Services
  891. started in this way have PrivateTmp= and RemoveIPC= enabled, so that
  892. any resources allocated by the service will be cleaned up when the
  893. service exits. They also have ProtectHome=read-only and
  894. ProtectSystem=strict enabled, so they are not able to make any
  895. permanent modifications to the system.
  896. * The nss-systemd module also always resolves root and nobody, making
  897. it possible to have no /etc/passwd or /etc/group files in minimal
  898. container or chroot environments.
  899. * Services may be started with their own user namespace using the new
  900. boolean PrivateUsers= option. Only root, nobody, and the uid/gid
  901. under which the service is running are mapped. All other users are
  902. mapped to nobody.
  903. * Support for the cgroup namespace has been added to systemd-nspawn. If
  904. supported by kernel, the container system started by systemd-nspawn
  905. will have its own view of the cgroup hierarchy. This new behaviour
  906. can be disabled using $SYSTEMD_NSPAWN_USE_CGNS environment variable.
  907. * The new MemorySwapMax= option can be used to limit the maximum swap
  908. usage under the unified cgroup hierarchy.
  909. * Support for the CPU controller in the unified cgroup hierarchy has
  910. been added, via the CPUWeight=, CPUStartupWeight=, CPUAccounting=
  911. options. This controller requires out-of-tree patches for the kernel
  912. and the support is provisional.
  913. * Mount and automount units may now be created transiently
  914. (i.e. dynamically at runtime via the bus API, instead of requiring
  915. unit files in the file system).
  916. * systemd-mount is a new tool which may mount file systems – much like
  917. mount(8), optionally pulling in additional dependencies through
  918. transient .mount and .automount units. For example, this tool
  919. automatically runs fsck on a backing block device before mounting,
  920. and allows the automount logic to be used dynamically from the
  921. command line for establishing mount points. This tool is particularly
  922. useful when dealing with removable media, as it will ensure fsck is
  923. run – if necessary – before the first access and that the file system
  924. is quickly unmounted after each access by utilizing the automount
  925. logic. This maximizes the chance that the file system on the
  926. removable media stays in a clean state, and if it isn't in a clean
  927. state is fixed automatically.
  928. * LazyUnmount=yes option for mount units has been added to expose the
  929. umount --lazy option. Similarly, ForceUnmount=yes exposes the --force
  930. option.
  931. * /efi will be used as the mount point of the EFI boot partition, if
  932. the directory is present, and the mount point was not configured
  933. through other means (e.g. fstab). If /efi directory does not exist,
  934. /boot will be used as before. This makes it easier to automatically
  935. mount the EFI partition on systems where /boot is used for something
  936. else.
  937. * When operating on GPT disk images for containers, systemd-nspawn will
  938. now mount the ESP to /boot or /efi according to the same rules as PID
  939. 1 running on a host. This allows tools like "bootctl" to operate
  940. correctly within such containers, in order to make container images
  941. bootable on physical systems.
  942. * disk/by-id and disk/by-path symlinks are now created for NVMe drives.
  943. * Two new user session targets have been added to support running
  944. graphical sessions under the systemd --user instance:
  945. and See
  946. systemd.special(7) for a description of how those targets should be
  947. used.
  948. * The vconsole initialization code has been significantly reworked to
  949. use KD_FONT_OP_GET/SET ioctls instead of KD_FONT_OP_COPY and better
  950. support unicode keymaps. Font and keymap configuration will now be
  951. copied to all allocated virtual consoles.
  952. * FreeBSD's bhyve virtualization is now detected.
  953. * Information recorded in the journal for core dumps now includes the
  954. contents of /proc/mountinfo and the command line of the process at
  955. the top of the process hierarchy (which is usually the init process
  956. of the container).
  957. * systemd-journal-gatewayd learned the --directory= option to serve
  958. files from the specified location.
  959. * journalctl --root=… can be used to peruse the journal in the
  960. /var/log/ directories inside of a container tree. This is similar to
  961. the existing --machine= option, but does not require the container to
  962. be active.
  963. * The hardware database has been extended to support
  964. ID_INPUT_TRACKBALL, used in addition to ID_INPUT_MOUSE to identify
  965. trackball devices.
  966. MOUSE_WHEEL_CLICK_ANGLE_HORIZONTAL hwdb property has been added to
  967. specify the click rate for mice which include a horizontal wheel with
  968. a click rate that is different than the one for the vertical wheel.
  969. * systemd-run gained a new --wait option that makes service execution
  970. synchronous. (Specifically, the command will not return until the
  971. specified service binary exited.)
  972. * systemctl gained a new --wait option that causes the start command to
  973. wait until the units being started have terminated again.
  974. * A new journal output mode "short-full" has been added which displays
  975. timestamps with abbreviated English day names and adds a timezone
  976. suffix. Those timestamps include more information than the default
  977. "short" output mode, and can be passed directly to journalctl's
  978. --since= and --until= options.
  979. * /etc/resolv.conf will be bind-mounted into containers started by
  980. systemd-nspawn, if possible, so any changes to resolv.conf contents
  981. are automatically propagated to the container.
  982. * The number of instances for socket-activated services originating
  983. from a single IP address can be limited with
  984. MaxConnectionsPerSource=, extending the existing setting of
  985. MaxConnections=.
  986. * systemd-networkd gained support for vcan ("Virtual CAN") interface
  987. configuration.
  988. * .netdev and .network configuration can now be extended through
  989. drop-ins.
  990. * UDP Segmentation Offload, TCP Segmentation Offload, Generic
  991. Segmentation Offload, Generic Receive Offload, Large Receive Offload
  992. can be enabled and disabled using the new UDPSegmentationOffload=,
  993. TCPSegmentationOffload=, GenericSegmentationOffload=,
  994. GenericReceiveOffload=, LargeReceiveOffload= options in the
  995. [Link] section of .link files.
  996. * The Spanning Tree Protocol, Priority, Aging Time, and the Default
  997. Port VLAN ID can be configured for bridge devices using the new STP=,
  998. Priority=, AgeingTimeSec=, and DefaultPVID= settings in the [Bridge]
  999. section of .netdev files.
  1000. * The route table to which routes received over DHCP or RA should be
  1001. added can be configured with the new RouteTable= option in the [DHCP]
  1002. and [IPv6AcceptRA] sections of .network files.
  1003. * The Address Resolution Protocol can be disabled on links managed by
  1004. systemd-networkd using the ARP=no setting in the [Link] section of
  1005. .network files.
  1006. * New environment variables $SERVICE_RESULT, $EXIT_CODE and
  1007. $EXIT_STATUS are set for ExecStop= and ExecStopPost= commands, and
  1008. encode information about the result and exit codes of the current
  1009. service runtime cycle.
  1010. * systemd-sysctl will now configure kernel parameters in the order
  1011. they occur in the configuration files. This matches what sysctl
  1012. has been traditionally doing.
  1013. * kernel-install "plugins" that are executed to perform various
  1014. tasks after a new kernel is added and before an old one is removed
  1015. can now return a special value to terminate the procedure and
  1016. prevent any later plugins from running.
  1017. * Journald's SplitMode=login setting has been deprecated. It has been
  1018. removed from documentation, and its use is discouraged. In a future
  1019. release it will be completely removed, and made equivalent to current
  1020. default of SplitMode=uid.
  1021. * Storage=both option setting in /etc/systemd/coredump.conf has been
  1022. removed. With fast LZ4 compression storing the core dump twice is not
  1023. useful.
  1024. * The --share-system systemd-nspawn option has been replaced with an
  1025. (undocumented) variable $SYSTEMD_NSPAWN_SHARE_SYSTEM, but the use of
  1026. this functionality is discouraged. In addition the variables
  1028. $SYSTEMD_NSPAWN_SHARE_NS_UTS may be used to control the unsharing of
  1029. individual namespaces.
  1030. * "machinectl list" now shows the IP address of running containers in
  1031. the output, as well as OS release information.
  1032. * "loginctl list" now shows the TTY of each session in the output.
  1033. * sd-bus gained new API calls sd_bus_track_set_recursive(),
  1034. sd_bus_track_get_recursive(), sd_bus_track_count_name(),
  1035. sd_bus_track_count_sender(). They permit usage of sd_bus_track peer
  1036. tracking objects in a "recursive" mode, where a single client can be
  1037. counted multiple times, if it takes multiple references.
  1038. * sd-bus gained new API calls sd_bus_set_exit_on_disconnect() and
  1039. sd_bus_get_exit_on_disconnect(). They may be used to to make a
  1040. process using sd-bus automatically exit if the bus connection is
  1041. severed.
  1042. * Bus clients of the service manager may now "pin" loaded units into
  1043. memory, by taking an explicit reference on them. This is useful to
  1044. ensure the client can retrieve runtime data about the service even
  1045. after the service completed execution. Taking such a reference is
  1046. available only for privileged clients and should be helpful to watch
  1047. running services in a race-free manner, and in particular collect
  1048. information about exit statuses and results.
  1049. * The nss-resolve module has been changed to strictly return UNAVAIL
  1050. when communication via D-Bus with resolved failed, and NOTFOUND when
  1051. a lookup completed but was negative. This means it is now possible to
  1052. neatly configure fallbacks using nsswitch.conf result checking
  1053. expressions. Taking benefit of this, the new recommended
  1054. configuration line for the "hosts" entry in /etc/nsswitch.conf is:
  1055. hosts: files mymachines resolve [!UNAVAIL=return] dns myhostname
  1056. * A new setting CtrlAltDelBurstAction= has been added to
  1057. /etc/systemd/system.conf which may be used to configure the precise
  1058. behaviour if the user on the console presses Ctrl-Alt-Del more often
  1059. than 7 times in 2s. Previously this would unconditionally result in
  1060. an expedited, immediate reboot. With this new setting the precise
  1061. operation may be configured in more detail, and also turned off
  1062. entirely.
  1063. * In .netdev files two new settings RemoteChecksumTx= and
  1064. RemoteChecksumRx= are now understood that permit configuring the
  1065. remote checksumming logic for VXLAN networks.
  1066. * The service manager learnt a new "invocation ID" concept for invoked
  1067. services. Each runtime cycle of a service will get a new invocation
  1068. ID (a 128bit random UUID) assigned that identifies the current
  1069. run of the service uniquely and globally. A new invocation ID
  1070. is generated each time a service starts up. The journal will store
  1071. the invocation ID of a service along with any logged messages, thus
  1072. making the invocation ID useful for matching the online runtime of a
  1073. service with the offline log data it generated in a safe way without
  1074. relying on synchronized timestamps. In many ways this new service
  1075. invocation ID concept is similar to the kernel's boot ID concept that
  1076. uniquely and globally identifies the runtime of each boot. The
  1077. invocation ID of a service is passed to the service itself via an
  1078. environment variable ($INVOCATION_ID). A new bus call
  1079. GetUnitByInvocationID() has been added that is similar to GetUnit()
  1080. but instead of retrieving the bus path for a unit by its name
  1081. retrieves it by its invocation ID. The returned path is valid only as
  1082. long as the passed invocation ID is current.
  1083. * systemd-resolved gained a new "DNSStubListener" setting in
  1084. resolved.conf. It either takes a boolean value or the special values
  1085. "udp" and "tcp", and configures whether to enable the stub DNS
  1086. listener on
  1087. * IP addresses configured via networkd may now carry additional
  1088. configuration settings supported by the kernel. New options include:
  1089. HomeAddress=, DuplicateAddressDetection=, ManageTemporaryAddress=,
  1090. PrefixRoute=, AutoJoin=.
  1091. * The PAM configuration fragment file for "user@.service" shipped with
  1092. systemd (i.e. the --user instance of systemd) has been stripped to
  1093. the minimum necessary to make the system boot. Previously, it
  1094. contained Fedora-specific stanzas that did not apply to other
  1095. distributions. It is expected that downstream distributions add
  1096. additional configuration lines, matching their needs to this file,
  1097. using it only as rough template of what systemd itself needs. Note
  1098. that this reduced fragment does not even include an invocation of
  1099. pam_limits which most distributions probably want to add, even though
  1100. systemd itself does not need it. (There's also the new build time
  1101. option --with-pamconfdir=no to disable installation of the PAM
  1102. fragment entirely.)
  1103. * If PrivateDevices=yes is set for a service the CAP_SYS_RAWIO
  1104. capability is now also dropped from its set (in addition to
  1105. CAP_SYS_MKNOD as before).
  1106. * In service unit files it is now possible to connect a specific named
  1107. file descriptor with stdin/stdout/stdout of an executed service. The
  1108. name may be specified in matching .socket units using the
  1109. FileDescriptorName= setting.
  1110. * A number of journal settings may now be configured on the kernel
  1111. command line. Specifically, the following options are now understood:
  1112. systemd.journald.max_level_console=,
  1113. systemd.journald.max_level_store=,
  1114. systemd.journald.max_level_syslog=, systemd.journald.max_level_kmsg=,
  1115. systemd.journald.max_level_wall=.
  1116. * "systemctl is-enabled --full" will now show by which symlinks a unit
  1117. file is enabled in the unit dependency tree.
  1118. * Support for VeraCrypt encrypted partitions has been added to the
  1119. "cryptsetup" logic and /etc/crypttab.
  1120. * systemd-detect-virt gained support for a new --private-users switch
  1121. that checks whether the invoking processes are running inside a user
  1122. namespace. Similar, a new special value "private-users" for the
  1123. existing ConditionVirtualization= setting has been added, permitting
  1124. skipping of specific units in user namespace environments.
  1125. Contributions from: Alban Crequy, Alexander Kuleshov, Alfie John,
  1126. Andreas Henriksson, Andrew Jeddeloh, Balázs Úr, Bart Rulon, Benjamin
  1127. Richter, Ben Gamari, Ben Harris, Brian J. Murrell, Christian Brauner,
  1128. Christian Rebischke, Clinton Roy, Colin Walters, Cristian Rodríguez,
  1129. Daniel Hahler, Daniel Mack, Daniel Maixner, Daniel Rusek, Dan Dedrick,
  1130. Davide Cavalca, David Herrmann, David Michael, Dennis Wassenberg,
  1131. Djalal Harouni, Dongsu Park, Douglas Christman, Elias Probst, Eric
  1132. Cook, Erik Karlsson, Evgeny Vereshchagin, Felipe Sateler, Felix Zhang,
  1133. Franck Bui, George Hilliard, Giuseppe Scrivano, HATAYAMA Daisuke,
  1134. Heikki Kemppainen, Hendrik Brueckner, hi117, Ismo Puustinen, Ivan
  1135. Shapovalov, Jakub Filak, Jakub Wilk, Jan Synacek, Jason Kölker,
  1136. Jean-Sébastien Bour, Jiří Pírko, Jonathan Boulle, Jorge Niedbalski,
  1137. Keith Busch, kristbaum, Kyle Russell, Lans Zhang, Lennart Poettering,
  1138. Leonardo Brondani Schenkel, Lucas Werkmeister, Luca Bruno, Lukáš
  1139. Nykrýn, Maciek Borzecki, Mantas Mikulėnas, Marc-Antoine Perennou,
  1140. Marcel Holtmann, Marcos Mello, Martin Ejdestig, Martin Pitt, Matej
  1141. Habrnal, Maxime de Roucy, Michael Biebl, Michael Chapman, Michael Hoy,
  1142. Michael Olbrich, Michael Pope, Michal Sekletar, Michal Soltys, Mike
  1143. Gilbert, Nick Owens, Patrik Flykt, Paweł Szewczyk, Peter Hutterer,
  1144. Piotr Drąg, Reid Price, Richard W.M. Jones, Roman Stingler, Ronny
  1145. Chevalier, Seraphime Kirkovski, Stefan Schweter, Steve Muir, Susant
  1146. Sahani, Tejun Heo, Thomas Blume, Thomas H. P. Andersen, Tiago Levit,
  1147. Tobias Jungel, Tomáš Janoušek, Topi Miettinen, Torstein Husebø, Umut
  1148. Tezduyar Lindskog, Vito Caputo, WaLyong Cho, Wilhelm Schuster, Yann
  1149. E. MORIN, Yi EungJun, Yuki Inoguchi, Yu Watanabe, Zbigniew
  1150. Jędrzejewski-Szmek, Zeal Jagannatha
  1151. — Santa Fe, 2016-11-03
  1152. CHANGES WITH 231:
  1153. * In service units the various ExecXYZ= settings have been extended
  1154. with an additional special character as first argument of the
  1155. assigned value: if the character '+' is used the specified command
  1156. line it will be run with full privileges, regardless of User=,
  1157. Group=, CapabilityBoundingSet= and similar options. The effect is
  1158. similar to the existing PermissionsStartOnly= option, but allows
  1159. configuration of this concept for each executed command line
  1160. independently.
  1161. * Services may now alter the service watchdog timeout at runtime by
  1162. sending a WATCHDOG_USEC= message via sd_notify().
  1163. * MemoryLimit= and related unit settings now optionally take percentage
  1164. specifications. The percentage is taken relative to the amount of
  1165. physical memory in the system (or in case of containers, the assigned
  1166. amount of memory). This allows scaling service resources neatly with
  1167. the amount of RAM available on the system. Similarly, systemd-logind's
  1168. RuntimeDirectorySize= option now also optionally takes percentage
  1169. values.
  1170. * In similar fashion TasksMax= takes percentage values now, too. The
  1171. value is taken relative to the configured maximum number of processes
  1172. on the system. The per-service task maximum has been changed to 15%
  1173. using this functionality. (Effectively this is an increase of 512 →
  1174. 4915 for service units, given the kernel's default pid_max setting.)
  1175. * Calendar time specifications in .timer units now understand a ".."
  1176. syntax for time ranges. Example: "4..7:10" may now be used for
  1177. defining a timer that is triggered at 4:10am, 5:10am, 6:10am and
  1178. 7:10am every day.
  1179. * The InaccessableDirectories=, ReadOnlyDirectories= and
  1180. ReadWriteDirectories= unit file settings have been renamed to
  1181. InaccessablePaths=, ReadOnlyPaths= and ReadWritePaths= and may now be
  1182. applied to all kinds of file nodes, and not just directories, with
  1183. the exception of symlinks. Specifically these settings may now be
  1184. used on block and character device nodes, UNIX sockets and FIFOS as
  1185. well as regular files. The old names of these settings remain
  1186. available for compatibility.
  1187. * systemd will now log about all service processes it kills forcibly
  1188. (using SIGKILL) because they remained after the clean shutdown phase
  1189. of the service completed. This should help identifying services that
  1190. shut down uncleanly. Moreover if KillUserProcesses= is enabled in
  1191. systemd-logind's configuration a similar log message is generated for
  1192. processes killed at the end of each session due to this setting.
  1193. * systemd will now set the $JOURNAL_STREAM environment variable for all
  1194. services whose stdout/stderr are connected to the Journal (which
  1195. effectively means by default: all services). The variable contains
  1196. the device and inode number of the file descriptor used for
  1197. stdout/stderr. This may be used by invoked programs to detect whether
  1198. their stdout/stderr is connected to the Journal, in which case they
  1199. can switch over to direct Journal communication, thus being able to
  1200. pass extended, structured metadata along with their log messages. As
  1201. one example, this is now used by glib's logging primitives.
  1202. * When using systemd's default tmp.mount unit for /tmp, the mount point
  1203. will now be established with the "nosuid" and "nodev" options. This
  1204. avoids privilege escalation attacks that put traps and exploits into
  1205. /tmp. However, this might cause problems if you e. g. put container
  1206. images or overlays into /tmp; if you need this, override tmp.mount's
  1207. "Options=" with a drop-in, or mount /tmp from /etc/fstab with your
  1208. desired options.
  1209. * systemd now supports the "memory" cgroup controller also on
  1210. cgroupsv2.
  1211. * The systemd-cgtop tool now optionally takes a control group path as
  1212. command line argument. If specified, the control group list shown is
  1213. limited to subgroups of that group.
  1214. * The SystemCallFilter= unit file setting gained support for
  1215. pre-defined, named system call filter sets. For example
  1216. SystemCallFilter=@clock is now an effective way to make all clock
  1217. changing-related system calls unavailable to a service. A number of
  1218. similar pre-defined groups are defined. Writing system call filters
  1219. for system services is simplified substantially with this new
  1220. concept. Accordingly, all of systemd's own, long-running services now
  1221. enable system call filtering based on this, by default.
  1222. * A new service setting MemoryDenyWriteExecute= has been added, taking
  1223. a boolean value. If turned on, a service may no longer create memory
  1224. mappings that are writable and executable at the same time. This
  1225. enhances security for services where this is enabled as it becomes
  1226. harder to dynamically write and then execute memory in exploited
  1227. service processes. This option has been enabled for all of systemd's
  1228. own long-running services.
  1229. * A new RestrictRealtime= service setting has been added, taking a
  1230. boolean argument. If set the service's processes may no longer
  1231. acquire realtime scheduling. This improves security as realtime
  1232. scheduling may otherwise be used to easily freeze the system.
  1233. * systemd-nspawn gained a new switch --notify-ready= taking a boolean
  1234. value. This may be used for requesting that the system manager inside
  1235. of the container reports start-up completion to nspawn which then
  1236. propagates this notification further to the service manager
  1237. supervising nspawn itself. A related option NotifyReady= in .nspawn
  1238. files has been added too. This functionality allows ordering of the
  1239. start-up of multiple containers using the usual systemd ordering
  1240. primitives.
  1241. * machinectl gained a new command "stop" that is an alias for
  1242. "terminate".
  1243. * systemd-resolved gained support for contacting DNS servers on
  1244. link-local IPv6 addresses.
  1245. * If systemd-resolved receives the SIGUSR2 signal it will now flush all
  1246. its caches. A method call for requesting the same operation has been
  1247. added to the bus API too, and is made available via "systemd-resolve
  1248. --flush-caches".
  1249. * systemd-resolve gained a new --status switch. If passed a brief
  1250. summary of the used DNS configuration with per-interface information
  1251. is shown.
  1252. * resolved.conf gained a new Cache= boolean option, defaulting to
  1253. on. If turned off local DNS caching is disabled. This comes with a
  1254. performance penalty in particular when DNSSEC is enabled. Note that
  1255. resolved disables its internal caching implicitly anyway, when the
  1256. configured DNS server is on a host-local IP address such as ::1 or
  1257., thus automatically avoiding double local caching.
  1258. * systemd-resolved now listens on the local IP address
  1259. for DNS requests. This improves compatibility with local programs
  1260. that do not use the libc NSS or systemd-resolved's bus APIs for name
  1261. resolution. This minimal DNS service is only available to local
  1262. programs and does not implement the full DNS protocol, but enough to
  1263. cover local DNS clients. A new, static resolv.conf file, listing just
  1264. this DNS server is now shipped in /usr/lib/systemd/resolv.conf. It is
  1265. now recommended to make /etc/resolv.conf a symlink to this file in
  1266. order to route all DNS lookups to systemd-resolved, regardless if
  1267. done via NSS, the bus API or raw DNS packets. Note that this local
  1268. DNS service is not as fully featured as the libc NSS or
  1269. systemd-resolved's bus APIs. For example, as unicast DNS cannot be
  1270. used to deliver link-local address information (as this implies
  1271. sending a local interface index along), LLMNR/mDNS support via this
  1272. interface is severely restricted. It is thus strongly recommended for
  1273. all applications to use the libc NSS API or native systemd-resolved
  1274. bus API instead.
  1275. * systemd-networkd's bridge support learned a new setting
  1276. VLANFiltering= for controlling VLAN filtering. Moreover a new section
  1277. in .network files has been added for configuring VLAN bridging in
  1278. more detail: VLAN=, EgressUntagged=, PVID= in [BridgeVLAN].
  1279. * systemd-networkd's IPv6 Router Advertisement code now makes use of
  1280. the DNSSL and RDNSS options. This means IPv6 DNS configuration may
  1281. now be acquired without relying on DHCPv6. Two new options
  1282. UseDomains= and UseDNS= have been added to configure this behaviour.
  1283. * systemd-networkd's IPv6AcceptRouterAdvertisements= option has been
  1284. renamed IPv6AcceptRA=, without altering its behaviour. The old
  1285. setting name remains available for compatibility reasons.
  1286. * The systemd-networkd VTI/VTI6 tunneling support gained new options
  1287. Key=, InputKey= and OutputKey=.
  1288. * systemd-networkd gained support for VRF ("Virtual Routing Function")
  1289. interface configuration.
  1290. * "systemctl edit" may now be used to create new unit files by
  1291. specifying the --force switch.
  1292. * sd-event gained a new function sd_event_get_iteration() for
  1293. requesting the current iteration counter of the event loop. It starts
  1294. at zero and is increased by one with each event loop iteration.
  1295. * A new rpm macro %systemd_ordering is provided by the macros.systemd
  1296. file. It can be used in lieu of %systemd_requires in packages which
  1297. don't use any systemd functionality and are intended to be installed
  1298. in minimal containers without systemd present. This macro provides
  1299. ordering dependencies to ensure that if the package is installed in
  1300. the same rpm transaction as systemd, systemd will be installed before
  1301. the scriptlets for the package are executed, allowing unit presets
  1302. to be handled.
  1303. New macros %_systemdgeneratordir and %_systemdusergeneratordir have
  1304. been added to simplify packaging of generators.
  1305. * The os-release file gained VERSION_CODENAME field for the
  1306. distribution nickname (e.g. VERSION_CODENAME=woody).
  1308. can be set to disable parsing of metadata and the creation
  1309. of persistent symlinks for that device.
  1310. * The v230 change to tag framebuffer devices (/dev/fb*) with "uaccess"
  1311. to make them available to logged-in users has been reverted.
  1312. * Much of the common code of the various systemd components is now
  1313. built into an internal shared library
  1314. (incorporating the systemd version number in the name, to be updated
  1315. with future releases) that the components link to. This should
  1316. decrease systemd footprint both in memory during runtime and on
  1317. disk. Note that the shared library is not for public use, and is
  1318. neither API not ABI stable, but is likely to change with every new
  1319. released update. Packagers need to make sure that binaries
  1320. linking to are updated in step with the
  1321. library.
  1322. * Configuration for "mkosi" is now part of the systemd
  1323. repository. mkosi is a tool to easily build legacy-free OS images,
  1324. and is available on github: If
  1325. "mkosi" is invoked in the build tree a new raw OS image is generated
  1326. incorporating the systemd sources currently being worked on and a
  1327. clean, fresh distribution installation. The generated OS image may be
  1328. booted up with "systemd-nspawn -b -i", qemu-kvm or on any physical
  1329. UEFI PC. This functionality is particularly useful to easily test
  1330. local changes made to systemd in a pristine, defined environment. See
  1331. HACKING for details.
  1332. * configure learned the --with-support-url= option to specify the
  1333. distribution's bugtracker.
  1334. Contributions from: Alban Crequy, Alessandro Puccetti, Alessio Igor
  1335. Bogani, Alexander Kuleshov, Alexander Kurtz, Alex Gaynor, Andika
  1336. Triwidada, Andreas Pokorny, Andreas Rammhold, Andrew Jeddeloh, Ansgar
  1337. Burchardt, Atrotors, Benjamin Drung, Brian Boylston, Christian Hesse,
  1338. Christian Rebischke, Daniele Medri, Daniel Mack, Dave Reisner, David
  1339. Herrmann, David Michael, Djalal Harouni, Douglas Christman, Elias
  1340. Probst, Evgeny Vereshchagin, Federico Mena Quintero, Felipe Sateler,
  1341. Franck Bui, Harald Hoyer, Ian Lee, Ivan Shapovalov, Jakub Wilk, Jan
  1342. Janssen, Jean-Sébastien Bour, John Paul Adrian Glaubitz, Jouke
  1343. Witteveen, Kai Ruhnau, kpengboy, Kyle Walker, Lénaïc Huard, Lennart
  1344. Poettering, Luca Bruno, Lukas Lösche, Lukáš Nykrýn, mahkoh, Marcel
  1345. Holtmann, Martin Pitt, Marty Plummer, Matthieu Codron, Max Prokhorov,
  1346. Michael Biebl, Michael Karcher, Michael Olbrich, Michał Bartoszkiewicz,
  1347. Michal Sekletar, Michal Soltys, Minkyung, Muhammet Kara, mulkieran,
  1348. Otto Wallenius, Pablo Lezaeta Reyes, Peter Hutterer, Ronny Chevalier,
  1349. Rusty Bird, Stef Walter, Susant Sahani, Tejun Heo, Thomas Blume, Thomas
  1350. Haller, Thomas H. P. Andersen, Tobias Jungel, Tom Gundersen, Tom Yan,
  1351. Topi Miettinen, Torstein Husebø, Valentin Vidić, Viktar Vaŭčkievič,
  1352. WaLyong Cho, Weng Xuetian, Werner Fink, Zbigniew Jędrzejewski-Szmek
  1353. — Berlin, 2016-07-25
  1354. CHANGES WITH 230:
  1355. * DNSSEC is now turned on by default in systemd-resolved (in
  1356. "allow-downgrade" mode), but may be turned off during compile time by
  1357. passing "--with-default-dnssec=no" to "configure" (and of course,
  1358. during runtime with DNSSEC= in resolved.conf). We recommend
  1359. downstreams to leave this on at least during development cycles and
  1360. report any issues with the DNSSEC logic upstream. We are very
  1361. interested in collecting feedback about the DNSSEC validator and its
  1362. limitations in the wild. Note however, that DNSSEC support is
  1363. probably nothing downstreams should turn on in stable distros just
  1364. yet, as it might create incompatibilities with a few DNS servers and
  1365. networks. We tried hard to make sure we downgrade to non-DNSSEC mode
  1366. automatically whenever we detect such incompatible setups, but there
  1367. might be systems we do not cover yet. Hence: please help us testing
  1368. the DNSSEC code, leave this on where you can, report back, but then
  1369. again don't consider turning this on in your stable, LTS or
  1370. production release just yet. (Note that you have to enable
  1371. nss-resolve in /etc/nsswitch.conf, to actually use systemd-resolved
  1372. and its DNSSEC mode for host name resolution from local
  1373. applications.)
  1374. * systemd-resolve conveniently resolves DANE records with the --tlsa
  1375. option and OPENPGPKEY records with the --openpgp option. It also
  1376. supports dumping raw DNS record data via the new --raw= switch.
  1377. * systemd-logind will now by default terminate user processes that are
  1378. part of the user session scope unit (session-XX.scope) when the user
  1379. logs out. This behavior is controlled by the KillUserProcesses=
  1380. setting in logind.conf, and the previous default of "no" is now
  1381. changed to "yes". This means that user sessions will be properly
  1382. cleaned up after, but additional steps are necessary to allow
  1383. intentionally long-running processes to survive logout.
  1384. While the user is logged in at least once, user@.service is running,
  1385. and any service that should survive the end of any individual login
  1386. session can be started at a user service or scope using systemd-run.
  1387. systemd-run(1) man page has been extended with an example which shows
  1388. how to run screen in a scope unit underneath user@.service. The same
  1389. command works for tmux.
  1390. After the user logs out of all sessions, user@.service will be
  1391. terminated too, by default, unless the user has "lingering" enabled.
  1392. To effectively allow users to run long-term tasks even if they are
  1393. logged out, lingering must be enabled for them. See loginctl(1) for
  1394. details. The default polkit policy was modified to allow users to
  1395. set lingering for themselves without authentication.
  1396. Previous defaults can be restored at compile time by the
  1397. --without-kill-user-processes option to "configure".
  1398. * systemd-logind gained new configuration settings SessionsMax= and
  1399. InhibitorsMax=, both with a default of 8192. It will not register new
  1400. user sessions or inhibitors above this limit.
  1401. * systemd-logind will now reload configuration on SIGHUP.
  1402. * The unified cgroup hierarchy added in Linux 4.5 is now supported.
  1403. Use systemd.unified_cgroup_hierarchy=1 on the kernel command line to
  1404. enable. Also, support for the "io" cgroup controller in the unified
  1405. hierarchy has been added, so that the "memory", "pids" and "io" are
  1406. now the controllers that are supported on the unified hierarchy.
  1407. WARNING: it is not possible to use previous systemd versions with
  1408. systemd.unified_cgroup_hierarchy=1 and the new kernel. Therefore it
  1409. is necessary to also update systemd in the initramfs if using the
  1410. unified hierarchy. An updated SELinux policy is also required.
  1411. * LLDP support has been extended, and both passive (receive-only) and
  1412. active (sender) modes are supported. Passive mode ("routers-only") is
  1413. enabled by default in systemd-networkd. Active LLDP mode is enabled
  1414. by default for containers on the internal network. The "networkctl
  1415. lldp" command may be used to list information gathered. "networkctl
  1416. status" will also show basic LLDP information on connected peers now.
  1417. * The IAID and DUID unique identifier sent in DHCP requests may now be
  1418. configured for the system and each .network file managed by
  1419. systemd-networkd using the DUIDType=, DUIDRawData=, IAID= options.
  1420. * systemd-networkd gained support for configuring proxy ARP support for
  1421. each interface, via the ProxyArp= setting in .network files. It also
  1422. gained support for configuring the multicast querier feature of
  1423. bridge devices, via the new MulticastQuerier= setting in .netdev
  1424. files. Similarly, snooping on the IGMP traffic can be controlled
  1425. via the new setting MulticastSnooping=.
  1426. A new setting PreferredLifetime= has been added for addresses
  1427. configured in .network file to configure the lifetime intended for an
  1428. address.
  1429. The systemd-networkd DHCP server gained the option EmitRouter=, which
  1430. defaults to yes, to configure whether the DHCP Option 3 (Router)
  1431. should be emitted.
  1432. * The testing tool /usr/lib/systemd/systemd-activate is renamed to
  1433. systemd-socket-activate and installed into /usr/bin. It is now fully
  1434. supported.
  1435. * systemd-journald now uses separate threads to flush changes to disk
  1436. when closing journal files, thus reducing impact of slow disk I/O on
  1437. logging performance.
  1438. * The sd-journal API gained two new calls
  1439. sd_journal_open_directory_fd() and sd_journal_open_files_fd() which
  1440. can be used to open journal files using file descriptors instead of
  1441. file or directory paths. sd_journal_open_container() has been
  1442. deprecated, sd_journal_open_directory_fd() should be used instead
  1443. with the flag SD_JOURNAL_OS_ROOT.
  1444. * journalctl learned a new output mode "-o short-unix" that outputs log
  1445. lines prefixed by their UNIX time (i.e. seconds since Jan 1st, 1970
  1446. UTC). It also gained support for a new --no-hostname setting to
  1447. suppress the hostname column in the family of "short" output modes.
  1448. * systemd-ask-password now optionally skips printing of the password to
  1449. stdout with --no-output which can be useful in scripts.
  1450. * Framebuffer devices (/dev/fb*) and 3D printers and scanners
  1451. (devices tagged with ID_MAKER_TOOL) are now tagged with
  1452. "uaccess" and are available to logged in users.
  1453. * The DeviceAllow= unit setting now supports specifiers (with "%").
  1454. * "systemctl show" gained a new --value switch, which allows print a
  1455. only the contents of a specific unit property, without also printing
  1456. the property's name. Similar support was added to "show*" verbs
  1457. of loginctl and machinectl that output "key=value" lists.
  1458. * A new unit type "generated" was added for files dynamically generated
  1459. by generator tools. Similarly, a new unit type "transient" is used
  1460. for unit files created using the runtime API. "systemctl enable" will
  1461. refuse to operate on such files.
  1462. * A new command "systemctl revert" has been added that may be used to
  1463. revert to the vendor version of a unit file, in case local changes
  1464. have been made by adding drop-ins or overriding the unit file.
  1465. * "machinectl clean" gained a new verb to automatically remove all or
  1466. just hidden container images.
  1467. * systemd-tmpfiles gained support for a new line type "e" for emptying
  1468. directories, if they exist, without creating them if they don't.
  1469. * systemd-nspawn gained support for automatically patching the UID/GIDs
  1470. of the owners and the ACLs of all files and directories in a
  1471. container tree to match the UID/GID user namespacing range selected
  1472. for the container invocation. This mode is enabled via the new
  1473. --private-users-chown switch. It also gained support for
  1474. automatically choosing a free, previously unused UID/GID range when
  1475. starting a container, via the new --private-users=pick setting (which
  1476. implies --private-users-chown). Together, these options for the first
  1477. time make user namespacing for nspawn containers fully automatic and
  1478. thus deployable. The systemd-nspawn@.service template unit file has
  1479. been changed to use this functionality by default.
  1480. * systemd-nspawn gained a new --network-zone= switch, that allows
  1481. creating ad-hoc virtual Ethernet links between multiple containers,
  1482. that only exist as long as at least one container referencing them is
  1483. running. This allows easy connecting of multiple containers with a
  1484. common link that implements an Ethernet broadcast domain. Each of
  1485. these network "zones" may be named relatively freely by the user, and
  1486. may be referenced by any number of containers, but each container may
  1487. only reference one of these "zones". On the lower level, this is
  1488. implemented by an automatically managed bridge network interface for
  1489. each zone, that is created when the first container referencing its
  1490. zone is created and removed when the last one referencing its zone
  1491. terminates.
  1492. * The default start timeout may now be configured on the kernel command
  1493. line via systemd.default_timeout_start_sec=. It was already
  1494. configurable via the DefaultTimeoutStartSec= option in
  1495. /etc/systemd/system.conf.
  1496. * Socket units gained a new TriggerLimitIntervalSec= and
  1497. TriggerLimitBurst= setting to configure a limit on the activation
  1498. rate of the socket unit.
  1499. * The LimitNICE= setting now optionally takes normal UNIX nice values
  1500. in addition to the raw integer limit value. If the specified
  1501. parameter is prefixed with "+" or "-" and is in the range -20..19 the
  1502. value is understood as UNIX nice value. If not prefixed like this it
  1503. is understood as raw RLIMIT_NICE limit.
  1504. * Note that the effect of the PrivateDevices= unit file setting changed
  1505. slightly with this release: the per-device /dev file system will be
  1506. mounted read-only from this version on, and will have "noexec"
  1507. set. This (minor) change of behavior might cause some (exceptional)
  1508. legacy software to break, when PrivateDevices=yes is set for its
  1509. service. Please leave PrivateDevices= off if you run into problems
  1510. with this.
  1511. * systemd-bootchart has been split out to a separate repository:
  1513. * systemd-bus-proxyd has been removed, as kdbus is unlikely to still be
  1514. merged into the kernel in its current form.
  1515. * The compatibility libraries,
  1516.,, and
  1517. which have been deprecated since systemd-209 have been removed along
  1518. with the corresponding pkg-config files. All symbols provided by
  1519. those libraries are provided by
  1520. * The Capabilities= unit file setting has been removed (it is ignored
  1521. for backwards compatibility). AmbientCapabilities= and
  1522. CapabilityBoundingSet= should be used instead.
  1523. * A new special target has been added,,
  1524. which creates a synchronization point for dependencies of the root
  1525. device in early userspace. Initramfs builders must ensure that this
  1526. target is now included in early userspace.
  1527. Contributions from: Alban Crequy, Alexander Kuleshov, Alexander Shopov,
  1528. Alex Crawford, Andre Klärner, Andrew Eikum, Beniamino Galvani, Benjamin
  1529. Robin, Biao Lu, Bjørnar Ness, Calvin Owens, Christian Hesse, Clemens
  1530. Gruber, Colin Guthrie, Daniel Drake, Daniele Medri, Daniel J Walsh,
  1531. Daniel Mack, Dan Nicholson, daurnimator, David Herrmann, David
  1532. R. Hedges, Elias Probst, Emmanuel Gil Peyrot, EMOziko, Evgeny
  1533. Vereshchagin, Federico, Felipe Sateler, Filipe Brandenburger, Franck
  1534. Bui, frankheckenbach, gdamjan, Georgia Brikis, Harald Hoyer, Hendrik
  1535. Brueckner, Hristo Venev, Iago López Galeiras, Ian Kelling, Ismo
  1536. Puustinen, Jakub Wilk, Jaroslav Škarvada, Jeff Huang, Joel Holdsworth,
  1537. John Paul Adrian Glaubitz, Jonathan Boulle, kayrus, Klearchos
  1538. Chaloulos, Kyle Russell, Lars Uebernickel, Lennart Poettering, Lubomir
  1539. Rintel, Lukáš Nykrýn, Mantas Mikulėnas, Marcel Holtmann, Martin Pitt,
  1540. Michael Biebl, michaelolbrich, Michał Bartoszkiewicz, Michal Koutný,
  1541. Michal Sekletar, Mike Frysinger, Mike Gilbert, Mingcong Bai, Ming Lin,
  1542. mulkieran, muzena, Nalin Dahyabhai, Naohiro Aota, Nathan McSween,
  1543. Nicolas Braud-Santoni, Patrik Flykt, Peter Hutterer, Peter Mattern,
  1544. Petr Lautrbach, Petros Angelatos, Piotr Drąg, Rabin Vincent, Robert
  1545. Węcławski, Ronny Chevalier, Samuel Tardieu, Stefan Saraev, Stefan
  1546. Schallenberg aka nafets227, Steven Siloti, Susant Sahani, Sylvain
  1547. Plantefève, Taylor Smock, Tejun Heo, Thomas Blume, Thomas Haller,
  1548. Thomas H. P. Andersen, Tobias Klauser, Tom Gundersen, topimiettinen,
  1549. Torstein Husebø, Umut Tezduyar Lindskog, Uwe Kleine-König, Victor Toso,
  1550. Vinay Kulkarni, Vito Caputo, Vittorio G (VittGam), Vladimir Panteleev,
  1551. Wieland Hoffmann, Wouter Verhelst, Yu Watanabe, Zbigniew
  1552. Jędrzejewski-Szmek
  1553. — Fairfax, 2016-05-21
  1554. CHANGES WITH 229:
  1555. * The systemd-resolved DNS resolver service has gained a substantial
  1556. set of new features, most prominently it may now act as a DNSSEC
  1557. validating stub resolver. DNSSEC mode is currently turned off by
  1558. default, but is expected to be turned on by default in one of the
  1559. next releases. For now, we invite everybody to test the DNSSEC logic
  1560. by setting DNSSEC=allow-downgrade in /etc/systemd/resolved.conf. The
  1561. service also gained a full set of D-Bus interfaces, including calls
  1562. to configure DNS and DNSSEC settings per link (for use by external
  1563. network management software). systemd-resolved and systemd-networkd
  1564. now distinguish between "search" and "routing" domains. The former
  1565. are used to qualify single-label names, the latter are used purely
  1566. for routing lookups within certain domains to specific links.
  1567. resolved now also synthesizes RRs for all entries from /etc/hosts.
  1568. * The systemd-resolve tool (which is a client utility for
  1569. systemd-resolved) has been improved considerably and is now fully
  1570. supported and documented. Hence it has moved from /usr/lib/systemd to
  1571. /usr/bin.
  1572. * /dev/disk/by-path/ symlink support has been (re-)added for virtio
  1573. devices.
  1574. * The coredump collection logic has been reworked: when a coredump is
  1575. collected it is now written to disk, compressed and processed
  1576. (including stacktrace extraction) from a new instantiated service
  1577. systemd-coredump@.service, instead of directly from the
  1578. /proc/sys/kernel/core_pattern hook we provide. This is beneficial as
  1579. processing large coredumps can take up a substantial amount of
  1580. resources and time, and this previously happened entirely outside of
  1581. systemd's service supervision. With the new logic the core_pattern
  1582. hook only does minimal metadata collection before passing off control
  1583. to the new instantiated service, which is configured with a time
  1584. limit, a nice level and other settings to minimize negative impact on
  1585. the rest of the system. Also note that the new logic will honour the
  1586. RLIMIT_CORE setting of the crashed process, which now allows users
  1587. and processes to turn off coredumping for their processes by setting
  1588. this limit.
  1589. * The RLIMIT_CORE resource limit now defaults to "unlimited" for PID 1
  1590. and all forked processes by default. Previously, PID 1 would leave
  1591. the setting at "0" for all processes, as set by the kernel. Note that
  1592. the resource limit traditionally has no effect on the generated
  1593. coredumps on the system if the /proc/sys/kernel/core_pattern hook
  1594. logic is used. Since the limit is now honoured (see above) its
  1595. default has been changed so that the coredumping logic is enabled by
  1596. default for all processes, while allowing specific opt-out.
  1597. * When the stacktrace is extracted from processes of system users, this
  1598. is now done as "systemd-coredump" user, in order to sandbox this
  1599. potentially security sensitive parsing operation. (Note that when
  1600. processing coredumps of normal users this is done under the user ID
  1601. of process that crashed, as before.) Packagers should take notice
  1602. that it is now necessary to create the "systemd-coredump" system user
  1603. and group at package installation time.
  1604. * The systemd-activate socket activation testing tool gained support
  1605. for SOCK_DGRAM and SOCK_SEQPACKET sockets using the new --datagram
  1606. and --seqpacket switches. It also has been extended to support both
  1607. new-style and inetd-style file descriptor passing. Use the new
  1608. --inetd switch to request inetd-style file descriptor passing.
  1609. * Most systemd tools now honor a new $SYSTEMD_COLORS environment
  1610. variable, which takes a boolean value. If set to false, ANSI color
  1611. output is disabled in the tools even when run on a terminal that
  1612. supports it.
  1613. * The VXLAN support in networkd now supports two new settings
  1614. DestinationPort= and PortRange=.
  1615. * A new systemd.machine_id= kernel command line switch has been added,
  1616. that may be used to set the machine ID in /etc/machine-id if it is
  1617. not initialized yet. This command line option has no effect if the
  1618. file is already initialized.
  1619. * systemd-nspawn gained a new --as-pid2 switch that invokes any
  1620. specified command line as PID 2 rather than PID 1 in the
  1621. container. In this mode PID 1 is a minimal stub init process that
  1622. implements the special POSIX and Linux semantics of PID 1 regarding
  1623. signal and child process management. Note that this stub init process
  1624. is implemented in nspawn itself and requires no support from the
  1625. container image. This new logic is useful to support running
  1626. arbitrary commands in the container, as normal processes are
  1627. generally not prepared to run as PID 1.
  1628. * systemd-nspawn gained a new --chdir= switch for setting the current
  1629. working directory for the process started in the container.
  1630. * "journalctl /dev/sda" will now output all kernel log messages for
  1631. specified device from the current boot, in addition to all devices
  1632. that are parents of it. This should make log output about devices
  1633. pretty useful, as long as kernel drivers attach enough metadata to
  1634. the log messages. (The usual SATA drivers do.)
  1635. * The sd-journal API gained two new calls
  1636. sd_journal_has_runtime_files() and sd_journal_has_persistent_files()
  1637. that report whether log data from /run or /var has been found.
  1638. * journalctl gained a new switch "--fields" that prints all journal
  1639. record field names currently in use in the journal. This is backed
  1640. by two new sd-journal API calls sd_journal_enumerate_fields() and
  1641. sd_journal_restart_fields().
  1642. * Most configurable timeouts in systemd now expect an argument of
  1643. "infinity" to turn them off, instead of "0" as before. The semantics
  1644. from now on is that a timeout of "0" means "now", and "infinity"
  1645. means "never". To maintain backwards compatibility, "0" continues to
  1646. turn off previously existing timeout settings.
  1647. * "systemctl reload-or-try-restart" has been renamed to "systemctl
  1648. try-reload-or-restart" to clarify what it actually does: the "try"
  1649. logic applies to both reloading and restarting, not just restarting.
  1650. The old name continues to be accepted for compatibility.
  1651. * On boot-up, when PID 1 detects that the system clock is behind the
  1652. release date of the systemd version in use, the clock is now set
  1653. to the latter. Previously, this was already done in timesyncd, in order
  1654. to avoid running with clocks set to the various clock epochs such as
  1655. 1902, 1938 or 1970. With this change the logic is now done in PID 1
  1656. in addition to timesyncd during early boot-up, so that it is enforced
  1657. before the first process is spawned by systemd. Note that the logic
  1658. in timesyncd remains, as it is more comprehensive and ensures
  1659. clock monotonicity by maintaining a persistent timestamp file in
  1660. /var. Since /var is generally not available in earliest boot or the
  1661. initrd, this part of the logic remains in timesyncd, and is not done
  1662. by PID 1.
  1663. * Support for tweaking details in net_cls.class_id through the
  1664. NetClass= configuration directive has been removed, as the kernel
  1665. people have decided to deprecate that controller in cgroup v2.
  1666. Userspace tools such as nftables are moving over to setting rules
  1667. that are specific to the full cgroup path of a task, which obsoletes
  1668. these controllers anyway. The NetClass= directive is kept around for
  1669. legacy compatibility reasons. For a more in-depth description of the
  1670. kernel change, please refer to the respective upstream commit:
  1672. * A new service setting RuntimeMaxSec= has been added that may be used
  1673. to specify a maximum runtime for a service. If the timeout is hit, the
  1674. service is terminated and put into a failure state.
  1675. * A new service setting AmbientCapabilities= has been added. It allows
  1676. configuration of additional Linux process capabilities that are
  1677. passed to the activated processes. This is only available on very
  1678. recent kernels.
  1679. * The process resource limit settings in service units may now be used
  1680. to configure hard and soft limits individually.
  1681. * The various libsystemd APIs such as sd-bus or sd-event now publicly
  1682. expose support for gcc's __attribute__((cleanup())) C extension.
  1683. Specifically, for many object destructor functions alternative
  1684. versions have been added that have names suffixed with "p" and take a
  1685. pointer to a pointer to the object to destroy, instead of just a
  1686. pointer to the object itself. This is useful because these destructor
  1687. functions may be used directly as parameters to the cleanup
  1688. construct. Internally, systemd has been a heavy user of this GCC
  1689. extension for a long time, and with this change similar support is
  1690. now available to consumers of the library outside of systemd. Note
  1691. that by using this extension in your sources compatibility with old
  1692. and strictly ANSI compatible C compilers is lost. However, all gcc or
  1693. LLVM versions of recent years support this extension.
  1694. * Timer units gained support for a new setting RandomizedDelaySec= that
  1695. allows configuring some additional randomized delay to the configured
  1696. time. This is useful to spread out timer events to avoid load peaks in
  1697. clusters or larger setups.
  1698. * Calendar time specifications now support sub-second accuracy.
  1699. * Socket units now support listening on SCTP and UDP-lite protocol
  1700. sockets.
  1701. * The sd-event API now comes with a full set of man pages.
  1702. * Older versions of systemd contained experimental support for
  1703. compressing journal files and coredumps with the LZ4 compressor that
  1704. was not compatible with the lz4 binary (due to API limitations of the
  1705. lz4 library). This support has been removed; only support for files
  1706. compatible with the lz4 binary remains. This LZ4 logic is now
  1707. officially supported and no longer considered experimental.
  1708. * The dkr image import logic has been removed again from importd. dkr's
  1709. micro-services focus doesn't fit into the machine image focus of
  1710. importd, and quickly got out of date with the upstream dkr API.
  1711. * Creation of the /run/lock/lockdev/ directory was dropped from
  1712. tmpfiles.d/legacy.conf. Better locking mechanisms like flock() have
  1713. been available for many years. If you still need this, you need to
  1714. create your own tmpfiles.d config file with:
  1715. d /run/lock/lockdev 0775 root lock -
  1716. Contributions from: Abdo Roig-Maranges, Alban Crequy, Aleksander
  1717. Adamowski, Alexander Kuleshov, Andreas Pokorny, Andrei Borzenkov,
  1718. Andrew Wilcox, Arthur Clement, Beniamino Galvani, Casey Schaufler,
  1719. Chris Atkinson, Chris Mayo, Christian Hesse, Damjan Georgievski, Dan
  1720. Dedrick, Daniele Medri, Daniel J Walsh, Daniel Korostil, Daniel Mack,
  1721. David Herrmann, Dimitri John Ledkov, Dominik Hannen, Douglas Christman,
  1722. Evgeny Vereshchagin, Filipe Brandenburger, Franck Bui, Gabor Kelemen,
  1723. Harald Hoyer, Hayden Walles, Helmut Grohne, Henrik Kaare Poulsen,
  1724. Hristo Venev, Hui Wang, Indrajit Raychaudhuri, Ismo Puustinen, Jakub
  1725. Wilk, Jan Alexander Steffens (heftig), Jan Engelhardt, Jan Synacek,
  1726. Joost Bremmer, Jorgen Schaefer, Karel Zak, Klearchos Chaloulos,
  1727. lc85446, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel
  1728. Holtmann, Martin Pitt, Michael Biebl, Michael Olbrich, Michael Scherer,
  1729. Michał Górny, Michal Sekletar, Nicolas Cornu, Nicolas Iooss, Nils
  1730. Carlson, nmartensen, nnz1024, Patrick Ohly, Peter Hutterer, Phillip Sz,
  1731. Ronny Chevalier, Samu Kallio, Shawn Landden, Stef Walter, Susant
  1732. Sahani, Sylvain Plantefève, Tadej Janež, Thomas Hindoe Paaboel
  1733. Andersen, Tom Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito
  1734. Caputo, WaLyong Cho, Yu Watanabe, Zbigniew Jędrzejewski-Szmek
  1735. — Berlin, 2016-02-11
  1736. CHANGES WITH 228:
  1737. * A number of properties previously only settable in unit
  1738. files are now also available as properties to set when
  1739. creating transient units programmatically via the bus, as it
  1740. is exposed with systemd-run's --property=
  1741. setting. Specifically, these are: SyslogIdentifier=,
  1742. SyslogLevelPrefix=, TimerSlackNSec=, OOMScoreAdjust=,
  1743. EnvironmentFile=, ReadWriteDirectories=,
  1744. ReadOnlyDirectories=, InaccessibleDirectories=,
  1745. ProtectSystem=, ProtectHome=, RuntimeDirectory=.
  1746. * When creating transient services via the bus API it is now
  1747. possible to pass in a set of file descriptors to use as
  1748. STDIN/STDOUT/STDERR for the invoked process.
  1749. * Slice units may now be created transiently via the bus APIs,
  1750. similar to the way service and scope units may already be
  1751. created transiently.
  1752. * Wherever systemd expects a calendar timestamp specification
  1753. (like in journalctl's --since= and --until= switches) UTC
  1754. timestamps are now supported. Timestamps suffixed with "UTC"
  1755. are now considered to be in Universal Time Coordinated
  1756. instead of the local timezone. Also, timestamps may now
  1757. optionally be specified with sub-second accuracy. Both of
  1758. these additions also apply to recurring calendar event
  1759. specification, such as OnCalendar= in timer units.
  1760. * journalctl gained a new "--sync" switch that asks the
  1761. journal daemon to write all so far unwritten log messages to
  1762. disk and sync the files, before returning.
  1763. * systemd-tmpfiles learned two new line types "q" and "Q" that
  1764. operate like "v", but also set up a basic btrfs quota
  1765. hierarchy when used on a btrfs file system with quota
  1766. enabled.
  1767. * tmpfiles' "v", "q" and "Q" will now create a plain directory
  1768. instead of a subvolume (even on a btrfs file system) if the
  1769. root directory is a plain directory, and not a
  1770. subvolume. This should simplify things with certain chroot()
  1771. environments which are not aware of the concept of btrfs
  1772. subvolumes.
  1773. * systemd-detect-virt gained a new --chroot switch to detect
  1774. whether execution takes place in a chroot() environment.
  1775. * CPUAffinity= now takes CPU index ranges in addition to
  1776. individual indexes.
  1777. * The various memory-related resource limit settings (such as
  1778. LimitAS=) now understand the usual K, M, G, ... suffixes to
  1779. the base of 1024 (IEC). Similar, the time-related resource
  1780. limit settings understand the usual min, h, day, ...
  1781. suffixes now.
  1782. * There's a new system.conf setting DefaultTasksMax= to
  1783. control the default TasksMax= setting for services and
  1784. scopes running on the system. (TasksMax= is the primary
  1785. setting that exposes the "pids" cgroup controller on systemd
  1786. and was introduced in the previous systemd release.) The
  1787. setting now defaults to 512, which means services that are
  1788. not explicitly configured otherwise will only be able to
  1789. create 512 processes or threads at maximum, from this
  1790. version on. Note that this means that thread- or
  1791. process-heavy services might need to be reconfigured to set
  1792. TasksMax= to a higher value. It is sufficient to set
  1793. TasksMax= in these specific unit files to a higher value, or
  1794. even "infinity". Similar, there's now a logind.conf setting
  1795. UserTasksMax= that defaults to 4096 and limits the total
  1796. number of processes or tasks each user may own
  1797. concurrently. nspawn containers also have the TasksMax=
  1798. value set by default now, to 8192. Note that all of this
  1799. only has an effect if the "pids" cgroup controller is
  1800. enabled in the kernel. The general benefit of these changes
  1801. should be a more robust and safer system, that provides a
  1802. certain amount of per-service fork() bomb protection.
  1803. * systemd-nspawn gained the new --network-veth-extra= switch
  1804. to define additional and arbitrarily-named virtual Ethernet
  1805. links between the host and the container.
  1806. * A new service execution setting PassEnvironment= has been
  1807. added that allows importing select environment variables
  1808. from PID1's environment block into the environment block of
  1809. the service.
  1810. * Timer units gained support for a new RemainAfterElapse=
  1811. setting which takes a boolean argument. It defaults to on,
  1812. exposing behaviour unchanged to previous releases. If set to
  1813. off, timer units are unloaded after they elapsed if they
  1814. cannot elapse again. This is particularly useful for
  1815. transient timer units, which shall not stay around longer
  1816. than until they first elapse.
  1817. * systemd will now bump the net.unix.max_dgram_qlen to 512 by
  1818. default now (the kernel default is 16). This is beneficial
  1819. for avoiding blocking on AF_UNIX/SOCK_DGRAM sockets since it
  1820. allows substantially larger numbers of queued
  1821. datagrams. This should increase the capability of systemd to
  1822. parallelize boot-up, as logging and sd_notify() are unlikely
  1823. to stall execution anymore. If you need to change the value
  1824. from the new defaults, use the usual sysctl.d/ snippets.
  1825. * The compression framing format used by the journal or
  1826. coredump processing has changed to be in line with what the
  1827. official LZ4 tools generate. LZ4 compression support in
  1828. systemd was considered unsupported previously, as the format
  1829. was not compatible with the normal tools. With this release
  1830. this has changed now, and it is hence safe for downstream
  1831. distributions to turn it on. While not compressing as well
  1832. as the XZ, LZ4 is substantially faster, which makes
  1833. it a good default choice for the compression logic in the
  1834. journal and in coredump handling.
  1835. * Any reference to /etc/mtab has been dropped from
  1836. systemd. The file has been obsolete since a while, but
  1837. systemd refused to work on systems where it was incorrectly
  1838. set up (it should be a symlink or non-existent). Please make
  1839. sure to update to util-linux 2.27.1 or newer in conjunction
  1840. with this systemd release, which also drops any reference to
  1841. /etc/mtab. If you maintain a distribution make sure that no
  1842. software you package still references it, as this is a
  1843. likely source of bugs. There's also a glibc bug pending,
  1844. asking for removal of any reference to this obsolete file:
  1846. Note that only util-linux versions built with
  1847. --enable-libmount-force-mountinfo are supported.
  1848. * Support for the ".snapshot" unit type has been removed. This
  1849. feature turned out to be little useful and little used, and
  1850. has now been removed from the core and from systemctl.
  1851. * The dependency types RequiresOverridable= and
  1852. RequisiteOverridable= have been removed from systemd. They
  1853. have been used only very sparingly to our knowledge and
  1854. other options that provide a similar effect (such as
  1855. systemctl --mode=ignore-dependencies) are much more useful
  1856. and commonly used. Moreover, they were only half-way
  1857. implemented as the option to control behaviour regarding
  1858. these dependencies was never added to systemctl. By removing
  1859. these dependency types the execution engine becomes a bit
  1860. simpler. Unit files that use these dependencies should be
  1861. changed to use the non-Overridable dependency types
  1862. instead. In fact, when parsing unit files with these
  1863. options, that's what systemd will automatically convert them
  1864. too, but it will also warn, asking users to fix the unit
  1865. files accordingly. Removal of these dependency types should
  1866. only affect a negligible number of unit files in the wild.
  1867. * Behaviour of networkd's IPForward= option changed
  1868. (again). It will no longer maintain a per-interface setting,
  1869. but propagate one way from interfaces where this is enabled
  1870. to the global kernel setting. The global setting will be
  1871. enabled when requested by a network that is set up, but
  1872. never be disabled again. This change was made to make sure
  1873. IPv4 and IPv6 behaviour regarding packet forwarding is
  1874. similar (as the Linux IPv6 stack does not support
  1875. per-interface control of this setting) and to minimize
  1876. surprises.
  1877. * In unit files the behaviour of %u, %U, %h, %s has
  1878. changed. These specifiers will now unconditionally resolve
  1879. to the various user database fields of the user that the
  1880. systemd instance is running as, instead of the user
  1881. configured in the specific unit via User=. Note that this
  1882. effectively doesn't change much, as resolving of these
  1883. specifiers was already turned off in the --system instance
  1884. of systemd, as we cannot do NSS lookups from PID 1. In the
  1885. --user instance of systemd these specifiers where correctly
  1886. resolved, but hardly made any sense, since the user instance
  1887. lacks privileges to do user switches anyway, and User= is
  1888. hence useless. Moreover, even in the --user instance of
  1889. systemd behaviour was awkward as it would only take settings
  1890. from User= assignment placed before the specifier into
  1891. account. In order to unify and simplify the logic around
  1892. this the specifiers will now always resolve to the
  1893. credentials of the user invoking the manager (which in case
  1894. of PID 1 is the root user).
  1895. Contributions from: Andrew Jones, Beniamino Galvani, Boyuan
  1896. Yang, Daniel Machon, Daniel Mack, David Herrmann, David
  1897. Reynolds, David Strauss, Dongsu Park, Evgeny Vereshchagin,
  1898. Felipe Sateler, Filipe Brandenburger, Franck Bui, Hristo
  1899. Venev, Iago López Galeiras, Jan Engelhardt, Jan Janssen, Jan
  1900. Synacek, Jesus Ornelas Aguayo, Karel Zak, kayrus, Kay Sievers,
  1901. Lennart Poettering, Liu Yuan Yuan, Mantas Mikulėnas, Marcel
  1902. Holtmann, Marcin Bachry, Marcos Alano, Marcos Mello, Mark
  1903. Theunissen, Martin Pitt, Michael Marineau, Michael Olbrich,
  1904. Michal Schmidt, Michal Sekletar, Mirco Tischler, Nick Owens,
  1905. Nicolas Cornu, Patrik Flykt, Peter Hutterer, reverendhomer,
  1906. Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Shawn Landden,
  1907. Susant Sahani, Thomas Haller, Thomas Hindoe Paaboel Andersen,
  1908. Tom Gundersen, Torstein Husebø, Vito Caputo, Zbigniew
  1909. Jędrzejewski-Szmek
  1910. — Berlin, 2015-11-18
  1911. CHANGES WITH 227:
  1912. * systemd now depends on util-linux v2.27. More specifically,
  1913. the newly added mount monitor feature in libmount now
  1914. replaces systemd's former own implementation.
  1915. * libmount mandates /etc/mtab not to be regular file, and
  1916. systemd now enforces this condition at early boot.
  1917. /etc/mtab has been deprecated and warned about for a very
  1918. long time, so systems running systemd should already have
  1919. stopped having this file around as anything else than a
  1920. symlink to /proc/self/mounts.
  1921. * Support for the "pids" cgroup controller has been added. It
  1922. allows accounting the number of tasks in a cgroup and
  1923. enforcing limits on it. This adds two new setting
  1924. TasksAccounting= and TasksMax= to each unit, as well as a
  1925. global option DefaultTasksAccounting=.
  1926. * Support for the "net_cls" cgroup controller has been added.
  1927. It allows assigning a net class ID to each task in the
  1928. cgroup, which can then be used in firewall rules and traffic
  1929. shaping configurations. Note that the kernel netfilter net
  1930. class code does not currently work reliably for ingress
  1931. packets on unestablished sockets.
  1932. This adds a new config directive called NetClass= to CGroup
  1933. enabled units. Allowed values are positive numbers for fixed
  1934. assignments and "auto" for picking a free value
  1935. automatically.
  1936. * 'systemctl is-system-running' now returns 'offline' if the
  1937. system is not booted with systemd. This command can now be
  1938. used as a substitute for 'systemd-notify --booted'.
  1939. * Watchdog timeouts have been increased to 3 minutes for all
  1940. in-tree service files. Apparently, disk IO issues are more
  1941. frequent than we hoped, and user reported >1 minute waiting
  1942. for disk IO.
  1943. * 'machine-id-commit' functionality has been merged into
  1944. 'machine-id-setup --commit'. The separate binary has been
  1945. removed.
  1946. * The WorkingDirectory= directive in unit files may now be set
  1947. to the special value '~'. In this case, the working
  1948. directory is set to the home directory of the user
  1949. configured in User=.
  1950. * "machinectl shell" will now open the shell in the home
  1951. directory of the selected user by default.
  1952. * The CrashChVT= configuration file setting is renamed to
  1953. CrashChangeVT=, following our usual logic of not
  1954. abbreviating unnecessarily. The old directive is still
  1955. supported for compat reasons. Also, this directive now takes
  1956. an integer value between 1 and 63, or a boolean value. The
  1957. formerly supported '-1' value for disabling stays around for
  1958. compat reasons.
  1959. * The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
  1960. NoNewPrivileges=, TTYPath=, WorkingDirectory= and
  1961. RootDirectory= properties can now be set for transient
  1962. units.
  1963. * The systemd-analyze tool gained a new "set-log-target" verb
  1964. to change the logging target the system manager logs to
  1965. dynamically during runtime. This is similar to how
  1966. "systemd-analyze set-log-level" already changes the log
  1967. level.
  1968. * In nspawn /sys is now mounted as tmpfs, with only a selected
  1969. set of subdirectories mounted in from the real sysfs. This
  1970. enhances security slightly, and is useful for ensuring user
  1971. namespaces work correctly.
  1972. * Support for USB FunctionFS activation has been added. This
  1973. allows implementation of USB gadget services that are
  1974. activated as soon as they are requested, so that they don't
  1975. have to run continuously, similar to classic socket
  1976. activation.
  1977. * The "systemctl exit" command now optionally takes an
  1978. additional parameter that sets the exit code to return from
  1979. the systemd manager when exiting. This is only relevant when
  1980. running the systemd user instance, or when running the
  1981. system instance in a container.
  1982. * sd-bus gained the new API calls sd_bus_path_encode_many()
  1983. and sd_bus_path_decode_many() that allow easy encoding and
  1984. decoding of multiple identifier strings inside a D-Bus
  1985. object path. Another new call sd_bus_default_flush_close()
  1986. has been added to flush and close per-thread default
  1987. connections.
  1988. * systemd-cgtop gained support for a -M/--machine= switch to
  1989. show the control groups within a certain container only.
  1990. * "systemctl kill" gained support for an optional --fail
  1991. switch. If specified the requested operation will fail of no
  1992. processes have been killed, because the unit had no
  1993. processes attached, or similar.
  1994. * A new systemd.crash_reboot=1 kernel command line option has
  1995. been added that triggers a reboot after crashing. This can
  1996. also be set through CrashReboot= in systemd.conf.
  1997. * The RuntimeDirectory= setting now understands unit
  1998. specifiers like %i or %f.
  1999. * A new (still internal) library API sd-ipv4acd has been added,
  2000. that implements address conflict detection for IPv4. It's
  2001. based on code from sd-ipv4ll, and will be useful for
  2002. detecting DHCP address conflicts.
  2003. * File descriptors passed during socket activation may now be
  2004. named. A new API sd_listen_fds_with_names() is added to
  2005. access the names. The default names may be overridden,
  2006. either in the .socket file using the FileDescriptorName=
  2007. parameter, or by passing FDNAME= when storing the file
  2008. descriptors using sd_notify().
  2009. * systemd-networkd gained support for:
  2010. - Setting the IPv6 Router Advertisement settings via
  2011. IPv6AcceptRouterAdvertisements= in .network files.
  2012. - Configuring the HelloTimeSec=, MaxAgeSec= and
  2013. ForwardDelaySec= bridge parameters in .netdev files.
  2014. - Configuring PreferredSource= for static routes in
  2015. .network files.
  2016. * The "ask-password" framework used to query for LUKS harddisk
  2017. passwords or SSL passwords during boot gained support for
  2018. caching passwords in the kernel keyring, if it is
  2019. available. This makes sure that the user only has to type in
  2020. a passphrase once if there are multiple objects to unlock
  2021. with the same one. Previously, such password caching was
  2022. available only when Plymouth was used; this moves the
  2023. caching logic into the systemd codebase itself. The
  2024. "systemd-ask-password" utility gained a new --keyname=
  2025. switch to control which kernel keyring key to use for
  2026. caching a password in. This functionality is also useful for
  2027. enabling display managers such as gdm to automatically
  2028. unlock the user's GNOME keyring if its passphrase, the
  2029. user's password and the harddisk password are the same, if
  2030. gdm-autologin is used.
  2031. * When downloading tar or raw images using "machinectl
  2032. pull-tar" or "machinectl pull-raw", a matching ".nspawn"
  2033. file is now also downloaded, if it is available and stored
  2034. next to the image file.
  2035. * Units of type ".socket" gained a new boolean setting
  2036. Writable= which is only useful in conjunction with
  2037. ListenSpecial=. If true, enables opening the specified
  2038. special file in O_RDWR mode rather than O_RDONLY mode.
  2039. * systemd-rfkill has been reworked to become a singleton
  2040. service that is activated through /dev/rfkill on each rfkill
  2041. state change and saves the settings to disk. This way,
  2042. systemd-rfkill is now compatible with devices that exist
  2043. only intermittendly, and even restores state if the previous
  2044. system shutdown was abrupt rather than clean.
  2045. * The journal daemon gained support for vacuuming old journal
  2046. files controlled by the number of files that shall remain,
  2047. in addition to the already existing control by size and by
  2048. date. This is useful as journal interleaving performance
  2049. degrades with too many separate journal files, and allows
  2050. putting an effective limit on them. The new setting defaults
  2051. to 100, but this may be changed by setting SystemMaxFiles=
  2052. and RuntimeMaxFiles= in journald.conf. Also, the
  2053. "journalctl" tool gained the new --vacuum-files= switch to
  2054. manually vacuum journal files to leave only the specified
  2055. number of files in place.
  2056. * udev will now create /dev/disk/by-path links for ATA devices
  2057. on kernels where that is supported.
  2058. * Galician, Serbian, Turkish and Korean translations were added.
  2059. Contributions from: Aaro Koskinen, Alban Crequy, Beniamino
  2060. Galvani, Benjamin Robin, Branislav Blaskovic, Chen-Han Hsiao
  2061. (Stanley), Daniel Buch, Daniel Machon, Daniel Mack, David
  2062. Herrmann, David Milburn, doubleodoug, Evgeny Vereshchagin,
  2063. Felipe Franciosi, Filipe Brandenburger, Fran Dieguez, Gabriel
  2064. de Perthuis, Georg Müller, Hans de Goede, Hendrik Brueckner,
  2065. Ivan Shapovalov, Jacob Keller, Jan Engelhardt, Jan Janssen,
  2066. Jan Synacek, Jens Kuske, Karel Zak, Kay Sievers, Krzesimir
  2067. Nowak, Krzysztof Kotlenga, Lars Uebernickel, Lennart
  2068. Poettering, Lukas Nykryn, Łukasz Stelmach, Maciej Wereski,
  2069. Marcel Holtmann, Marius Thesing, Martin Pitt, Michael Biebl,
  2070. Michael Gebetsroither, Michal Schmidt, Michal Sekletar, Mike
  2071. Gilbert, Muhammet Kara, nazgul77, Nicolas Cornu, NoXPhasma,
  2072. Olof Johansson, Patrik Flykt, Pawel Szewczyk, reverendhomer,
  2073. Ronny Chevalier, Sangjung Woo, Seong-ho Cho, Susant Sahani,
  2074. Sylvain Plantefève, Thomas Haller, Thomas Hindoe Paaboel
  2075. Andersen, Tom Gundersen, Tom Lyon, Viktar Vauchkevich,
  2076. Zbigniew Jędrzejewski-Szmek, Марко М. Костић
  2077. — Berlin, 2015-10-07
  2078. CHANGES WITH 226:
  2079. * The DHCP implementation of systemd-networkd gained a set of
  2080. new features:
  2081. - The DHCP server now supports emitting DNS and NTP
  2082. information. It may be enabled and configured via
  2083. EmitDNS=, DNS=, EmitNTP=, and NTP=. If transmission of DNS
  2084. and NTP information is enabled, but no servers are
  2085. configured, the corresponding uplink information (if there
  2086. is any) is propagated.
  2087. - Server and client now support transmission and reception
  2088. of timezone information. It can be configured via the
  2089. newly introduced network options UseTimezone=,
  2090. EmitTimezone=, and Timezone=. Transmission of timezone
  2091. information is enabled between host and containers by
  2092. default now: the container will change its local timezone
  2093. to what the host has set.
  2094. - Lease timeouts can now be configured via
  2095. MaxLeaseTimeSec= and DefaultLeaseTimeSec=.
  2096. - The DHCP server improved on the stability of
  2097. leases. Clients are more likely to get the same lease
  2098. information back, even if the server loses state.
  2099. - The DHCP server supports two new configuration options to
  2100. control the lease address pool metrics, PoolOffset= and
  2101. PoolSize=.
  2102. * The encapsulation limit of tunnels in systemd-networkd may
  2103. now be configured via 'EncapsulationLimit='. It allows
  2104. modifying the maximum additional levels of encapsulation
  2105. that are permitted to be prepended to a packet.
  2106. * systemd now supports the concept of user buses replacing
  2107. session buses, if used with dbus-1.10 (and enabled via dbus
  2108. --enable-user-session). It previously only supported this on
  2109. kdbus-enabled systems, and this release expands this to
  2110. 'dbus-daemon' systems.
  2111. * systemd-networkd now supports predictable interface names
  2112. for virtio devices.
  2113. * systemd now optionally supports the new Linux kernel
  2114. "unified" control group hierarchy. If enabled via the kernel
  2115. command-line option 'systemd.unified_cgroup_hierarchy=1',
  2116. systemd will try to mount the unified cgroup hierarchy
  2117. directly on /sys/fs/cgroup. If not enabled, or not
  2118. available, systemd will fall back to the legacy cgroup
  2119. hierarchy setup, as before. Host system and containers can
  2120. mix and match legacy and unified hierarchies as they
  2121. wish. nspawn understands the $UNIFIED_CGROUP_HIERARCHY
  2122. environment variable to individually select the hierarchy to
  2123. use for executed containers. By default, nspawn will use the
  2124. unified hierarchy for the containers if the host uses the
  2125. unified hierarchy, and the legacy hierarchy otherwise.
  2126. Please note that at this point the unified hierarchy is an
  2127. experimental kernel feature and is likely to change in one
  2128. of the next kernel releases. Therefore, it should not be
  2129. enabled by default in downstream distributions yet. The
  2130. minimum required kernel version for the unified hierarchy to
  2131. work is 4.2. Note that when the unified hierarchy is used
  2132. for the first time delegated access to controllers is
  2133. safe. Because of this systemd-nspawn containers will get
  2134. access to controllers now, as will systemd user
  2135. sessions. This means containers and user sessions may now
  2136. manage their own resources, partitioning up what the system
  2137. grants them.
  2138. * A new special scope unit "init.scope" has been introduced
  2139. that encapsulates PID 1 of the system. It may be used to
  2140. determine resource usage and enforce resource limits on PID
  2141. 1 itself. PID 1 hence moved out of the root of the control
  2142. group tree.
  2143. * The cgtop tool gained support for filtering out kernel
  2144. threads when counting tasks in a control group. Also, the
  2145. count of processes is now recursively summed up by
  2146. default. Two options -k and --recursive= have been added to
  2147. revert to old behaviour. The tool has also been updated to
  2148. work correctly in containers now.
  2149. * systemd-nspawn's --bind= and --bind-ro= options have been
  2150. extended to allow creation of non-recursive bind mounts.
  2151. * libsystemd gained two new calls sd_pid_get_cgroup() and
  2152. sd_peer_get_cgroup() which return the control group path of
  2153. a process or peer of a connected AF_UNIX socket. This
  2154. function call is particularly useful when implementing
  2155. delegated subtrees support in the control group hierarchy.
  2156. * The "sd-event" event loop API of libsystemd now supports
  2157. correct dequeuing of real-time signals, without losing
  2158. signal events.
  2159. * When systemd requests a PolicyKit decision when managing
  2160. units it will now add additional fields to the request,
  2161. including unit name and desired operation. This enables more
  2162. powerful PolicyKit policies, that make decisions depending
  2163. on these parameters.
  2164. * nspawn learnt support for .nspawn settings files, that may
  2165. accompany the image files or directories of containers, and
  2166. may contain additional settings for the container. This is
  2167. an alternative to configuring container parameters via the
  2168. nspawn command line.
  2169. Contributions from: Cristian Rodríguez, Daniel Mack, David
  2170. Herrmann, Eugene Yakubovich, Evgeny Vereshchagin, Filipe
  2171. Brandenburger, Hans de Goede, Jan Alexander Steffens, Jan
  2172. Synacek, Kay Sievers, Lennart Poettering, Mangix, Marcel
  2173. Holtmann, Martin Pitt, Michael Biebl, Michael Chapman, Michal
  2174. Sekletar, Peter Hutterer, Piotr Drąg, reverendhomer, Robin
  2175. Hack, Susant Sahani, Sylvain Pasche, Thomas Hindoe Paaboel
  2176. Andersen, Tom Gundersen, Torstein Husebø
  2177. — Berlin, 2015-09-08
  2178. CHANGES WITH 225:
  2179. * machinectl gained a new verb 'shell' which opens a fresh
  2180. shell on the target container or the host. It is similar to
  2181. the existing 'login' command of machinectl, but spawns the
  2182. shell directly without prompting for username or
  2183. password. The pseudo machine '.host' now refers to the local
  2184. host and is used by default. Hence, 'machinectl shell' can
  2185. be used as replacement for 'su -' which spawns a session as
  2186. a fresh systemd unit in a way that is fully isolated from
  2187. the originating session.
  2188. * systemd-networkd learned to cope with private-zone DHCP
  2189. options and allows other programs to query the values.
  2190. * SELinux access control when enabling/disabling units is no
  2191. longer enforced with this release. The previous
  2192. implementation was incorrect, and a new corrected
  2193. implementation is not yet available. As unit file operations
  2194. are still protected via PolicyKit and D-Bus policy this is
  2195. not a security problem. Yet, distributions which care about
  2196. optimal SELinux support should probably not stabilize on
  2197. this release.
  2198. * sd-bus gained support for matches of type "arg0has=", that
  2199. test for membership of strings in string arrays sent in bus
  2200. messages.
  2201. * systemd-resolved now dumps the contents of its DNS and LLMNR
  2202. caches to the logs on reception of the SIGUSR1 signal. This
  2203. is useful to debug DNS behaviour.
  2204. * The coredumpctl tool gained a new --directory= option to
  2205. operate on journal files in a specific directory.
  2206. * "systemctl reboot" and related commands gained a new
  2207. "--message=" option which may be used to set a free-text
  2208. wall message when shutting down or rebooting the
  2209. system. This message is also logged, which is useful for
  2210. figuring out the reason for a reboot or shutdown a
  2211. posteriori.
  2212. * The "systemd-resolve-host" tool's -i switch now takes
  2213. network interface numbers as alternative to interface names.
  2214. * A new unit file setting for services has been introduced:
  2215. UtmpMode= allows configuration of how precisely systemd
  2216. handles utmp and wtmp entries for the service if this is
  2217. enabled. This allows writing services that appear similar to
  2218. user sessions in the output of the "w", "who", "last" and
  2219. "lastlog" tools.
  2220. * systemd-resolved will now locally synthesize DNS resource
  2221. records for the "localhost" and "gateway" domains as well as
  2222. the local hostname. This should ensure that clients querying
  2223. RRs via resolved will get similar results as those going via
  2224. NSS, if nss-myhostname is enabled.
  2225. Contributions from: Alastair Hughes, Alex Crawford, Daniel
  2226. Mack, David Herrmann, Dimitri John Ledkov, Eric Kostrowski,
  2227. Evgeny Vereshchagin, Felipe Sateler, HATAYAMA Daisuke, Jan
  2228. Pokorný, Jan Synacek, Johnny Robeson, Karel Zak, Kay Sievers,
  2229. Kefeng Wang, Lennart Poettering, Major Hayden, Marcel
  2230. Holtmann, Markus Elfring, Martin Mikkelsen, Martin Pitt, Matt
  2231. Turner, Maxim Mikityanskiy, Michael Biebl, Namhyung Kim,
  2232. Nicolas Cornu, Owen W. Taylor, Patrik Flykt, Peter Hutterer,
  2233. reverendhomer, Richard Maw, Ronny Chevalier, Seth Jennings,
  2234. Stef Walter, Susant Sahani, Thomas Blume, Thomas Hindoe
  2235. Paaboel Andersen, Thomas Meyer, Tom Gundersen, Vincent Batts,
  2236. WaLyong Cho, Zbigniew Jędrzejewski-Szmek
  2237. — Berlin, 2015-08-27
  2238. CHANGES WITH 224:
  2239. * The systemd-efi-boot-generator functionality was merged into
  2240. systemd-gpt-auto-generator.
  2241. * systemd-networkd now supports Group Policy for vxlan
  2242. devices. It can be enabled via the new boolean configuration
  2243. option called 'GroupPolicyExtension='.
  2244. Contributions from: Andreas Kempf, Christian Hesse, Daniel Mack, David
  2245. Herrmann, Herman Fries, Johannes Nixdorf, Kay Sievers, Lennart
  2246. Poettering, Peter Hutterer, Susant Sahani, Tom Gundersen
  2247. — Berlin, 2015-07-31
  2248. CHANGES WITH 223:
  2249. * The python-systemd code has been removed from the systemd repository.
  2250. A new repository has been created which accommodates the code from
  2251. now on, and we kindly ask distributions to create a separate package
  2252. for this:
  2253. * The systemd daemon will now reload its main configuration
  2254. (/etc/systemd/system.conf) on daemon-reload.
  2255. * sd-dhcp now exposes vendor specific extensions via
  2256. sd_dhcp_lease_get_vendor_specific().
  2257. * systemd-networkd gained a number of new configuration options.
  2258. - A new boolean configuration option for TAP devices called
  2259. 'VNetHeader='. If set, the IFF_VNET_HDR flag is set for the
  2260. device, thus allowing to send and receive GSO packets.
  2261. - A new tunnel configuration option called 'CopyDSCP='.
  2262. If enabled, the DSCP field of ip6 tunnels is copied into the
  2263. decapsulated packet.
  2264. - A set of boolean bridge configuration options were added.
  2265. 'UseBPDU=', 'HairPin=', 'FastLeave=', 'AllowPortToBeRoot=',
  2266. and 'UnicastFlood=' are now parsed by networkd and applied to the
  2267. respective bridge link device via the respective IFLA_BRPORT_*
  2268. netlink attribute.
  2269. - A new string configuration option to override the hostname sent
  2270. to a DHCP server, called 'Hostname='. If set and 'SendHostname='
  2271. is true, networkd will use the configured hostname instead of the
  2272. system hostname when sending DHCP requests.
  2273. - A new tunnel configuration option called 'IPv6FlowLabel='. If set,
  2274. networkd will configure the IPv6 flow-label of the tunnel device
  2275. according to RFC2460.
  2276. - The 'macvtap' virtual network devices are now supported, similar to
  2277. the already supported 'macvlan' devices.
  2278. * systemd-resolved now implements RFC5452 to improve resilience against
  2279. cache poisoning. Additionally, source port randomization is enabled
  2280. by default to further protect against DNS spoofing attacks.
  2281. * nss-mymachines now supports translating UIDs and GIDs of running
  2282. containers with user-namespaces enabled. If a container 'foo'
  2283. translates a host uid 'UID' to the container uid 'TUID', then
  2284. nss-mymachines will also map uid 'UID' to/from username 'vu-foo-TUID'
  2285. (with 'foo' and 'TUID' replaced accordingly). Similarly, groups are
  2286. mapped as 'vg-foo-TGID'.
  2287. Contributions from: Beniamino Galvani, cee1, Christian Hesse, Daniel
  2288. Buch, Daniel Mack, daurnimator, David Herrmann, Dimitri John Ledkov,
  2289. HATAYAMA Daisuke, Ivan Shapovalov, Jan Alexander Steffens (heftig),
  2290. Johan Ouwerkerk, Jose Carlos Venegas Munoz, Karel Zak, Kay Sievers,
  2291. Lennart Poettering, Lidong Zhong, Martin Pitt, Michael Biebl, Michael
  2292. Olbrich, Michal Schmidt, Michal Sekletar, Mike Gilbert, Namhyung Kim,
  2293. Nick Owens, Peter Hutterer, Richard Maw, Steven Allen, Sungbae Yoo,
  2294. Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel Andersen, Tom
  2295. Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Vito Caputo,
  2296. Vivenzio Pagliari, Zbigniew Jędrzejewski-Szmek
  2297. — Berlin, 2015-07-29
  2298. CHANGES WITH 222:
  2299. * udev does not longer support the WAIT_FOR_SYSFS= key in udev rules.
  2300. There are no known issues with current sysfs, and udev does not need
  2301. or should be used to work around such bugs.
  2302. * udev does no longer enable USB HID power management. Several reports
  2303. indicate, that some devices cannot handle that setting.
  2304. * The udev accelerometer helper was removed. The functionality
  2305. is now fully included in iio-sensor-proxy. But this means,
  2306. older iio-sensor-proxy versions will no longer provide
  2307. accelerometer/orientation data with this systemd version.
  2308. Please upgrade iio-sensor-proxy to version 1.0.
  2309. * networkd gained a new configuration option IPv6PrivacyExtensions=
  2310. which enables IPv6 privacy extensions (RFC 4941, "Privacy Extensions
  2311. for Stateless Address") on selected networks.
  2312. * For the sake of fewer build-time dependencies and less code in the
  2313. main repository, the python bindings are about to be removed in the
  2314. next release. A new repository has been created which accommodates
  2315. the code from now on, and we kindly ask distributions to create a
  2316. separate package for this. The removal will take place in v223.
  2318. Contributions from: Abdo Roig-Maranges, Andrew Eikum, Bastien Nocera,
  2319. Cédric Delmas, Christian Hesse, Christos Trochalakis, Daniel Mack,
  2320. daurnimator, David Herrmann, Dimitri John Ledkov, Eric Biggers, Eric
  2321. Cook, Felipe Sateler, Geert Jansen, Gerd Hoffmann, Gianpaolo Macario,
  2322. Greg Kroah-Hartman, Iago López Galeiras, Jan Alexander Steffens
  2323. (heftig), Jan Engelhardt, Jay Strict, Kay Sievers, Lennart Poettering,
  2324. Markus Knetschke, Martin Pitt, Michael Biebl, Michael Marineau, Michal
  2325. Sekletar, Miguel Bernal Marin, Peter Hutterer, Richard Maw, rinrinne,
  2326. Susant Sahani, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
  2327. Husebø, Vedran Miletić, WaLyong Cho, Zbigniew Jędrzejewski-Szmek
  2328. — Berlin, 2015-07-07
  2329. CHANGES WITH 221:
  2330. * The sd-bus.h and sd-event.h APIs have now been declared
  2331. stable and have been added to the official interface of
  2332. sd-bus implements an alternative D-Bus client
  2333. library, that is relatively easy to use, very efficient and
  2334. supports both classic D-Bus as well as kdbus as transport
  2335. backend. sd-event is a generic event loop abstraction that
  2336. is built around Linux epoll, but adds features such as event
  2337. prioritization or efficient timer handling. Both APIs are good
  2338. choices for C programs looking for a bus and/or event loop
  2339. implementation that is minimal and does not have to be
  2340. portable to other kernels.
  2341. * kdbus support is no longer compile-time optional. It is now
  2342. always built-in. However, it can still be disabled at
  2343. runtime using the kdbus=0 kernel command line setting, and
  2344. that setting may be changed to default to off, by specifying
  2345. --disable-kdbus at build-time. Note though that the kernel
  2346. command line setting has no effect if the kdbus.ko kernel
  2347. module is not installed, in which case kdbus is (obviously)
  2348. also disabled. We encourage all downstream distributions to
  2349. begin testing kdbus by adding it to the kernel images in the
  2350. development distributions, and leaving kdbus support in
  2351. systemd enabled.
  2352. * The minimal required util-linux version has been bumped to
  2353. 2.26.
  2354. * Support for chkconfig (--enable-chkconfig) was removed in
  2355. favor of calling an abstraction tool
  2356. /lib/systemd/systemd-sysv-install. This needs to be
  2357. implemented for your distribution. See "SYSV INIT.D SCRIPTS"
  2358. in README for details.
  2359. * If there's a systemd unit and a SysV init script for the
  2360. same service name, and the user executes "systemctl enable"
  2361. for it (or a related call), then this will now enable both
  2362. (or execute the related operation on both), not just the
  2363. unit.
  2364. * The libudev API documentation has been converted from gtkdoc
  2365. into man pages.
  2366. * gudev has been removed from the systemd tree, it is now an
  2367. external project.
  2368. * The systemd-cgtop tool learnt a new --raw switch to generate
  2369. "raw" (machine parsable) output.
  2370. * networkd's IPForwarding= .network file setting learnt the
  2371. new setting "kernel", which ensures that networkd does not
  2372. change the IP forwarding sysctl from the default kernel
  2373. state.
  2374. * The systemd-logind bus API now exposes a new boolean
  2375. property "Docked" that reports whether logind considers the
  2376. system "docked", i.e. connected to a docking station or not.
  2377. Contributions from: Alex Crawford, Andreas Pokorny, Andrei
  2378. Borzenkov, Charles Duffy, Colin Guthrie, Cristian Rodríguez,
  2379. Daniele Medri, Daniel Hahler, Daniel Mack, David Herrmann,
  2380. David Mohr, Dimitri John Ledkov, Djalal Harouni, dslul, Ed
  2381. Swierk, Eric Cook, Filipe Brandenburger, Gianpaolo Macario,
  2382. Harald Hoyer, Iago López Galeiras, Igor Vuk, Jan Synacek,
  2383. Jason Pleau, Jason S. McMullan, Jean Delvare, Jeff Huang,
  2384. Jonathan Boulle, Karel Zak, Kay Sievers, kloun, Lennart
  2385. Poettering, Marc-Antoine Perennou, Marcel Holtmann, Mario
  2386. Limonciello, Martin Pitt, Michael Biebl, Michael Olbrich,
  2387. Michal Schmidt, Mike Gilbert, Nick Owens, Pablo Lezaeta Reyes,
  2388. Patrick Donnelly, Pavel Odvody, Peter Hutterer, Philip
  2389. Withnall, Ronny Chevalier, Simon McVittie, Susant Sahani,
  2390. Thomas Hindoe Paaboel Andersen, Tom Gundersen, Torstein
  2391. Husebø, Umut Tezduyar Lindskog, Viktar Vauchkevich, Werner
  2392. Fink, Zbigniew Jędrzejewski-Szmek
  2393. — Berlin, 2015-06-19
  2394. CHANGES WITH 220:
  2395. * The gudev library has been extracted into a separate repository
  2396. available at:
  2397. It is now managed as part of the Gnome project. Distributions
  2398. are recommended to pass --disable-gudev to systemd and use
  2399. gudev from the Gnome project instead. gudev is still included
  2400. in systemd, for now. It will be removed soon, though. Please
  2401. also see the announcement-thread on systemd-devel:
  2403. * systemd now exposes a CPUUsageNSec= property for each
  2404. service unit on the bus, that contains the overall consumed
  2405. CPU time of a service (the sum of what each process of the
  2406. service consumed). This value is only available if
  2407. CPUAccounting= is turned on for a service, and is then shown
  2408. in the "systemctl status" output.
  2409. * Support for configuring alternative mappings of the old SysV
  2410. runlevels to systemd targets has been removed. They are now
  2411. hardcoded in a way that runlevels 2, 3, 4 all map to
  2412. and 5 to (which
  2413. previously was already the default behaviour).
  2414. * The auto-mounter logic gained support for mount point
  2415. expiry, using a new TimeoutIdleSec= setting in .automount
  2416. units. (Also available as x-systemd.idle-timeout= in /etc/fstab).
  2417. * The EFI System Partition (ESP) as mounted to /boot by
  2418. systemd-efi-boot-generator will now be unmounted
  2419. automatically after 2 minutes of not being used. This should
  2420. minimize the risk of ESP corruptions.
  2421. * New /etc/fstab options x-systemd.requires= and
  2422. x-systemd.requires-mounts-for= are now supported to express
  2423. additional dependencies for mounts. This is useful for
  2424. journalling file systems that support external journal
  2425. devices or overlay file systems that require underlying file
  2426. systems to be mounted.
  2427. * systemd does not support direct live-upgrades (via systemctl
  2428. daemon-reexec) from versions older than v44 anymore. As no
  2429. distribution we are aware of shipped such old versions in a
  2430. stable release this should not be problematic.
  2431. * When systemd forks off a new per-connection service instance
  2432. it will now set the $REMOTE_ADDR environment variable to the
  2433. remote IP address, and $REMOTE_PORT environment variable to
  2434. the remote IP port. This behaviour is similar to the
  2435. corresponding environment variables defined by CGI.
  2436. * systemd-networkd gained support for uplink failure
  2437. detection. The BindCarrier= option allows binding interface
  2438. configuration dynamically to the link sense of other
  2439. interfaces. This is useful to achieve behaviour like in
  2440. network switches.
  2441. * systemd-networkd gained support for configuring the DHCP
  2442. client identifier to use when requesting leases.
  2443. * systemd-networkd now has a per-network UseNTP= option to
  2444. configure whether NTP server information acquired via DHCP
  2445. is passed on to services like systemd-timesyncd.
  2446. * systemd-networkd gained support for vti6 tunnels.
  2447. * Note that systemd-networkd manages the sysctl variable
  2448. /proc/sys/net/ipv[46]/conf/*/forwarding for each interface
  2449. it is configured for since v219. The variable controls IP
  2450. forwarding, and is a per-interface alternative to the global
  2451. /proc/sys/net/ipv[46]/ip_forward. This setting is
  2452. configurable in the IPForward= option, which defaults to
  2453. "no". This means if networkd is used for an interface it is
  2454. no longer sufficient to set the global sysctl option to turn
  2455. on IP forwarding! Instead, the .network file option
  2456. IPForward= needs to be turned on! Note that the
  2457. implementation of this behaviour was broken in v219 and has
  2458. been fixed in v220.
  2459. * Many bonding and vxlan options are now configurable in
  2460. systemd-networkd.
  2461. * systemd-nspawn gained a new --property= setting to set unit
  2462. properties for the container scope. This is useful for
  2463. setting resource parameters (e.g. "CPUShares=500") on
  2464. containers started from the command line.
  2465. * systemd-nspawn gained a new --private-users= switch to make
  2466. use of user namespacing available on recent Linux kernels.
  2467. * systemd-nspawn may now be called as part of a shell pipeline
  2468. in which case the pipes used for stdin and stdout are passed
  2469. directly to the process invoked in the container, without
  2470. indirection via a pseudo tty.
  2471. * systemd-nspawn gained a new switch to control the UNIX
  2472. signal to use when killing the init process of the container
  2473. when shutting down.
  2474. * systemd-nspawn gained a new --overlay= switch for mounting
  2475. overlay file systems into the container using the new kernel
  2476. overlayfs support.
  2477. * When a container image is imported via systemd-importd and
  2478. the host file system is not btrfs, a loopback block device
  2479. file is created in /var/lib/machines.raw with a btrfs file
  2480. system inside. It is then mounted to /var/lib/machines to
  2481. enable btrfs features for container management. The loopback
  2482. file and btrfs file system is grown as needed when container
  2483. images are imported via systemd-importd.
  2484. * systemd-machined/systemd-importd gained support for btrfs
  2485. quota, to enforce container disk space limits on disk. This
  2486. is exposed in "machinectl set-limit".
  2487. * systemd-importd now can import containers from local .tar,
  2488. .raw and .qcow2 images, and export them to .tar and .raw. It
  2489. can also import dkr v2 images now from the network (on top
  2490. of v1 as before).
  2491. * systemd-importd gained support for verifying downloaded
  2492. images with gpg2 (previously only gpg1 was supported).
  2493. * systemd-machined, systemd-logind, systemd: most bus calls
  2494. are now accessible to unprivileged processes via
  2495. PolicyKit. Also, systemd-logind will now allow users to kill
  2496. their own sessions without further privileges or
  2497. authorization.
  2498. * systemd-shutdownd has been removed. This service was
  2499. previously responsible for implementing scheduled shutdowns
  2500. as exposed in /usr/bin/shutdown's time parameter. This
  2501. functionality has now been moved into systemd-logind and is
  2502. accessible via a bus interface.
  2503. * "systemctl reboot" gained a new switch --firmware-setup that
  2504. can be used to reboot into the EFI firmware setup, if that
  2505. is available. systemd-logind now exposes an API on the bus
  2506. to trigger such reboots, in case graphical desktop UIs want
  2507. to cover this functionality.
  2508. * "systemctl enable", "systemctl disable" and "systemctl mask"
  2509. now support a new "--now" switch. If specified the units
  2510. that are enabled will also be started, and the ones
  2511. disabled/masked also stopped.
  2512. * The Gummiboot EFI boot loader tool has been merged into
  2513. systemd, and renamed to "systemd-boot". The bootctl tool has been
  2514. updated to support systemd-boot.
  2515. * An EFI kernel stub has been added that may be used to create
  2516. kernel EFI binaries that contain not only the actual kernel,
  2517. but also an initrd, boot splash, command line and OS release
  2518. information. This combined binary can then be signed as a
  2519. single image, so that the firmware can verify it all in one
  2520. step. systemd-boot has special support for EFI binaries created
  2521. like this and can extract OS release information from them
  2522. and show them in the boot menu. This functionality is useful
  2523. to implement cryptographically verified boot schemes.
  2524. * Optional support has been added to systemd-fsck to pass
  2525. fsck's progress report to an AF_UNIX socket in the file
  2526. system.
  2527. * udev will no longer create device symlinks for all block
  2528. devices by default. A blacklist for excluding special block
  2529. devices from this logic has been turned into a whitelist
  2530. that requires picking block devices explicitly that require
  2531. device symlinks.
  2532. * A new (currently still internal) API sd-device.h has been
  2533. added to libsystemd. This modernized API is supposed to
  2534. replace libudev eventually. In fact, already much of libudev
  2535. is now just a wrapper around sd-device.h.
  2536. * A new hwdb database for storing metadata about pointing
  2537. stick devices has been added.
  2538. * systemd-tmpfiles gained support for setting file attributes
  2539. similar to the "chattr" tool with new 'h' and 'H' lines.
  2540. * systemd-journald will no longer unconditionally set the
  2541. btrfs NOCOW flag on new journal files. This is instead done
  2542. with tmpfiles snippet using the new 'h' line type. This
  2543. allows easy disabling of this logic, by masking the
  2544. journal-nocow.conf tmpfiles file.
  2545. * systemd-journald will now translate audit message types to
  2546. human readable identifiers when writing them to the
  2547. journal. This should improve readability of audit messages.
  2548. * The LUKS logic gained support for the offset= and skip=
  2549. options in /etc/crypttab, as previously implemented by
  2550. Debian.
  2551. * /usr/lib/os-release gained a new optional field VARIANT= for
  2552. distributions that support multiple variants (such as a
  2553. desktop edition, a server edition, ...)
  2554. Contributions from: Aaro Koskinen, Adam Goode, Alban Crequy,
  2555. Alberto Fanjul Alonso, Alexander Sverdlin, Alex Puchades, Alin
  2556. Rauta, Alison Chaiken, Andrew Jones, Arend van Spriel,
  2557. Benedikt Morbach, Benjamin Franzke, Benjamin Tissoires, Blaž
  2558. Tomažič, Chris Morgan, Chris Morin, Colin Walters, Cristian
  2559. Rodríguez, Daniel Buch, Daniel Drake, Daniele Medri, Daniel
  2560. Mack, Daniel Mustieles, daurnimator, Davide Bettio, David
  2561. Herrmann, David Strauss, Didier Roche, Dimitri John Ledkov,
  2562. Eric Cook, Gavin Li, Goffredo Baroncelli, Hannes Reinecke,
  2563. Hans de Goede, Hans-Peter Deifel, Harald Hoyer, Iago López
  2564. Galeiras, Ivan Shapovalov, Jan Engelhardt, Jan Janssen, Jan
  2565. Pazdziora, Jan Synacek, Jasper St. Pierre, Jay Faulkner, John
  2566. Paul Adrian Glaubitz, Jonathon Gilbert, Karel Zak, Kay
  2567. Sievers, Koen Kooi, Lennart Poettering, Lubomir Rintel, Lucas
  2568. De Marchi, Lukas Nykryn, Lukas Rusak, Lukasz Skalski, Łukasz
  2569. Stelmach, Mantas Mikulėnas, Marc-Antoine Perennou, Marcel
  2570. Holtmann, Martin Pitt, Mathieu Chevrier, Matthew Garrett,
  2571. Michael Biebl, Michael Marineau, Michael Olbrich, Michal
  2572. Schmidt, Michal Sekletar, Mirco Tischler, Nir Soffer, Patrik
  2573. Flykt, Pavel Odvody, Peter Hutterer, Peter Lemenkov, Peter
  2574. Waller, Piotr Drąg, Raul Gutierrez S, Richard Maw, Ronny
  2575. Chevalier, Ross Burton, Sebastian Rasmussen, Sergey Ptashnick,
  2576. Seth Jennings, Shawn Landden, Simon Farnsworth, Stefan Junker,
  2577. Stephen Gallagher, Susant Sahani, Sylvain Plantefève, Thomas
  2578. Haller, Thomas Hindoe Paaboel Andersen, Tobias Hunger, Tom
  2579. Gundersen, Torstein Husebø, Umut Tezduyar Lindskog, Will
  2580. Woods, Zachary Cook, Zbigniew Jędrzejewski-Szmek
  2581. — Berlin, 2015-05-22
  2582. CHANGES WITH 219:
  2583. * Introduce a new API "sd-hwdb.h" for querying the hardware
  2584. metadata database. With this minimal interface one can query
  2585. and enumerate the udev hwdb, decoupled from the old libudev
  2586. library. libudev's interface for this is now only a wrapper
  2587. around sd-hwdb. A new tool systemd-hwdb has been added to
  2588. interface with and update the database.
  2589. * When any of systemd's tools copies files (for example due to
  2590. tmpfiles' C lines) a btrfs reflink will attempted first,
  2591. before bytewise copying is done.
  2592. * systemd-nspawn gained a new --ephemeral switch. When
  2593. specified a btrfs snapshot is taken of the container's root
  2594. directory, and immediately removed when the container
  2595. terminates again. Thus, a container can be started whose
  2596. changes never alter the container's root directory, and are
  2597. lost on container termination. This switch can also be used
  2598. for starting a container off the root file system of the
  2599. host without affecting the host OS. This switch is only
  2600. available on btrfs file systems.
  2601. * systemd-nspawn gained a new --template= switch. It takes the
  2602. path to a container tree to use as template for the tree
  2603. specified via --directory=, should that directory be
  2604. missing. This allows instantiating containers dynamically,
  2605. on first run. This switch is only available on btrfs file
  2606. systems.
  2607. * When a .mount unit refers to a mount point on which multiple
  2608. mounts are stacked, and the .mount unit is stopped all of
  2609. the stacked mount points will now be unmounted until no
  2610. mount point remains.
  2611. * systemd now has an explicit notion of supported and
  2612. unsupported unit types. Jobs enqueued for unsupported unit
  2613. types will now fail with an "unsupported" error code. More
  2614. specifically .swap, .automount and .device units are not
  2615. supported in containers, .busname units are not supported on
  2616. non-kdbus systems. .swap and .automount are also not
  2617. supported if their respective kernel compile time options
  2618. are disabled.
  2619. * machinectl gained support for two new "copy-from" and
  2620. "copy-to" commands for copying files from a running
  2621. container to the host or vice versa.
  2622. * machinectl gained support for a new "bind" command to bind
  2623. mount host directories into local containers. This is
  2624. currently only supported for nspawn containers.
  2625. * networkd gained support for configuring bridge forwarding
  2626. database entries (fdb) from .network files.
  2627. * A new tiny daemon "systemd-importd" has been added that can
  2628. download container images in tar, raw, qcow2 or dkr formats,
  2629. and make them available locally in /var/lib/machines, so
  2630. that they can run as nspawn containers. The daemon can GPG
  2631. verify the downloads (not supported for dkr, since it has no
  2632. provisions for verifying downloads). It will transparently
  2633. decompress bz2, xz, gzip compressed downloads if necessary,
  2634. and restore sparse files on disk. The daemon uses privilege
  2635. separation to ensure the actual download logic runs with
  2636. fewer privileges than the daemon itself. machinectl has
  2637. gained new commands "pull-tar", "pull-raw" and "pull-dkr" to
  2638. make the functionality of importd available to the
  2639. user. With this in place the Fedora and Ubuntu "Cloud"
  2640. images can be downloaded and booted as containers unmodified
  2641. (the Fedora images lack the appropriate GPG signature files
  2642. currently, so they cannot be verified, but this will change
  2643. soon, hopefully). Note that downloading images is currently
  2644. only fully supported on btrfs.
  2645. * machinectl is now able to list container images found in
  2646. /var/lib/machines, along with some metadata about sizes of
  2647. disk and similar. If the directory is located on btrfs and
  2648. quota is enabled, this includes quota display. A new command
  2649. "image-status" has been added that shows additional
  2650. information about images.
  2651. * machinectl is now able to clone container images
  2652. efficiently, if the underlying file system (btrfs) supports
  2653. it, with the new "machinectl clone" command. It also
  2654. gained commands for renaming and removing images, as well as
  2655. marking them read-only or read-write (supported also on
  2656. legacy file systems).
  2657. * networkd gained support for collecting LLDP network
  2658. announcements, from hardware that supports this. This is
  2659. shown in networkctl output.
  2660. * systemd-run gained support for a new -t (--pty) switch for
  2661. invoking a binary on a pty whose input and output is
  2662. connected to the invoking terminal. This allows executing
  2663. processes as system services while interactively
  2664. communicating with them via the terminal. Most interestingly
  2665. this is supported across container boundaries. Invoking
  2666. "systemd-run -t /bin/bash" is an alternative to running a
  2667. full login session, the difference being that the former
  2668. will not register a session, nor go through the PAM session
  2669. setup.
  2670. * tmpfiles gained support for a new "v" line type for creating
  2671. btrfs subvolumes. If the underlying file system is a legacy
  2672. file system, this automatically degrades to creating a
  2673. normal directory. Among others /var/lib/machines is now
  2674. created like this at boot, should it be missing.
  2675. * The directory /var/lib/containers/ has been deprecated and
  2676. been replaced by /var/lib/machines. The term "machines" has
  2677. been used in the systemd context as generic term for both
  2678. VMs and containers, and hence appears more appropriate for
  2679. this, as the directory can also contain raw images bootable
  2680. via qemu/kvm.
  2681. * systemd-nspawn when invoked with -M but without --directory=
  2682. or --image= is now capable of searching for the container
  2683. root directory, subvolume or disk image automatically, in
  2684. /var/lib/machines. systemd-nspawn@.service has been updated
  2685. to make use of this, thus allowing it to be used for raw
  2686. disk images, too.
  2687. * A new unit has been introduced that is
  2688. supposed to group all containers/VMs invoked as services on
  2689. the system. systemd-nspawn@.service has been updated to
  2690. integrate with that.
  2691. * machinectl gained a new "start" command, for invoking a
  2692. container as a service. "machinectl start foo" is mostly
  2693. equivalent to "systemctl start systemd-nspawn@foo.service",
  2694. but handles escaping in a nicer way.
  2695. * systemd-nspawn will now mount most of the cgroupfs tree
  2696. read-only into each container, with the exception of the
  2697. container's own subtree in the name=systemd hierarchy.
  2698. * journald now sets the special FS_NOCOW file flag for its
  2699. journal files. This should improve performance on btrfs, by
  2700. avoiding heavy fragmentation when journald's write-pattern
  2701. is used on COW file systems. It degrades btrfs' data
  2702. integrity guarantees for the files to the same levels as for
  2703. ext3/ext4 however. This should be OK though as journald does
  2704. its own data integrity checks and all its objects are
  2705. checksummed on disk. Also, journald should handle btrfs disk
  2706. full events a lot more gracefully now, by processing SIGBUS
  2707. errors, and not relying on fallocate() anymore.
  2708. * When journald detects that journal files it is writing to
  2709. have been deleted it will immediately start new journal
  2710. files.
  2711. * systemd now provides a way to store file descriptors
  2712. per-service in PID 1. This is useful for daemons to ensure
  2713. that fds they require are not lost during a daemon
  2714. restart. The fds are passed to the daemon on the next
  2715. invocation in the same way socket activation fds are
  2716. passed. This is now used by journald to ensure that the
  2717. various sockets connected to all the system's stdout/stderr
  2718. are not lost when journald is restarted. File descriptors
  2719. may be stored in PID 1 via the sd_pid_notify_with_fds() API,
  2720. an extension to sd_notify(). Note that a limit is enforced
  2721. on the number of fds a service can store in PID 1, and it
  2722. defaults to 0, so that no fds may be stored, unless this is
  2723. explicitly turned on.
  2724. * The default TERM variable to use for units connected to a
  2725. terminal, when no other value is explicitly is set is now
  2726. vt220 rather than vt102. This should be fairly safe still,
  2727. but allows PgUp/PgDn work.
  2728. * The /etc/crypttab option header= as known from Debian is now
  2729. supported.
  2730. * "loginctl user-status" and "loginctl session-status" will
  2731. now show the last 10 lines of log messages of the
  2732. user/session following the status output. Similar,
  2733. "machinectl status" will show the last 10 log lines
  2734. associated with a virtual machine or container
  2735. service. (Note that this is usually not the log messages
  2736. done in the VM/container itself, but simply what the
  2737. container manager logs. For nspawn this includes all console
  2738. output however.)
  2739. * "loginctl session-status" without further argument will now
  2740. show the status of the session of the caller. Similar,
  2741. "lock-session", "unlock-session", "activate",
  2742. "enable-linger", "disable-linger" may now be called without
  2743. session/user parameter in which case they apply to the
  2744. caller's session/user.
  2745. * An X11 session scriptlet is now shipped that uploads
  2746. $DISPLAY and $XAUTHORITY into the environment of the systemd
  2747. --user daemon if a session begins. This should improve
  2748. compatibility with X11 enabled applications run as systemd
  2749. user services.
  2750. * Generators are now subject to masking via /etc and /run, the
  2751. same way as unit files.
  2752. * networkd .network files gained support for configuring
  2753. per-link IPv4/IPv6 packet forwarding as well as IPv4
  2754. masquerading. This is by default turned on for veth links to
  2755. containers, as registered by systemd-nspawn. This means that
  2756. nspawn containers run with --network-veth will now get
  2757. automatic routed access to the host's networks without any
  2758. further configuration or setup, as long as networkd runs on
  2759. the host.
  2760. * systemd-nspawn gained the --port= (-p) switch to expose TCP
  2761. or UDP posts of a container on the host. With this in place
  2762. it is possible to run containers with private veth links
  2763. (--network-veth), and have their functionality exposed on
  2764. the host as if their services were running directly on the
  2765. host.
  2766. * systemd-nspawn's --network-veth switch now gained a short
  2767. version "-n", since with the changes above it is now truly
  2768. useful out-of-the-box. The systemd-nspawn@.service has been
  2769. updated to make use of it too by default.
  2770. * systemd-nspawn will now maintain a per-image R/W lock, to
  2771. ensure that the same image is not started more than once
  2772. writable. (It's OK to run an image multiple times
  2773. simultaneously in read-only mode.)
  2774. * systemd-nspawn's --image= option is now capable of
  2775. dissecting and booting MBR and GPT disk images that contain
  2776. only a single active Linux partition. Previously it
  2777. supported only GPT disk images with proper GPT type
  2778. IDs. This allows running cloud images from major
  2779. distributions directly with systemd-nspawn, without
  2780. modification.
  2781. * In addition to collecting mouse dpi data in the udev
  2782. hardware database, there's now support for collecting angle
  2783. information for mouse scroll wheels. The database is
  2784. supposed to guarantee similar scrolling behavior on mice
  2785. that it knows about. There's also support for collecting
  2786. information about Touchpad types.
  2787. * udev's input_id built-in will now also collect touch screen
  2788. dimension data and attach it to probed devices.
  2789. * /etc/os-release gained support for a Distribution Privacy
  2790. Policy link field.
  2791. * networkd gained support for creating "ipvlan", "gretap",
  2792. "ip6gre", "ip6gretap" and "ip6tnl" network devices.
  2793. * systemd-tmpfiles gained support for "a" lines for setting
  2794. ACLs on files.
  2795. * systemd-nspawn will now mount /tmp in the container to
  2796. tmpfs, automatically.
  2797. * systemd now exposes the memory.usage_in_bytes cgroup
  2798. attribute and shows it for each service in the "systemctl
  2799. status" output, if available.
  2800. * When the user presses Ctrl-Alt-Del more than 7x within 2s an
  2801. immediate reboot is triggered. This useful if shutdown is
  2802. hung and is unable to complete, to expedite the
  2803. operation. Note that this kind of reboot will still unmount
  2804. all file systems, and hence should not result in fsck being
  2805. run on next reboot.
  2806. * A .device unit for an optical block device will now be
  2807. considered active only when a medium is in the drive. Also,
  2808. mount units are now bound to their backing devices thus
  2809. triggering automatic unmounting when devices become
  2810. unavailable. With this in place systemd will now
  2811. automatically unmount left-over mounts when a CD-ROM is
  2812. ejected or an USB stick is yanked from the system.
  2813. * networkd-wait-online now has support for waiting for
  2814. specific interfaces only (with globbing), and for giving up
  2815. after a configurable timeout.
  2816. * networkd now exits when idle. It will be automatically
  2817. restarted as soon as interfaces show up, are removed or
  2818. change state. networkd will stay around as long as there is
  2819. at least one DHCP state machine or similar around, that keep
  2820. it non-idle.
  2821. * networkd may now configure IPv6 link-local addressing in
  2822. addition to IPv4 link-local addressing.
  2823. * The IPv6 "token" for use in SLAAC may now be configured for
  2824. each .network interface in networkd.
  2825. * Routes configured with networkd may now be assigned a scope
  2826. in .network files.
  2827. * networkd's [Match] sections now support globbing and lists
  2828. of multiple space-separated matches per item.
  2829. Contributions from: Alban Crequy, Alin Rauta, Andrey Chaser,
  2830. Bastien Nocera, Bruno Bottazzini, Carlos Garnacho, Carlos
  2831. Morata Castillo, Chris Atkinson, Chris J. Arges, Christian
  2832. Kirbach, Christian Seiler, Christoph Brill, Colin Guthrie,
  2833. Colin Walters, Cristian Rodríguez, Daniele Medri, Daniel Mack,
  2834. Dave Reisner, David Herrmann, Djalal Harouni, Erik Auerswald,
  2835. Filipe Brandenburger, Frank Theile, Gabor Kelemen, Gabriel de
  2836. Perthuis, Harald Hoyer, Hui Wang, Ivan Shapovalov, Jan
  2837. Engelhardt, Jan Synacek, Jay Faulkner, Johannes Hölzl, Jonas
  2838. Ådahl, Jonathan Boulle, Josef Andersson, Kay Sievers, Ken
  2839. Werner, Lennart Poettering, Lucas De Marchi, Lukas Märdian,
  2840. Lukas Nykryn, Lukasz Skalski, Luke Shumaker, Mantas Mikulėnas,
  2841. Manuel Mendez, Marcel Holtmann, Marc Schmitzer, Marko
  2842. Myllynen, Martin Pitt, Maxim Mikityanskiy, Michael Biebl,
  2843. Michael Marineau, Michael Olbrich, Michal Schmidt, Mindaugas
  2844. Baranauskas, Moez Bouhlel, Naveen Kumar, Patrik Flykt, Paul
  2845. Martin, Peter Hutterer, Peter Mattern, Philippe De Swert,
  2846. Piotr Drąg, Rafael Ferreira, Rami Rosen, Robert Milasan, Ronny
  2847. Chevalier, Sangjung Woo, Sebastien Bacher, Sergey Ptashnick,
  2848. Shawn Landden, Stéphane Graber, Susant Sahani, Sylvain
  2849. Plantefève, Thomas Hindoe Paaboel Andersen, Tim JP, Tom
  2850. Gundersen, Topi Miettinen, Torstein Husebø, Umut Tezduyar
  2851. Lindskog, Veres Lajos, Vincent Batts, WaLyong Cho, Wieland
  2852. Hoffmann, Zbigniew Jędrzejewski-Szmek
  2853. — Berlin, 2015-02-16
  2854. CHANGES WITH 218:
  2855. * When querying unit file enablement status (for example via
  2856. "systemctl is-enabled"), a new state "indirect" is now known
  2857. which indicates that a unit might not be enabled itself, but
  2858. another unit listed in its Also= setting might be.
  2859. * Similar to the various existing ConditionXYZ= settings for
  2860. units, there are now matching AssertXYZ= settings. While
  2861. failing conditions cause a unit to be skipped, but its job
  2862. to succeed, failing assertions declared like this will cause
  2863. a unit start operation and its job to fail.
  2864. * hostnamed now knows a new chassis type "embedded".
  2865. * systemctl gained a new "edit" command. When used on a unit
  2866. file, this allows extending unit files with .d/ drop-in
  2867. configuration snippets or editing the full file (after
  2868. copying it from /usr/lib to /etc). This will invoke the
  2869. user's editor (as configured with $EDITOR), and reload the
  2870. modified configuration after editing.
  2871. * "systemctl status" now shows the suggested enablement state
  2872. for a unit, as declared in the (usually vendor-supplied)
  2873. system preset files.
  2874. * nss-myhostname will now resolve the single-label host name
  2875. "gateway" to the locally configured default IP routing
  2876. gateways, ordered by their metrics. This assigns a stable
  2877. name to the used gateways, regardless which ones are
  2878. currently configured. Note that the name will only be
  2879. resolved after all other name sources (if nss-myhostname is
  2880. configured properly) and should hence not negatively impact
  2881. systems that use the single-label host name "gateway" in
  2882. other contexts.
  2883. * systemd-inhibit now allows filtering by mode when listing
  2884. inhibitors.
  2885. * Scope and service units gained a new "Delegate" boolean
  2886. property, which, when set, allows processes running inside the
  2887. unit to further partition resources. This is primarily
  2888. useful for systemd user instances as well as container
  2889. managers.
  2890. * journald will now pick up audit messages directly from
  2891. the kernel, and log them like any other log message. The
  2892. audit fields are split up and fully indexed. This means that
  2893. journalctl in many ways is now a (nicer!) alternative to
  2894. ausearch, the traditional audit client. Note that this
  2895. implements only a minimal audit client. If you want the
  2896. special audit modes like reboot-on-log-overflow, please use
  2897. the traditional auditd instead, which can be used in
  2898. parallel to journald.
  2899. * The ConditionSecurity= unit file option now understands the
  2900. special string "audit" to check whether auditing is
  2901. available.
  2902. * journalctl gained two new commands --vacuum-size= and
  2903. --vacuum-time= to delete old journal files until the
  2904. remaining ones take up no more than the specified size on disk,
  2905. or are not older than the specified time.
  2906. * A new, native PPPoE library has been added to sd-network,
  2907. systemd's library of light-weight networking protocols. This
  2908. library will be used in a future version of networkd to
  2909. enable PPPoE communication without an external pppd daemon.
  2910. * The busctl tool now understands a new "capture" verb that
  2911. works similar to "monitor", but writes a packet capture
  2912. trace to STDOUT that can be redirected to a file which is
  2913. compatible with libcap's capture file format. This can then
  2914. be loaded in Wireshark and similar tools to inspect bus
  2915. communication.
  2916. * The busctl tool now understands a new "tree" verb that shows
  2917. the object trees of a specific service on the bus, or of all
  2918. services.
  2919. * The busctl tool now understands a new "introspect" verb that
  2920. shows all interfaces and members of objects on the bus,
  2921. including their signature and values. This is particularly
  2922. useful to get more information about bus objects shown by
  2923. the new "busctl tree" command.
  2924. * The busctl tool now understands new verbs "call",
  2925. "set-property" and "get-property" for invoking bus method
  2926. calls, setting and getting bus object properties in a
  2927. friendly way.
  2928. * busctl gained a new --augment-creds= argument that controls
  2929. whether the tool shall augment credential information it
  2930. gets from the bus with data from /proc, in a possibly
  2931. race-ful way.
  2932. * nspawn's --link-journal= switch gained two new values
  2933. "try-guest" and "try-host" that work like "guest" and
  2934. "host", but do not fail if the host has no persistent
  2935. journalling enabled. -j is now equivalent to
  2936. --link-journal=try-guest.
  2937. * macvlan network devices created by nspawn will now have
  2938. stable MAC addresses.
  2939. * A new SmackProcessLabel= unit setting has been added, which
  2940. controls the SMACK security label processes forked off by
  2941. the respective unit shall use.
  2942. * If compiled with --enable-xkbcommon, systemd-localed will
  2943. verify x11 keymap settings by compiling the given keymap. It
  2944. will spew out warnings if the compilation fails. This
  2945. requires libxkbcommon to be installed.
  2946. * When a coredump is collected, a larger number of metadata
  2947. fields is now collected and included in the journal records
  2948. created for it. More specifically, control group membership,
  2949. environment variables, memory maps, working directory,
  2950. chroot directory, /proc/$PID/status, and a list of open file
  2951. descriptors is now stored in the log entry.
  2952. * The udev hwdb now contains DPI information for mice. For
  2953. details see:
  2955. * All systemd programs that read standalone configuration
  2956. files in /etc now also support a corresponding series of
  2957. .conf.d configuration directories in /etc/, /run/,
  2958. /usr/local/lib/, /usr/lib/, and (if configured with
  2959. --enable-split-usr) /lib/. In particular, the following
  2960. configuration files now have corresponding configuration
  2961. directories: system.conf user.conf, logind.conf,
  2962. journald.conf, sleep.conf, bootchart.conf, coredump.conf,
  2963. resolved.conf, timesyncd.conf, journal-remote.conf, and
  2964. journal-upload.conf. Note that distributions should use the
  2965. configuration directories in /usr/lib/; the directories in
  2966. /etc/ are reserved for the system administrator.
  2967. * systemd-rfkill will no longer take the rfkill device name
  2968. into account when storing rfkill state on disk, as the name
  2969. might be dynamically assigned and not stable. Instead, the
  2970. ID_PATH udev variable combined with the rfkill type (wlan,
  2971. bluetooth, ...) is used.
  2972. * A new service systemd-machine-id-commit.service has been
  2973. added. When used on systems where /etc is read-only during
  2974. boot, and /etc/machine-id is not initialized (but an empty
  2975. file), this service will copy the temporary machine ID
  2976. created as replacement into /etc after the system is fully
  2977. booted up. This is useful for systems that are freshly
  2978. installed with a non-initialized machine ID, but should get
  2979. a fixed machine ID for subsequent boots.
  2980. * networkd's .netdev files now provide a large set of
  2981. configuration parameters for VXLAN devices. Similarly, the
  2982. bridge port cost parameter is now configurable in .network
  2983. files. There's also new support for configuring IP source
  2984. routing. networkd .link files gained support for a new
  2985. OriginalName= match that is useful to match against the
  2986. original interface name the kernel assigned. .network files
  2987. may include MTU= and MACAddress= fields for altering the MTU
  2988. and MAC address while being connected to a specific network
  2989. interface.
  2990. * The LUKS logic gained supported for configuring
  2991. UUID-specific key files. There's also new support for naming
  2992. LUKS device from the kernel command line, using the new
  2993. argument.
  2994. * Timer units may now be transiently created via the bus API
  2995. (this was previously already available for scope and service
  2996. units). In addition it is now possible to create multiple
  2997. transient units at the same time with a single bus call. The
  2998. "systemd-run" tool has been updated to make use of this for
  2999. running commands on a specified time, in at(1)-style.
  3000. * tmpfiles gained support for "t" lines, for assigning
  3001. extended attributes to files. Among other uses this may be
  3002. used to assign SMACK labels to files.
  3003. Contributions from: Alin Rauta, Alison Chaiken, Andrej
  3004. Manduch, Bastien Nocera, Chris Atkinson, Chris Leech, Chris
  3005. Mayo, Colin Guthrie, Colin Walters, Cristian Rodríguez,
  3006. Daniele Medri, Daniel Mack, Dan Williams, Dan Winship, Dave
  3007. Reisner, David Herrmann, Didier Roche, Felipe Sateler, Gavin
  3008. Li, Hans de Goede, Harald Hoyer, Iago López Galeiras, Ivan
  3009. Shapovalov, Jakub Filak, Jan Janssen, Jan Synacek, Joe
  3010. Lawrence, Josh Triplett, Kay Sievers, Lennart Poettering,
  3011. Lukas Nykryn, Łukasz Stelmach, Maciej Wereski, Mantas
  3012. Mikulėnas, Marcel Holtmann, Martin Pitt, Maurizio Lombardi,
  3013. Michael Biebl, Michael Chapman, Michael Marineau, Michal
  3014. Schmidt, Michal Sekletar, Olivier Brunel, Patrik Flykt, Peter
  3015. Hutterer, Przemyslaw Kedzierski, Rami Rosen, Ray Strode,
  3016. Richard Schütz, Richard W.M. Jones, Ronny Chevalier, Ross
  3017. Lagerwall, Sean Young, Stanisław Pitucha, Susant Sahani,
  3018. Thomas Haller, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
  3019. Torstein Husebø, Umut Tezduyar Lindskog, Vicente Olivert
  3020. Riera, WaLyong Cho, Wesley Dawson, Zbigniew Jędrzejewski-Szmek
  3021. — Berlin, 2014-12-10
  3022. CHANGES WITH 217:
  3023. * journalctl gained the new options -t/--identifier= to match
  3024. on the syslog identifier (aka "tag"), as well as --utc to
  3025. show log timestamps in the UTC timezone. journalctl now also
  3026. accepts -n/--lines=all to disable line capping in a pager.
  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.
  3031. * Services can notify the manager before they start a reload
  3032. (by sending RELOADING=1) or shutdown (by sending
  3033. STOPPING=1). This allows the manager to track and show the
  3034. internal state of daemons and closes a race condition when
  3035. the process is still running but has closed its D-Bus
  3036. connection.
  3037. * Services with Type=oneshot do not have to have any ExecStart
  3038. commands anymore.
  3039. * User units are now loaded also from
  3040. $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
  3041. /run/systemd/user directory that was already previously
  3042. supported, but is under the control of the user.
  3043. * Job timeouts (i.e. time-outs on the time a job that is
  3044. queued stays in the run queue) can now optionally result in
  3045. immediate reboot or power-off actions (JobTimeoutAction= and
  3046. JobTimeoutRebootArgument=). This is useful on ".target"
  3047. units, to limit the maximum time a target remains
  3048. undispatched in the run queue, and to trigger an emergency
  3049. operation in such a case. This is now used by default to
  3050. turn off the system if boot-up (as defined by everything in
  3051. hangs and does not complete for at least
  3052. 15min. Also, if power-off or reboot hang for at least 30min
  3053. an immediate power-off/reboot operation is triggered. This
  3054. functionality is particularly useful to increase reliability
  3055. on embedded devices, but also on laptops which might
  3056. accidentally get powered on when carried in a backpack and
  3057. whose boot stays stuck in a hard disk encryption passphrase
  3058. question.
  3059. * systemd-logind can be configured to also handle lid switch
  3060. events even when the machine is docked or multiple displays
  3061. are attached (HandleLidSwitchDocked= option).
  3062. * A helper binary and a service have been added which can be
  3063. used to resume from hibernation in the initramfs. A
  3064. generator will parse the resume= option on the kernel
  3065. command line to trigger resume.
  3066. * A user console daemon systemd-consoled has been
  3067. added. Currently, it is a preview, and will so far open a
  3068. single terminal on each session of the user marked as
  3069. Desktop=systemd-console.
  3070. * Route metrics can be specified for DHCP routes added by
  3071. systemd-networkd.
  3072. * The SELinux context of socket-activated services can be set
  3073. from the information provided by the networking stack
  3074. (SELinuxContextFromNet= option).
  3075. * Userspace firmware loading support has been removed and
  3076. the minimum supported kernel version is thus bumped to 3.7.
  3077. * Timeout for udev workers has been increased from 1 to 3
  3078. minutes, but a warning will be printed after 1 minute to
  3079. help diagnose kernel modules that take a long time to load.
  3080. * Udev rules can now remove tags on devices with TAG-="foobar".
  3081. * systemd's readahead implementation has been removed. In many
  3082. circumstances it didn't give expected benefits even for
  3083. rotational disk drives and was becoming less relevant in the
  3084. age of SSDs. As none of the developers has been using
  3085. rotating media anymore, and nobody stepped up to actively
  3086. maintain this component of systemd it has now been removed.
  3087. * Swap units can use Options= to specify discard options.
  3088. Discard options specified for swaps in /etc/fstab are now
  3089. respected.
  3090. * Docker containers are now detected as a separate type of
  3091. virtualization.
  3092. * The Password Agent protocol gained support for queries where
  3093. the user input is shown, useful e.g. for user names.
  3094. systemd-ask-password gained a new --echo option to turn that
  3095. on.
  3096. * The default sysctl.d/ snippets will now set:
  3097. net.core.default_qdisc = fq_codel
  3098. This selects Fair Queuing Controlled Delay as the default
  3099. queuing discipline for network interfaces. fq_codel helps
  3100. fight the network bufferbloat problem. It is believed to be
  3101. a good default with no tuning required for most workloads.
  3102. Downstream distributions may override this choice. On 10Gbit
  3103. servers that do not do forwarding, "fq" may perform better.
  3104. Systems without a good clocksource should use "pfifo_fast".
  3105. * If kdbus is enabled during build a new option BusPolicy= is
  3106. available for service units, that allows locking all service
  3107. processes into a stricter bus policy, in order to limit
  3108. access to various bus services, or even hide most of them
  3109. from the service's view entirely.
  3110. * networkctl will now show the .network and .link file
  3111. networkd has applied to a specific interface.
  3112. * sd-login gained a new API call sd_session_get_desktop() to
  3113. query which desktop environment has been selected for a
  3114. session.
  3115. * UNIX utmp support is now compile-time optional to support
  3116. legacy-free systems.
  3117. * systemctl gained two new commands "add-wants" and
  3118. "add-requires" for pulling in units from specific targets
  3119. easily.
  3120. * If the word "rescue" is specified on the kernel command line
  3121. the system will now boot into rescue mode (aka
  3122., which was previously available only by
  3123. specifying "1" or "" on the kernel
  3124. command line. This new kernel command line option nicely
  3125. mirrors the already existing "emergency" kernel command line
  3126. option.
  3127. * New kernel command line options mount.usr=, mount.usrflags=,
  3128. mount.usrfstype= have been added that match root=, rootflags=,
  3129. rootfstype= but allow mounting a specific file system to
  3130. /usr.
  3131. * The $NOTIFY_SOCKET is now also passed to control processes of
  3132. services, not only the main process.
  3133. * This version reenables support for fsck's -l switch. This
  3134. means at least version v2.25 of util-linux is required for
  3135. operation, otherwise dead-locks on device nodes may
  3136. occur. Again: you need to update util-linux to at least
  3137. v2.25 when updating systemd to v217.
  3138. * The "multi-seat-x" tool has been removed from systemd, as
  3139. its functionality has been integrated into X servers 1.16,
  3140. and the tool is hence redundant. It is recommended to update
  3141. display managers invoking this tool to simply invoke X
  3142. directly from now on, again.
  3143. * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
  3144. message flag has been added for all of systemd's PolicyKit
  3145. authenticated method calls has been added. In particular
  3146. this now allows optional interactive authorization via
  3147. PolicyKit for many of PID1's privileged operations such as
  3148. unit file enabling and disabling.
  3149. * "udevadm hwdb --update" learnt a new switch "--usr" for
  3150. placing the rebuilt hardware database in /usr instead of
  3151. /etc. When used only hardware database entries stored in
  3152. /usr will be used, and any user database entries in /etc are
  3153. ignored. This functionality is useful for vendors to ship a
  3154. pre-built database on systems where local configuration is
  3155. unnecessary or unlikely.
  3156. * Calendar time specifications in .timer units now also
  3157. understand the strings "semi-annually", "quarterly" and
  3158. "minutely" as shortcuts (in addition to the preexisting
  3159. "anually", "hourly", ...).
  3160. * systemd-tmpfiles will now correctly create files in /dev
  3161. at boot which are marked for creation only at boot. It is
  3162. recommended to always create static device nodes with 'c!'
  3163. and 'b!', so that they are created only at boot and not
  3164. overwritten at runtime.
  3165. * When the watchdog logic is used for a service (WatchdogSec=)
  3166. and the watchdog timeout is hit the service will now be
  3167. terminated with SIGABRT (instead of just SIGTERM), in order
  3168. to make sure a proper coredump and backtrace is
  3169. generated. This ensures that hanging services will result in
  3170. similar coredump/backtrace behaviour as services that hit a
  3171. segmentation fault.
  3172. Contributions from: Andreas Henriksson, Andrei Borzenkov,
  3173. Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
  3174. Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
  3175. Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
  3176. Herrmann, David Sommerseth, David Strauss, Emil Renner
  3177. Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
  3178. Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
  3179. Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
  3180. Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
  3181. Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
  3182. Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
  3183. Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
  3184. Michael Marineau, Michael Olbrich, Michael Scherer, Michal
  3185. Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
  3186. Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
  3187. Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
  3188. Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
  3189. Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
  3190. Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
  3191. Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
  3192. Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
  3193. Jędrzejewski-Szmek
  3194. — Berlin, 2014-10-28
  3195. CHANGES WITH 216:
  3196. * timedated no longer reads NTP implementation unit names from
  3197. /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
  3198. implementations should add a
  3199. Conflicts=systemd-timesyncd.service
  3200. to their unit files to take over and replace systemd's NTP
  3201. default functionality.
  3202. * systemd-sysusers gained a new line type "r" for configuring
  3203. which UID/GID ranges to allocate system users/groups
  3204. from. Lines of type "u" may now add an additional column
  3205. that specifies the home directory for the system user to be
  3206. created. Also, systemd-sysusers may now optionally read user
  3207. information from STDIN instead of a file. This is useful for
  3208. invoking it from RPM preinst scriptlets that need to create
  3209. users before the first RPM file is installed since these
  3210. files might need to be owned by them. A new
  3211. %sysusers_create_inline RPM macro has been introduced to do
  3212. just that. systemd-sysusers now updates the shadow files as
  3213. well as the user/group databases, which should enhance
  3214. compatibility with certain tools like grpck.
  3215. * A number of bus APIs of PID 1 now optionally consult
  3216. PolicyKit to permit access for otherwise unprivileged
  3217. clients under certain conditions. Note that this currently
  3218. doesn't support interactive authentication yet, but this is
  3219. expected to be added eventually, too.
  3220. * /etc/machine-info now has new fields for configuring the
  3221. deployment environment of the machine, as well as the
  3222. location of the machine. hostnamectl has been updated with
  3223. new command to update these fields.
  3224. * systemd-timesyncd has been updated to automatically acquire
  3225. NTP server information from systemd-networkd, which might
  3226. have been discovered via DHCP.
  3227. * systemd-resolved now includes a caching DNS stub resolver
  3228. and a complete LLMNR name resolution implementation. A new
  3229. NSS module "nss-resolve" has been added which can be used
  3230. instead of glibc's own "nss-dns" to resolve hostnames via
  3231. systemd-resolved. Hostnames, addresses and arbitrary RRs may
  3232. be resolved via systemd-resolved D-Bus APIs. In contrast to
  3233. the glibc internal resolver systemd-resolved is aware of
  3234. multi-homed system, and keeps DNS server and caches separate
  3235. and per-interface. Queries are sent simultaneously on all
  3236. interfaces that have DNS servers configured, in order to
  3237. properly handle VPNs and local LANs which might resolve
  3238. separate sets of domain names. systemd-resolved may acquire
  3239. DNS server information from systemd-networkd automatically,
  3240. which in turn might have discovered them via DHCP. A tool
  3241. "systemd-resolve-host" has been added that may be used to
  3242. query the DNS logic in resolved. systemd-resolved implements
  3243. IDNA and automatically uses IDNA or UTF-8 encoding depending
  3244. on whether classic DNS or LLMNR is used as transport. In the
  3245. next releases we intend to add a DNSSEC and mDNS/DNS-SD
  3246. implementation to systemd-resolved.
  3247. * A new NSS module nss-mymachines has been added, that
  3248. automatically resolves the names of all local registered
  3249. containers to their respective IP addresses.
  3250. * A new client tool "networkctl" for systemd-networkd has been
  3251. added. It currently is entirely passive and will query
  3252. networking configuration from udev, rtnetlink and networkd,
  3253. and present it to the user in a very friendly
  3254. way. Eventually, we hope to extend it to become a full
  3255. control utility for networkd.
  3256. * .socket units gained a new DeferAcceptSec= setting that
  3257. controls the kernels' TCP_DEFER_ACCEPT sockopt for
  3258. TCP. Similarly, support for controlling TCP keep-alive
  3259. settings has been added (KeepAliveTimeSec=,
  3260. KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
  3261. turning off Nagle's algorithm on TCP has been added
  3262. (NoDelay=).
  3263. * logind learned a new session type "web", for use in projects
  3264. like Cockpit which register web clients as PAM sessions.
  3265. * timer units with at least one OnCalendar= setting will now
  3266. be started only after has been
  3267. reached. This way they will not elapse before the system
  3268. clock has been corrected by a local NTP client or
  3269. similar. This is particular useful on RTC-less embedded
  3270. machines, that come up with an invalid system clock.
  3271. * systemd-nspawn's --network-veth= switch should now result in
  3272. stable MAC addresses for both the outer and the inner side
  3273. of the link.
  3274. * systemd-nspawn gained a new --volatile= switch for running
  3275. container instances with /etc or /var unpopulated.
  3276. * The kdbus client code has been updated to use the new Linux
  3277. 3.17 memfd subsystem instead of the old kdbus-specific one.
  3278. * systemd-networkd's DHCP client and server now support
  3279. FORCERENEW. There are also new configuration options to
  3280. configure the vendor client identifier and broadcast mode
  3281. for DHCP.
  3282. * systemd will no longer inform the kernel about the current
  3283. timezone, as this is necessarily incorrect and racy as the
  3284. kernel has no understanding of DST and similar
  3285. concepts. This hence means FAT timestamps will be always
  3286. considered UTC, similar to what Android is already
  3287. doing. Also, when the RTC is configured to the local time
  3288. (rather than UTC) systemd will never synchronize back to it,
  3289. as this might confuse Windows at a later boot.
  3290. * systemd-analyze gained a new command "verify" for offline
  3291. validation of unit files.
  3292. * systemd-networkd gained support for a couple of additional
  3293. settings for bonding networking setups. Also, the metric for
  3294. statically configured routes may now be configured. For
  3295. network interfaces where this is appropriate the peer IP
  3296. address may now be configured.
  3297. * systemd-networkd's DHCP client will no longer request
  3298. broadcasting by default, as this tripped up some networks.
  3299. For hardware where broadcast is required the feature should
  3300. be switched back on using RequestBroadcast=yes.
  3301. * systemd-networkd will now set up IPv4LL addresses (when
  3302. enabled) even if DHCP is configured successfully.
  3303. * udev will now default to respect network device names given
  3304. by the kernel when the kernel indicates that these are
  3305. predictable. This behavior can be tweaked by changing
  3306. NamePolicy= in the relevant .link file.
  3307. * A new library systemd-terminal has been added that
  3308. implements full TTY stream parsing and rendering. This
  3309. library is supposed to be used later on for implementing a
  3310. full userspace VT subsystem, replacing the current kernel
  3311. implementation.
  3312. * A new tool systemd-journal-upload has been added to push
  3313. journal data to a remote system running
  3314. systemd-journal-remote.
  3315. * journald will no longer forward all local data to another
  3316. running syslog daemon. This change has been made because
  3317. rsyslog (which appears to be the most commonly used syslog
  3318. implementation these days) no longer makes use of this, and
  3319. instead pulls the data out of the journal on its own. Since
  3320. forwarding the messages to a non-existent syslog server is
  3321. more expensive than we assumed we have now turned this
  3322. off. If you run a syslog server that is not a recent rsyslog
  3323. version, you have to turn this option on again
  3324. (ForwardToSyslog= in journald.conf).
  3325. * journald now optionally supports the LZ4 compressor for
  3326. larger journal fields. This compressor should perform much
  3327. better than XZ which was the previous default.
  3328. * machinectl now shows the IP addresses of local containers,
  3329. if it knows them, plus the interface name of the container.
  3330. * A new tool "systemd-escape" has been added that makes it
  3331. easy to escape strings to build unit names and similar.
  3332. * sd_notify() messages may now include a new ERRNO= field
  3333. which is parsed and collected by systemd and shown among the
  3334. "systemctl status" output for a service.
  3335. * A new component "systemd-firstboot" has been added that
  3336. queries the most basic systemd information (timezone,
  3337. hostname, root password) interactively on first
  3338. boot. Alternatively it may also be used to provision these
  3339. things offline on OS images installed into directories.
  3340. * The default sysctl.d/ snippets will now set
  3341. net.ipv4.conf.default.promote_secondaries=1
  3342. This has the benefit of no flushing secondary IP addresses
  3343. when primary addresses are removed.
  3344. Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
  3345. Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
  3346. Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
  3347. Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
  3348. Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
  3349. B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
  3350. Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
  3351. Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
  3352. Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
  3353. Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
  3354. Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
  3355. Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
  3356. Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
  3357. Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
  3358. Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
  3359. — Berlin, 2014-08-19
  3360. CHANGES WITH 215:
  3361. * A new tool systemd-sysusers has been added. This tool
  3362. creates system users and groups in /etc/passwd and
  3363. /etc/group, based on static declarative system user/group
  3364. definitions in /usr/lib/sysusers.d/. This is useful to
  3365. enable factory resets and volatile systems that boot up with
  3366. an empty /etc directory, and thus need system users and
  3367. groups created during early boot. systemd now also ships
  3368. with two default sysusers.d/ files for the most basic
  3369. users and groups systemd and the core operating system
  3370. require.
  3371. * A new tmpfiles snippet has been added that rebuilds the
  3372. essential files in /etc on boot, should they be missing.
  3373. * A directive for ensuring automatic clean-up of
  3374. /var/cache/man/ has been removed from the default
  3375. configuration. This line should now be shipped by the man
  3376. implementation. The necessary change has been made to the
  3377. man-db implementation. Note that you need to update your man
  3378. implementation to one that ships this line, otherwise no
  3379. automatic clean-up of /var/cache/man will take place.
  3380. * A new condition ConditionNeedsUpdate= has been added that
  3381. may conditionalize services to only run when /etc or /var
  3382. are "older" than the vendor operating system resources in
  3383. /usr. This is useful for reconstructing or updating /etc
  3384. after an offline update of /usr or a factory reset, on the
  3385. next reboot. Services that want to run once after such an
  3386. update or reset should use this condition and order
  3387. themselves before the new systemd-update-done.service, which
  3388. will mark the two directories as fully updated. A number of
  3389. service files have been added making use of this, to rebuild
  3390. the udev hardware database, the journald message catalog and
  3391. dynamic loader cache (ldconfig). The systemd-sysusers tool
  3392. described above also makes use of this now. With this in
  3393. place it is now possible to start up a minimal operating
  3394. system with /etc empty cleanly. For more information on the
  3395. concepts involved see this recent blog story:
  3397. * A new system group "input" has been introduced, and all
  3398. input device nodes get this group assigned. This is useful
  3399. for system-level software to get access to input devices. It
  3400. complements what is already done for "audio" and "video".
  3401. * systemd-networkd learnt minimal DHCPv4 server support in
  3402. addition to the existing DHCPv4 client support. It also
  3403. learnt DHCPv6 client and IPv6 Router Solicitation client
  3404. support. The DHCPv4 client gained support for static routes
  3405. passed in from the server. Note that the [DHCPv4] section
  3406. known in older systemd-networkd versions has been renamed to
  3407. [DHCP] and is now also used by the DHCPv6 client. Existing
  3408. .network files using settings of this section should be
  3409. updated, though compatibility is maintained. Optionally, the
  3410. client hostname may now be sent to the DHCP server.
  3411. * networkd gained support for vxlan virtual networks as well
  3412. as tun/tap and dummy devices.
  3413. * networkd gained support for automatic allocation of address
  3414. ranges for interfaces from a system-wide pool of
  3415. addresses. This is useful for dynamically managing a large
  3416. number of interfaces with a single network configuration
  3417. file. In particular this is useful to easily assign
  3418. appropriate IP addresses to the veth links of a large number
  3419. of nspawn instances.
  3420. * RPM macros for processing sysusers, sysctl and binfmt
  3421. drop-in snippets at package installation time have been
  3422. added.
  3423. * The /etc/os-release file should now be placed in
  3424. /usr/lib/os-release. The old location is automatically
  3425. created as symlink. /usr/lib is the more appropriate
  3426. location of this file, since it shall actually describe the
  3427. vendor operating system shipped in /usr, and not the
  3428. configuration stored in /etc.
  3429. * .mount units gained a new boolean SloppyOptions= setting
  3430. that maps to mount(8)'s -s option which enables permissive
  3431. parsing of unknown mount options.
  3432. * tmpfiles learnt a new "L+" directive which creates a symlink
  3433. but (unlike "L") deletes a pre-existing file first, should
  3434. it already exist and not already be the correct
  3435. symlink. Similarly, "b+", "c+" and "p+" directives have been
  3436. added as well, which create block and character devices, as
  3437. well as fifos in the filesystem, possibly removing any
  3438. pre-existing files of different types.
  3439. * For tmpfiles' "L", "L+", "C" and "C+" directives the final
  3440. 'argument' field (which so far specified the source to
  3441. symlink/copy the files from) is now optional. If omitted the
  3442. same file os copied from /usr/share/factory/ suffixed by the
  3443. full destination path. This is useful for populating /etc
  3444. with essential files, by copying them from vendor defaults
  3445. shipped in /usr/share/factory/etc.
  3446. * A new command "systemctl preset-all" has been added that
  3447. applies the service preset settings to all installed unit
  3448. files. A new switch --preset-mode= has been added that
  3449. controls whether only enable or only disable operations
  3450. shall be executed.
  3451. * A new command "systemctl is-system-running" has been added
  3452. that allows checking the overall state of the system, for
  3453. example whether it is fully up and running.
  3454. * When the system boots up with an empty /etc, the equivalent
  3455. to "systemctl preset-all" is executed during early boot, to
  3456. make sure all default services are enabled after a factory
  3457. reset.
  3458. * systemd now contains a minimal preset file that enables the
  3459. most basic services systemd ships by default.
  3460. * Unit files' [Install] section gained a new DefaultInstance=
  3461. field for defining the default instance to create if a
  3462. template unit is enabled with no instance specified.
  3463. * A new passive target has been added
  3464. that may be used by services that need to make they run and
  3465. finish before the first LUKS cryptographic device is set up.
  3466. * The /dev/loop-control and /dev/btrfs-control device nodes
  3467. are now owned by the "disk" group by default, opening up
  3468. access to this group.
  3469. * systemd-coredump will now automatically generate a
  3470. stack trace of all core dumps taking place on the system,
  3471. based on elfutils' libdw library. This stack trace is logged
  3472. to the journal.
  3473. * systemd-coredump may now optionally store coredumps directly
  3474. on disk (in /var/lib/systemd/coredump, possibly compressed),
  3475. instead of storing them unconditionally in the journal. This
  3476. mode is the new default. A new configuration file
  3477. /etc/systemd/coredump.conf has been added to configure this
  3478. and other parameters of systemd-coredump.
  3479. * coredumpctl gained a new "info" verb to show details about a
  3480. specific coredump. A new switch "-1" has also been added
  3481. that makes sure to only show information about the most
  3482. recent entry instead of all entries. Also, as the tool is
  3483. generally useful now the "systemd-" prefix of the binary
  3484. name has been removed. Distributions that want to maintain
  3485. compatibility with the old name should add a symlink from
  3486. the old name to the new name.
  3487. * journald's SplitMode= now defaults to "uid". This makes sure
  3488. that unprivileged users can access their own coredumps with
  3489. coredumpctl without restrictions.
  3490. * New kernel command line options "systemd.wants=" (for
  3491. pulling an additional unit during boot), "systemd.mask="
  3492. (for masking a specific unit for the boot), and
  3493. "systemd.debug-shell" (for enabling the debug shell on tty9)
  3494. have been added. This is implemented in the new generator
  3495. "systemd-debug-generator".
  3496. * systemd-nspawn will now by default filter a couple of
  3497. syscalls for containers, among them those required for
  3498. kernel module loading, direct x86 IO port access, swap
  3499. management, and kexec. Most importantly though
  3500. open_by_handle_at() is now prohibited for containers,
  3501. closing a hole similar to a recently discussed vulnerability
  3502. in docker regarding access to files on file hierarchies the
  3503. container should normally not have access to. Note that, for
  3504. nspawn, we generally make no security claims anyway (and
  3505. this is explicitly documented in the man page), so this is
  3506. just a fix for one of the most obvious problems.
  3507. * A new man page file-hierarchy(7) has been added that
  3508. contains a minimized, modernized version of the file system
  3509. layout systemd expects, similar in style to the FHS
  3510. specification or hier(5). A new tool systemd-path(1) has
  3511. been added to query many of these paths for the local
  3512. machine and user.
  3513. * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
  3514. longer done. Since the directory now has a per-user size
  3515. limit, and is cleaned on logout this appears unnecessary,
  3516. in particular since this now brings the lifecycle of this
  3517. directory closer in line with how IPC objects are handled.
  3518. * systemd.pc now exports a number of additional directories,
  3519. including $libdir (which is useful to identify the library
  3520. path for the primary architecture of the system), and a
  3521. couple of drop-in directories.
  3522. * udev's predictable network interface names now use the dev_port
  3523. sysfs attribute, introduced in linux 3.15 instead of dev_id to
  3524. distinguish between ports of the same PCI function. dev_id should
  3525. only be used for ports using the same HW address, hence the need
  3526. for dev_port.
  3527. * machined has been updated to export the OS version of a
  3528. container (read from /etc/os-release and
  3529. /usr/lib/os-release) on the bus. This is now shown in
  3530. "machinectl status" for a machine.
  3531. * A new service setting RestartForceExitStatus= has been
  3532. added. If configured to a set of exit signals or process
  3533. return values, the service will be restarted when the main
  3534. daemon process exits with any of them, regardless of the
  3535. Restart= setting.
  3536. * systemctl's -H switch for connecting to remote systemd
  3537. machines has been extended so that it may be used to
  3538. directly connect to a specific container on the
  3539. host. "systemctl -H root@foobar:waldi" will now connect as
  3540. user "root" to host "foobar", and then proceed directly to
  3541. the container named "waldi". Note that currently you have to
  3542. authenticate as user "root" for this to work, as entering
  3543. containers is a privileged operation.
  3544. Contributions from: Andreas Henriksson, Benjamin Steinwender,
  3545. Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
  3546. Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
  3547. Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
  3548. Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
  3549. Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
  3550. Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
  3551. Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
  3552. Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
  3553. Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
  3554. Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
  3555. Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
  3556. — Berlin, 2014-07-03
  3557. CHANGES WITH 214:
  3558. * As an experimental feature, udev now tries to lock the
  3559. disk device node (flock(LOCK_SH|LOCK_NB)) while it
  3560. executes events for the disk or any of its partitions.
  3561. Applications like partitioning programs can lock the
  3562. disk device node (flock(LOCK_EX)) and claim temporary
  3563. device ownership that way; udev will entirely skip all event
  3564. handling for this disk and its partitions. If the disk
  3565. was opened for writing, the close will trigger a partition
  3566. table rescan in udev's "watch" facility, and if needed
  3567. synthesize "change" events for the disk and all its partitions.
  3568. This is now unconditionally enabled, and if it turns out to
  3569. cause major problems, we might turn it on only for specific
  3570. devices, or might need to disable it entirely. Device Mapper
  3571. devices are excluded from this logic.
  3572. * We temporarily dropped the "-l" switch for fsck invocations,
  3573. since they collide with the flock() logic above. util-linux
  3574. upstream has been changed already to avoid this conflict,
  3575. and we will readd "-l" as soon as util-linux with this
  3576. change has been released.
  3577. * The dependency on libattr has been removed. Since a long
  3578. time, the extended attribute calls have moved to glibc, and
  3579. libattr is thus unnecessary.
  3580. * Virtualization detection works without privileges now. This
  3581. means the systemd-detect-virt binary no longer requires
  3582. CAP_SYS_PTRACE file capabilities, and our daemons can run
  3583. with fewer privileges.
  3584. * systemd-networkd now runs under its own "systemd-network"
  3585. user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
  3586. CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
  3587. loses the ability to write to files owned by root this way.
  3588. * Similarly, systemd-resolved now runs under its own
  3589. "systemd-resolve" user with no capabilities remaining.
  3590. * Similarly, systemd-bus-proxyd now runs under its own
  3591. "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
  3592. * systemd-networkd gained support for setting up "veth"
  3593. virtual Ethernet devices for container connectivity, as well
  3594. as GRE and VTI tunnels.
  3595. * systemd-networkd will no longer automatically attempt to
  3596. manually load kernel modules necessary for certain tunnel
  3597. transports. Instead, it is assumed the kernel loads them
  3598. automatically when required. This only works correctly on
  3599. very new kernels. On older kernels, please consider adding
  3600. the kernel modules to /etc/modules-load.d/ as a work-around.
  3601. * The resolv.conf file systemd-resolved generates has been
  3602. moved to /run/systemd/resolve/. If you have a symlink from
  3603. /etc/resolv.conf, it might be necessary to correct it.
  3604. * Two new service settings, ProtectHome= and ProtectSystem=,
  3605. have been added. When enabled, they will make the user data
  3606. (such as /home) inaccessible or read-only and the system
  3607. (such as /usr) read-only, for specific services. This allows
  3608. very light-weight per-service sandboxing to avoid
  3609. modifications of user data or system files from
  3610. services. These two new switches have been enabled for all
  3611. of systemd's long-running services, where appropriate.
  3612. * Socket units gained new SocketUser= and SocketGroup=
  3613. settings to set the owner user and group of AF_UNIX sockets
  3614. and FIFOs in the file system.
  3615. * Socket units gained a new RemoveOnStop= setting. If enabled,
  3616. all FIFOS and sockets in the file system will be removed
  3617. when the specific socket unit is stopped.
  3618. * Socket units gained a new Symlinks= setting. It takes a list
  3619. of symlinks to create to file system sockets or FIFOs
  3620. created by the specific Unix sockets. This is useful to
  3621. manage symlinks to socket nodes with the same life-cycle as
  3622. the socket itself.
  3623. * The /dev/log socket and /dev/initctl FIFO have been moved to
  3624. /run, and have been replaced by symlinks. This allows
  3625. connecting to these facilities even if PrivateDevices=yes is
  3626. used for a service (which makes /dev/log itself unavailable,
  3627. but /run is left). This also has the benefit of ensuring
  3628. that /dev only contains device nodes, directories and
  3629. symlinks, and nothing else.
  3630. * sd-daemon gained two new calls sd_pid_notify() and
  3631. sd_pid_notifyf(). They are similar to sd_notify() and
  3632. sd_notifyf(), but allow overriding of the source PID of
  3633. notification messages if permissions permit this. This is
  3634. useful to send notify messages on behalf of a different
  3635. process (for example, the parent process). The
  3636. systemd-notify tool has been updated to make use of this
  3637. when sending messages (so that notification messages now
  3638. originate from the shell script invoking systemd-notify and
  3639. not the systemd-notify process itself. This should minimize
  3640. a race where systemd fails to associate notification
  3641. messages to services when the originating process already
  3642. vanished.
  3643. * A new "on-abnormal" setting for Restart= has been added. If
  3644. set, it will result in automatic restarts on all "abnormal"
  3645. reasons for a process to exit, which includes unclean
  3646. signals, core dumps, timeouts and watchdog timeouts, but
  3647. does not include clean and unclean exit codes or clean
  3648. signals. Restart=on-abnormal is an alternative for
  3649. Restart=on-failure for services that shall be able to
  3650. terminate and avoid restarts on certain errors, by
  3651. indicating so with an unclean exit code. Restart=on-failure
  3652. or Restart=on-abnormal is now the recommended setting for
  3653. all long-running services.
  3654. * If the InaccessibleDirectories= service setting points to a
  3655. mount point (or if there are any submounts contained within
  3656. it), it is now attempted to completely unmount it, to make
  3657. the file systems truly unavailable for the respective
  3658. service.
  3659. * The ReadOnlyDirectories= service setting and
  3660. systemd-nspawn's --read-only parameter are now recursively
  3661. applied to all submounts, too.
  3662. * Mount units may now be created transiently via the bus APIs.
  3663. * The support for SysV and LSB init scripts has been removed
  3664. from the systemd daemon itself. Instead, it is now
  3665. implemented as a generator that creates native systemd units
  3666. from these scripts when needed. This enables us to remove a
  3667. substantial amount of legacy code from PID 1, following the
  3668. fact that many distributions only ship a very small number
  3669. of LSB/SysV init scripts nowadays.
  3670. * Privileged Xen (dom0) domains are not considered
  3671. virtualization anymore by the virtualization detection
  3672. logic. After all, they generally have unrestricted access to
  3673. the hardware and usually are used to manage the unprivileged
  3674. (domU) domains.
  3675. * systemd-tmpfiles gained a new "C" line type, for copying
  3676. files or entire directories.
  3677. * systemd-tmpfiles "m" lines are now fully equivalent to "z"
  3678. lines. So far, they have been non-globbing versions of the
  3679. latter, and have thus been redundant. In future, it is
  3680. recommended to only use "z". "m" has hence been removed
  3681. from the documentation, even though it stays supported.
  3682. * A tmpfiles snippet to recreate the most basic structure in
  3683. /var has been added. This is enough to create the /var/run →
  3684. /run symlink and create a couple of structural
  3685. directories. This allows systems to boot up with an empty or
  3686. volatile /var. Of course, while with this change, the core OS
  3687. now is capable with dealing with a volatile /var, not all
  3688. user services are ready for it. However, we hope that sooner
  3689. or later, many service daemons will be changed upstream so
  3690. that they are able to automatically create their necessary
  3691. directories in /var at boot, should they be missing. This is
  3692. the first step to allow state-less systems that only require
  3693. the vendor image for /usr to boot.
  3694. * systemd-nspawn has gained a new --tmpfs= switch to mount an
  3695. empty tmpfs instance to a specific directory. This is
  3696. particularly useful for making use of the automatic
  3697. reconstruction of /var (see above), by passing --tmpfs=/var.
  3698. * Access modes specified in tmpfiles snippets may now be
  3699. prefixed with "~", which indicates that they shall be masked
  3700. by whether the existing file or directory is currently
  3701. writable, readable or executable at all. Also, if specified,
  3702. the sgid/suid/sticky bits will be masked for all
  3703. non-directories.
  3704. * A new passive target unit "" has been
  3705. added which is useful for services that shall run before any
  3706. network is configured, for example firewall scripts.
  3707. * The "floppy" group that previously owned the /dev/fd*
  3708. devices is no longer used. The "disk" group is now used
  3709. instead. Distributions should probably deprecate usage of
  3710. this group.
  3711. Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
  3712. King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
  3713. Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
  3714. Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
  3715. Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
  3716. Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
  3717. Jędrzejewski-Szmek
  3718. — Berlin, 2014-06-11
  3719. CHANGES WITH 213:
  3720. * A new "systemd-timesyncd" daemon has been added for
  3721. synchronizing the system clock across the network. It
  3722. implements an SNTP client. In contrast to NTP
  3723. implementations such as chrony or the NTP reference server,
  3724. this only implements a client side, and does not bother with
  3725. the full NTP complexity, focusing only on querying time from
  3726. one remote server and synchronizing the local clock to
  3727. it. Unless you intend to serve NTP to networked clients or
  3728. want to connect to local hardware clocks, this simple NTP
  3729. client should be more than appropriate for most
  3730. installations. The daemon runs with minimal privileges, and
  3731. has been hooked up with networkd to only operate when
  3732. network connectivity is available. The daemon saves the
  3733. current clock to disk every time a new NTP sync has been
  3734. acquired, and uses this to possibly correct the system clock
  3735. early at bootup, in order to accommodate for systems that
  3736. lack an RTC such as the Raspberry Pi and embedded devices,
  3737. and to make sure that time monotonically progresses on these
  3738. systems, even if it is not always correct. To make use of
  3739. this daemon, a new system user and group "systemd-timesync"
  3740. needs to be created on installation of systemd.
  3741. * The queue "seqnum" interface of libudev has been disabled, as
  3742. it was generally incompatible with device namespacing as
  3743. sequence numbers of devices go "missing" if the devices are
  3744. part of a different namespace.
  3745. * "systemctl list-timers" and "systemctl list-sockets" gained
  3746. a --recursive switch for showing units of these types also
  3747. for all local containers, similar in style to the already
  3748. supported --recursive switch for "systemctl list-units".
  3749. * A new RebootArgument= setting has been added for service
  3750. units, which may be used to specify a kernel reboot argument
  3751. to use when triggering reboots with StartLimitAction=.
  3752. * A new FailureAction= setting has been added for service
  3753. units which may be used to specify an operation to trigger
  3754. when a service fails. This works similarly to
  3755. StartLimitAction=, but unlike it, controls what is done
  3756. immediately rather than only after several attempts to
  3757. restart the service in question.
  3758. * hostnamed got updated to also expose the kernel name,
  3759. release, and version on the bus. This is useful for
  3760. executing commands like hostnamectl with the -H switch.
  3761. systemd-analyze makes use of this to properly display
  3762. details when running non-locally.
  3763. * The bootchart tool can now show cgroup information in the
  3764. graphs it generates.
  3765. * The CFS CPU quota cgroup attribute is now exposed for
  3766. services. The new CPUQuota= switch has been added for this
  3767. which takes a percentage value. Setting this will have the
  3768. result that a service may never get more CPU time than the
  3769. specified percentage, even if the machine is otherwise idle.
  3770. * systemd-networkd learned IPIP and SIT tunnel support.
  3771. * LSB init scripts exposing a dependency on $network will now
  3772. get a dependency on rather than simply
  3773. This should bring LSB handling closer to
  3774. what it was on SysV systems.
  3775. * A new kernel option has been added to control
  3776. how fsck shall deal with unclean file systems at boot.
  3777. * The (.ini) configuration file parser will now silently
  3778. ignore sections whose name begins with "X-". This may be
  3779. used to maintain application-specific extension sections in unit
  3780. files.
  3781. * machined gained a new API to query the IP addresses of
  3782. registered containers. "machinectl status" has been updated
  3783. to show these addresses in its output.
  3784. * A new call sd_uid_get_display() has been added to the
  3785. sd-login APIs for querying the "primary" session of a
  3786. user. The "primary" session of the user is elected from the
  3787. user's sessions and generally a graphical session is
  3788. preferred over a text one.
  3789. * A minimal systemd-resolved daemon has been added. It
  3790. currently simply acts as a companion to systemd-networkd and
  3791. manages resolv.conf based on per-interface DNS
  3792. configuration, possibly supplied via DHCP. In the long run
  3793. we hope to extend this into a local DNSSEC enabled DNS and
  3794. mDNS cache.
  3795. * The systemd-networkd-wait-online tool is now enabled by
  3796. default. It will delay until a network
  3797. connection has been configured. The tool primarily integrates
  3798. with networkd, but will also make a best effort to make sense
  3799. of network configuration performed in some other way.
  3800. * Two new service options StartupCPUShares= and
  3801. StartupBlockIOWeight= have been added that work similarly to
  3802. CPUShares= and BlockIOWeight= however only apply during
  3803. system startup. This is useful to prioritize certain services
  3804. differently during bootup than during normal runtime.
  3805. * hostnamed has been changed to prefer the statically
  3806. configured hostname in /etc/hostname (unless set to
  3807. 'localhost' or empty) over any dynamic one supplied by
  3808. dhcp. With this change, the rules for picking the hostname
  3809. match more closely the rules of other configuration settings
  3810. where the local administrator's configuration in /etc always
  3811. overrides any other settings.
  3812. Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
  3813. den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
  3814. Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
  3815. David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
  3816. Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
  3817. Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
  3818. Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
  3819. Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
  3820. Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
  3821. Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
  3822. Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
  3823. Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
  3824. Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
  3825. Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
  3826. Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
  3827. Lindskog, WaLyong Cho, Will Woods, Zbigniew
  3828. Jędrzejewski-Szmek
  3829. — Beijing, 2014-05-28
  3830. CHANGES WITH 212:
  3831. * When restoring the screen brightness at boot, stay away from
  3832. the darkest setting or from the lowest 5% of the available
  3833. range, depending on which is the larger value of both. This
  3834. should effectively protect the user from rebooting into a
  3835. black screen, should the brightness have been set to minimum
  3836. by accident.
  3837. * sd-login gained a new sd_machine_get_class() call to
  3838. determine the class ("vm" or "container") of a machine
  3839. registered with machined.
  3840. * sd-login gained new calls
  3841. sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
  3842. to query the identity of the peer of a local AF_UNIX
  3843. connection. They operate similarly to their sd_pid_get_xyz()
  3844. counterparts.
  3845. * PID 1 will now maintain a system-wide system state engine
  3846. with the states "starting", "running", "degraded",
  3847. "maintenance", "stopping". These states are bound to system
  3848. startup, normal runtime, runtime with at least one failed
  3849. service, rescue/emergency mode and system shutdown. This
  3850. state is shown in the "systemctl status" output when no unit
  3851. name is passed. It is useful to determine system state, in
  3852. particularly when doing so for many systems or containers at
  3853. once.
  3854. * A new command "list-machines" has been added to "systemctl"
  3855. that lists all local OS containers and shows their system
  3856. state (see above), if systemd runs inside of them.
  3857. * systemctl gained a new "-r" switch to recursively enumerate
  3858. units on all local containers, when used with the
  3859. "list-unit" command (which is the default one that is
  3860. executed when no parameters are specified).
  3861. * The GPT automatic partition discovery logic will now honour
  3862. two GPT partition flags: one may be set on a partition to
  3863. cause it to be mounted read-only, and the other may be set
  3864. on a partition to ignore it during automatic discovery.
  3865. * Two new GPT type UUIDs have been added for automatic root
  3866. partition discovery, for 32-bit and 64-bit ARM. This is not
  3867. particularly useful for discovering the root directory on
  3868. these architectures during bare-metal boots (since UEFI is
  3869. not common there), but still very useful to allow booting of
  3870. ARM disk images in nspawn with the -i option.
  3871. * MAC addresses of interfaces created with nspawn's
  3872. --network-interface= switch will now be generated from the
  3873. machine name, and thus be stable between multiple invocations
  3874. of the container.
  3875. * logind will now automatically remove all IPC objects owned
  3876. by a user if she or he fully logs out. This makes sure that
  3877. users who are logged out cannot continue to consume IPC
  3878. resources. This covers SysV memory, semaphores and message
  3879. queues as well as POSIX shared memory and message
  3880. queues. Traditionally, SysV and POSIX IPC had no life-cycle
  3881. limits. With this functionality, that is corrected. This may
  3882. be turned off by using the RemoveIPC= switch of logind.conf.
  3883. * The systemd-machine-id-setup and tmpfiles tools gained a
  3884. --root= switch to operate on a specific root directory,
  3885. instead of /.
  3886. * journald can now forward logged messages to the TTYs of all
  3887. logged in users ("wall"). This is the default for all
  3888. emergency messages now.
  3889. * A new tool systemd-journal-remote has been added to stream
  3890. journal log messages across the network.
  3891. * /sys/fs/cgroup/ is now mounted read-only after all cgroup
  3892. controller trees are mounted into it. Note that the
  3893. directories mounted beneath it are not read-only. This is a
  3894. security measure and is particularly useful because glibc
  3895. actually includes a search logic to pick any tmpfs it can
  3896. find to implement shm_open() if /dev/shm is not available
  3897. (which it might very well be in namespaced setups).
  3898. * machinectl gained a new "poweroff" command to cleanly power
  3899. down a local OS container.
  3900. * The PrivateDevices= unit file setting will now also drop the
  3901. CAP_MKNOD capability from the capability bound set, and
  3902. imply DevicePolicy=closed.
  3903. * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
  3904. comprehensively on all long-running systemd services where
  3905. this is appropriate.
  3906. * systemd-udevd will now run in a disassociated mount
  3907. namespace. To mount directories from udev rules, make sure to
  3908. pull in mount units via SYSTEMD_WANTS properties.
  3909. * The kdbus support gained support for uploading policy into
  3910. the kernel. sd-bus gained support for creating "monitoring"
  3911. connections that can eavesdrop into all bus communication
  3912. for debugging purposes.
  3913. * Timestamps may now be specified in seconds since the UNIX
  3914. epoch Jan 1st, 1970 by specifying "@" followed by the value
  3915. in seconds.
  3916. * Native tcpwrap support in systemd has been removed. tcpwrap
  3917. is old code, not really maintained anymore and has serious
  3918. shortcomings, and better options such as firewalls
  3919. exist. For setups that require tcpwrap usage, please
  3920. consider invoking your socket-activated service via tcpd,
  3921. like on traditional inetd.
  3922. * A new system.conf configuration option
  3923. DefaultTimerAccuracySec= has been added that controls the
  3924. default AccuracySec= setting of .timer units.
  3925. * Timer units gained a new WakeSystem= switch. If enabled,
  3926. timers configured this way will cause the system to resume
  3927. from system suspend (if the system supports that, which most
  3928. do these days).
  3929. * Timer units gained a new Persistent= switch. If enabled,
  3930. timers configured this way will save to disk when they have
  3931. been last triggered. This information is then used on next
  3932. reboot to possible execute overdue timer events, that
  3933. could not take place because the system was powered off.
  3934. This enables simple anacron-like behaviour for timer units.
  3935. * systemctl's "list-timers" will now also list the time a
  3936. timer unit was last triggered in addition to the next time
  3937. it will be triggered.
  3938. * systemd-networkd will now assign predictable IPv4LL
  3939. addresses to its local interfaces.
  3940. Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
  3941. Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
  3942. Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
  3943. Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
  3944. Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
  3945. Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
  3946. Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
  3947. Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
  3948. Jędrzejewski-Szmek
  3949. — Berlin, 2014-03-25
  3950. CHANGES WITH 211:
  3951. * A new unit file setting RestrictAddressFamilies= has been
  3952. added to restrict which socket address families unit
  3953. processes gain access to. This takes address family names
  3954. like "AF_INET" or "AF_UNIX", and is useful to minimize the
  3955. attack surface of services via exotic protocol stacks. This
  3956. is built on seccomp system call filters.
  3957. * Two new unit file settings RuntimeDirectory= and
  3958. RuntimeDirectoryMode= have been added that may be used to
  3959. manage a per-daemon runtime directories below /run. This is
  3960. an alternative for setting up directory permissions with
  3961. tmpfiles snippets, and has the advantage that the runtime
  3962. directory's lifetime is bound to the daemon runtime and that
  3963. the daemon starts up with an empty directory each time. This
  3964. is particularly useful when writing services that drop
  3965. privileges using the User= or Group= setting.
  3966. * The DeviceAllow= unit setting now supports globbing for
  3967. matching against device group names.
  3968. * The systemd configuration file system.conf gained new
  3969. settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
  3970. DefaultMemoryAccounting= to globally turn on/off accounting
  3971. for specific resources (cgroups) for all units. These
  3972. settings may still be overridden individually in each unit
  3973. though.
  3974. * systemd-gpt-auto-generator is now able to discover /srv and
  3975. root partitions in addition to /home and swap partitions. It
  3976. also supports LUKS-encrypted partitions now. With this in
  3977. place, automatic discovery of partitions to mount following
  3978. the Discoverable Partitions Specification
  3979. (
  3980. is now a lot more complete. This allows booting without
  3981. /etc/fstab and without root= on the kernel command line on
  3982. systems prepared appropriately.
  3983. * systemd-nspawn gained a new --image= switch which allows
  3984. booting up disk images and Linux installations on any block
  3985. device that follow the Discoverable Partitions Specification
  3986. (see above). This means that installations made with
  3987. appropriately updated installers may now be started and
  3988. deployed using container managers, completely
  3989. unmodified. (We hope that libvirt-lxc will add support for
  3990. this feature soon, too.)
  3991. * systemd-nspawn gained a new --network-macvlan= setting to
  3992. set up a private macvlan interface for the
  3993. container. Similarly, systemd-networkd gained a new
  3994. Kind=macvlan setting in .netdev files.
  3995. * systemd-networkd now supports configuring local addresses
  3996. using IPv4LL.
  3997. * A new tool systemd-network-wait-online has been added to
  3998. synchronously wait for network connectivity using
  3999. systemd-networkd.
  4000. * The sd-bus.h bus API gained a new sd_bus_track object for
  4001. tracking the life-cycle of bus peers. Note that sd-bus.h is
  4002. still not a public API though (unless you specify
  4003. --enable-kdbus on the configure command line, which however
  4004. voids your warranty and you get no API stability guarantee).
  4005. * The $XDG_RUNTIME_DIR runtime directories for each user are
  4006. now individual tmpfs instances, which has the benefit of
  4007. introducing separate pools for each user, with individual
  4008. size limits, and thus making sure that unprivileged clients
  4009. can no longer negatively impact the system or other users by
  4010. filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
  4011. RuntimeDirectorySize= has been introduced that allows
  4012. controlling the default size limit for all users. It
  4013. defaults to 10% of the available physical memory. This is no
  4014. replacement for quotas on tmpfs though (which the kernel
  4015. still does not support), as /dev/shm and /tmp are still
  4016. shared resources used by both the system and unprivileged
  4017. users.
  4018. * logind will now automatically turn off automatic suspending
  4019. on laptop lid close when more than one display is
  4020. connected. This was previously expected to be implemented
  4021. individually in desktop environments (such as GNOME),
  4022. however has been added to logind now, in order to fix a
  4023. boot-time race where a desktop environment might not have
  4024. been started yet and thus not been able to take an inhibitor
  4025. lock at the time where logind already suspends the system
  4026. due to a closed lid.
  4027. * logind will now wait at least 30s after each system
  4028. suspend/resume cycle, and 3min after system boot before
  4029. suspending the system due to a closed laptop lid. This
  4030. should give USB docking stations and similar enough time to
  4031. be probed and configured after system resume and boot in
  4032. order to then act as suspend blocker.
  4033. * systemd-run gained a new --property= setting which allows
  4034. initialization of resource control properties (and others)
  4035. for the created scope or service unit. Example: "systemd-run
  4036. --property=BlockIOWeight=10 updatedb" may be used to run
  4037. updatedb at a low block IO scheduling weight.
  4038. * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
  4039. now also work in --scope mode.
  4040. * When systemd is compiled with kdbus support, basic support
  4041. for enforced policies is now in place. (Note that enabling
  4042. kdbus still voids your warranty and no API compatibility
  4043. promises are made.)
  4044. Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
  4045. K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
  4046. Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
  4047. Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
  4048. Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
  4049. Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
  4050. Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
  4051. Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
  4052. Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
  4053. Zbigniew Jędrzejewski-Szmek
  4054. — Berlin, 2014-03-12
  4055. CHANGES WITH 210:
  4056. * systemd will now relabel /dev after loading the SMACK policy
  4057. according to SMACK rules.
  4058. * A new unit file option AppArmorProfile= has been added to
  4059. set the AppArmor profile for the processes of a unit.
  4060. * A new condition check ConditionArchitecture= has been added
  4061. to conditionalize units based on the system architecture, as
  4062. reported by uname()'s "machine" field.
  4063. * systemd-networkd now supports matching on the system
  4064. virtualization, architecture, kernel command line, host name
  4065. and machine ID.
  4066. * logind is now a lot more aggressive when suspending the
  4067. machine due to a closed laptop lid. Instead of acting only
  4068. on the lid close action, it will continuously watch the lid
  4069. status and act on it. This is useful for laptops where the
  4070. power button is on the outside of the chassis so that it can
  4071. be reached without opening the lid (such as the Lenovo
  4072. Yoga). On those machines, logind will now immediately
  4073. re-suspend the machine if the power button has been
  4074. accidentally pressed while the laptop was suspended and in a
  4075. backpack or similar.
  4076. * logind will now watch SW_DOCK switches and inhibit reaction
  4077. to the lid switch if it is pressed. This means that logind
  4078. will not suspend the machine anymore if the lid is closed
  4079. and the system is docked, if the laptop supports SW_DOCK
  4080. notifications via the input layer. Note that ACPI docking
  4081. stations do not generate this currently. Also note that this
  4082. logic is usually not fully sufficient and Desktop
  4083. Environments should take a lid switch inhibitor lock when an
  4084. external display is connected, as systemd will not watch
  4085. this on its own.
  4086. * nspawn will now make use of the devices cgroup controller by
  4087. default, and only permit creation of and access to the usual
  4088. API device nodes like /dev/null or /dev/random, as well as
  4089. access to (but not creation of) the pty devices.
  4090. * We will now ship a default .network file for
  4091. systemd-networkd that automatically configures DHCP for
  4092. network interfaces created by nspawn's --network-veth or
  4093. --network-bridge= switches.
  4094. * systemd will now understand the usual M, K, G, T suffixes
  4095. according to SI conventions (i.e. to the base 1000) when
  4096. referring to throughput and hardware metrics. It will stay
  4097. with IEC conventions (i.e. to the base 1024) for software
  4098. metrics, according to what is customary according to
  4099. Wikipedia. We explicitly document which base applies for
  4100. each configuration option.
  4101. * The DeviceAllow= setting in unit files now supports a syntax
  4102. to whitelist an entire group of devices node majors at once,
  4103. based on the /proc/devices listing. For example, with the
  4104. string "char-pts", it is now possible to whitelist all
  4105. current and future pseudo-TTYs at once.
  4106. * sd-event learned a new "post" event source. Event sources of
  4107. this type are triggered by the dispatching of any event
  4108. source of a type that is not "post". This is useful for
  4109. implementing clean-up and check event sources that are
  4110. triggered by other work being done in the program.
  4111. * systemd-networkd is no longer statically enabled, but uses
  4112. the usual [Install] sections so that it can be
  4113. enabled/disabled using systemctl. It still is enabled by
  4114. default however.
  4115. * When creating a veth interface pair with systemd-nspawn, the
  4116. host side will now be prefixed with "vb-" if
  4117. --network-bridge= is used, and with "ve-" if --network-veth
  4118. is used. This way, it is easy to distinguish these cases on
  4119. the host, for example to apply different configuration to
  4120. them with systemd-networkd.
  4121. * The compatibility libraries for,
  4122., and
  4123. do not make use of IFUNC
  4124. anymore. Instead, we now build multiple times
  4125. under these alternative names. This means that the footprint
  4126. is drastically increased, but given that these are
  4127. transitional compatibility libraries, this should not matter
  4128. much. This change has been made necessary to support the ARM
  4129. platform for these compatibility libraries, as the ARM
  4130. toolchain is not really at the same level as the toolchain
  4131. for other architectures like x86 and does not support
  4132. IFUNC. Please make sure to use --enable-compat-libs only
  4133. during a transitional period!
  4134. Contributions from: Andreas Fuchs, Armin K., Colin Walters,
  4135. Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
  4136. Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
  4137. St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
  4138. Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
  4139. Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
  4140. Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
  4141. Zbigniew Jędrzejewski-Szmek
  4142. — Berlin, 2014-02-24
  4143. CHANGES WITH 209:
  4144. * A new component "systemd-networkd" has been added that can
  4145. be used to configure local network interfaces statically or
  4146. via DHCP. It is capable of bringing up bridges, VLANs, and
  4147. bonding. Currently, no hook-ups for interactive network
  4148. configuration are provided. Use this for your initrd,
  4149. container, embedded, or server setup if you need a simple,
  4150. yet powerful, network configuration solution. This
  4151. configuration subsystem is quite nifty, as it allows wildcard
  4152. hotplug matching in interfaces. For example, with a single
  4153. configuration snippet, you can configure that all Ethernet
  4154. interfaces showing up are automatically added to a bridge,
  4155. or similar. It supports link-sensing and more.
  4156. * A new tool "systemd-socket-proxyd" has been added which can
  4157. act as a bidirectional proxy for TCP sockets. This is
  4158. useful for adding socket activation support to services that
  4159. do not actually support socket activation, including virtual
  4160. machines and the like.
  4161. * Add a new tool to save/restore rfkill state on
  4162. shutdown/boot.
  4163. * Save/restore state of keyboard backlights in addition to
  4164. display backlights on shutdown/boot.
  4165. * udev learned a new SECLABEL{} construct to label device
  4166. nodes with a specific security label when they appear. For
  4167. now, only SECLABEL{selinux} is supported, but the syntax is
  4168. prepared for additional security frameworks.
  4169. * udev gained a new scheme to configure link-level attributes
  4170. from files in /etc/systemd/network/*.link. These files can
  4171. match against MAC address, device path, driver name and type,
  4172. and will apply attributes like the naming policy, link speed,
  4173. MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
  4174. address assignment policy (randomized, ...).
  4175. * The configuration of network interface naming rules for
  4176. "permanent interface names" has changed: a new NamePolicy=
  4177. setting in the [Link] section of .link files determines the
  4178. priority of possible naming schemes (onboard, slot, MAC,
  4179. path). The default value of this setting is determined by
  4180. /usr/lib/net/links/ Old
  4181. 80-net-name-slot.rules udev configuration file has been
  4182. removed, so local configuration overriding this file should
  4183. be adapted to override instead.
  4184. * When the User= switch is used in a unit file, also
  4185. initialize $SHELL= based on the user database entry.
  4186. * systemd no longer depends on libdbus. All communication is
  4187. now done with sd-bus, systemd's low-level bus library
  4188. implementation.
  4189. * kdbus support has been added to PID 1 itself. When kdbus is
  4190. enabled, this causes PID 1 to set up the system bus and
  4191. enable support for a new ".busname" unit type that
  4192. encapsulates bus name activation on kdbus. It works a little
  4193. bit like ".socket" units, except for bus names. A new
  4194. generator has been added that converts classic dbus1 service
  4195. activation files automatically into native systemd .busname
  4196. and .service units.
  4197. * sd-bus: add a light-weight vtable implementation that allows
  4198. defining objects on the bus with a simple static const
  4199. vtable array of its methods, signals and properties.
  4200. * systemd will not generate or install static dbus
  4201. introspection data anymore to /usr/share/dbus-1/interfaces,
  4202. as the precise format of these files is unclear, and
  4203. nothing makes use of it.
  4204. * A proxy daemon is now provided to proxy clients connecting
  4205. via classic D-Bus AF_UNIX sockets to kdbus, to provide full
  4206. compatibility with classic D-Bus.
  4207. * A bus driver implementation has been added that supports the
  4208. classic D-Bus bus driver calls on kdbus, also for
  4209. compatibility purposes.
  4210. * A new API "sd-event.h" has been added that implements a
  4211. minimal event loop API built around epoll. It provides a
  4212. couple of features that direct epoll usage is lacking:
  4213. prioritization of events, scales to large numbers of timer
  4214. events, per-event timer slack (accuracy), system-wide
  4215. coalescing of timer events, exit handlers, watchdog
  4216. supervision support using systemd's sd_notify() API, child
  4217. process handling.
  4218. * A new API "sd-rntl.h" has been added that provides an API
  4219. around the route netlink interface of the kernel, similar in
  4220. style to "sd-bus.h".
  4221. * A new API "sd-dhcp-client.h" has been added that provides a
  4222. small DHCPv4 client-side implementation. This is used by
  4223. "systemd-networkd".
  4224. * There is a new kernel command line option
  4225. "systemd.restore_state=0|1". When set to "0", none of the
  4226. systemd tools will restore saved runtime state to hardware
  4227. devices. More specifically, the rfkill and backlight states
  4228. are not restored.
  4229. * The FsckPassNo= compatibility option in mount/service units
  4230. has been removed. The fstab generator will now add the
  4231. necessary dependencies automatically, and does not require
  4232. PID1's support for that anymore.
  4233. * journalctl gained a new switch, --list-boots, that lists
  4234. recent boots with their times and boot IDs.
  4235. * The various tools like systemctl, loginctl, timedatectl,
  4236. busctl, systemd-run, ... have gained a new switch "-M" to
  4237. connect to a specific, local OS container (as direct
  4238. connection, without requiring SSH). This works on any
  4239. container that is registered with machined, such as those
  4240. created by libvirt-lxc or nspawn.
  4241. * systemd-run and systemd-analyze also gained support for "-H"
  4242. to connect to remote hosts via SSH. This is particularly
  4243. useful for systemd-run because it enables queuing of jobs
  4244. onto remote systems.
  4245. * machinectl gained a new command "login" to open a getty
  4246. login in any local container. This works with any container
  4247. that is registered with machined (such as those created by
  4248. libvirt-lxc or nspawn), and which runs systemd inside.
  4249. * machinectl gained a new "reboot" command that may be used to
  4250. trigger a reboot on a specific container that is registered
  4251. with machined. This works on any container that runs an init
  4252. system of some kind.
  4253. * systemctl gained a new "list-timers" command to print a nice
  4254. listing of installed timer units with the times they elapse
  4255. next.
  4256. * Alternative reboot() parameters may now be specified on the
  4257. "systemctl reboot" command line and are passed to the
  4258. reboot() system call.
  4259. * systemctl gained a new --job-mode= switch to configure the
  4260. mode to queue a job with. This is a more generic version of
  4261. --fail, --irreversible, and --ignore-dependencies, which are
  4262. still available but not advertised anymore.
  4263. * /etc/systemd/system.conf gained new settings to configure
  4264. various default timeouts of units, as well as the default
  4265. start limit interval and burst. These may still be overridden
  4266. within each Unit.
  4267. * PID1 will now export on the bus profile data of the security
  4268. policy upload process (such as the SELinux policy upload to
  4269. the kernel).
  4270. * journald: when forwarding logs to the console, include
  4271. timestamps (following the setting in
  4272. /sys/module/printk/parameters/time).
  4273. * OnCalendar= in timer units now understands the special
  4274. strings "yearly" and "annually". (Both are equivalent)
  4275. * The accuracy of timer units is now configurable with the new
  4276. AccuracySec= setting. It defaults to 1min.
  4277. * A new dependency type JoinsNamespaceOf= has been added that
  4278. allows running two services within the same /tmp and network
  4279. namespace, if PrivateNetwork= or PrivateTmp= are used.
  4280. * A new command "cat" has been added to systemctl. It outputs
  4281. the original unit file of a unit, and concatenates the
  4282. contents of additional "drop-in" unit file snippets, so that
  4283. the full configuration is shown.
  4284. * systemctl now supports globbing on the various "list-xyz"
  4285. commands, like "list-units" or "list-sockets", as well as on
  4286. those commands which take multiple unit names.
  4287. * journalctl's --unit= switch gained support for globbing.
  4288. * All systemd daemons now make use of the watchdog logic so
  4289. that systemd automatically notices when they hang.
  4290. * If the $container_ttys environment variable is set,
  4291. getty-generator will automatically spawn a getty for each
  4292. listed tty. This is useful for container managers to request
  4293. login gettys to be spawned on as many ttys as needed.
  4294. * %h, %s, %U specifier support is not available anymore when
  4295. used in unit files for PID 1. This is because NSS calls are
  4296. not safe from PID 1. They stay available for --user
  4297. instances of systemd, and as special case for the root user.
  4298. * loginctl gained a new "--no-legend" switch to turn off output
  4299. of the legend text.
  4300. * The "sd-login.h" API gained three new calls:
  4301. sd_session_is_remote(), sd_session_get_remote_user(),
  4302. sd_session_get_remote_host() to query information about
  4303. remote sessions.
  4304. * The udev hardware database now also carries vendor/product
  4305. information of SDIO devices.
  4306. * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
  4307. determine whether watchdog notifications are requested by
  4308. the system manager.
  4309. * Socket-activated per-connection services now include a
  4310. short description of the connection parameters in the
  4311. description.
  4312. * tmpfiles gained a new "--boot" option. When this is not used,
  4313. only lines where the command character is not suffixed with
  4314. "!" are executed. When this option is specified, those
  4315. options are executed too. This partitions tmpfiles
  4316. directives into those that can be safely executed at any
  4317. time, and those which should be run only at boot (for
  4318. example, a line that creates /run/nologin).
  4319. * A new API "sd-resolve.h" has been added which provides a simple
  4320. asynchronous wrapper around glibc NSS host name resolution
  4321. calls, such as getaddrinfo(). In contrast to glibc's
  4322. getaddrinfo_a(), it does not use signals. In contrast to most
  4323. other asynchronous name resolution libraries, this one does
  4324. not reimplement DNS, but reuses NSS, so that alternate
  4325. host name resolution systems continue to work, such as mDNS,
  4326. LDAP, etc. This API is based on libasyncns, but it has been
  4327. cleaned up for inclusion in systemd.
  4328. * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
  4329. "sd-daemon.h" are no longer found in individual libraries
  4331., Instead, we have
  4332. merged them into a single library,, which
  4333. provides all symbols. The reason for this is cyclic
  4334. dependencies, as these libraries tend to use each other's
  4335. symbols. So far, we have managed to workaround that by linking
  4336. a copy of a good part of our code into each of these
  4337. libraries again and again, which, however, makes certain
  4338. things hard to do, like sharing static variables. Also, it
  4339. substantially increases footprint. With this change, there
  4340. is only one library for the basic APIs systemd
  4341. provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
  4342. "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
  4343. library as well, however are subject to the --enable-kdbus
  4344. switch (see below). Note that "sd-dhcp-client.h" is not part
  4345. of this library (this is because it only consumes, never
  4346. provides, services of/to other APIs). To make the transition
  4347. easy from the separate libraries to the unified one, we
  4348. provide the --enable-compat-libs compile-time switch which
  4349. will generate stub libraries that are compatible with the
  4350. old ones but redirect all calls to the new one.
  4351. * All of the kdbus logic and the new APIs "sd-bus.h",
  4352. "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
  4353. and "sd-utf8.h" are compile-time optional via the
  4354. "--enable-kdbus" switch, and they are not compiled in by
  4355. default. To make use of kdbus, you have to explicitly enable
  4356. the switch. Note however, that neither the kernel nor the
  4357. userspace API for all of this is considered stable yet. We
  4358. want to maintain the freedom to still change the APIs for
  4359. now. By specifying this build-time switch, you acknowledge
  4360. that you are aware of the instability of the current
  4361. APIs.
  4362. * Also, note that while kdbus is pretty much complete,
  4363. it lacks one thing: proper policy support. This means you
  4364. can build a fully working system with all features; however,
  4365. it will be highly insecure. Policy support will be added in
  4366. one of the next releases, at the same time that we will
  4367. declare the APIs stable.
  4368. * When the kernel command line argument "kdbus" is specified,
  4369. systemd will automatically load the kdbus.ko kernel module. At
  4370. this stage of development, it is only useful for testing kdbus
  4371. and should not be used in production. Note: if "--enable-kdbus"
  4372. is specified, and the kdbus.ko kernel module is available, and
  4373. "kdbus" is added to the kernel command line, the entire system
  4374. runs with kdbus instead of dbus-daemon, with the above mentioned
  4375. problem of missing the system policy enforcement. Also a future
  4376. version of kdbus.ko or a newer systemd will not be compatible with
  4377. each other, and will unlikely be able to boot the machine if only
  4378. one of them is updated.
  4379. * systemctl gained a new "import-environment" command which
  4380. uploads the caller's environment (or parts thereof) into the
  4381. service manager so that it is inherited by services started
  4382. by the manager. This is useful to upload variables like
  4383. $DISPLAY into the user service manager.
  4384. * A new PrivateDevices= switch has been added to service units
  4385. which allows running a service with a namespaced /dev
  4386. directory that does not contain any device nodes for
  4387. physical devices. More specifically, it only includes devices
  4388. such as /dev/null, /dev/urandom, and /dev/zero which are API
  4389. entry points.
  4390. * logind has been extended to support behaviour like VT
  4391. switching on seats that do not support a VT. This makes
  4392. multi-session available on seats that are not the first seat
  4393. (seat0), and on systems where kernel support for VTs has
  4394. been disabled at compile-time.
  4395. * If a process holds a delay lock for system sleep or shutdown
  4396. and fails to release it in time, we will now log its
  4397. identity. This makes it easier to identify processes that
  4398. cause slow suspends or power-offs.
  4399. * When parsing /etc/crypttab, support for a new key-slot=
  4400. option as supported by Debian is added. It allows indicating
  4401. which LUKS slot to use on disk, speeding up key loading.
  4402. * The sd_journald_sendv() API call has been checked and
  4403. officially declared to be async-signal-safe so that it may
  4404. be invoked from signal handlers for logging purposes.
  4405. * Boot-time status output is now enabled automatically after a
  4406. short timeout if boot does not progress, in order to give
  4407. the user an indication what she or he is waiting for.
  4408. * The boot-time output has been improved to show how much time
  4409. remains until jobs expire.
  4410. * The KillMode= switch in service units gained a new possible
  4411. value "mixed". If set, and the unit is shut down, then the
  4412. initial SIGTERM signal is sent only to the main daemon
  4413. process, while the following SIGKILL signal is sent to
  4414. all remaining processes of the service.
  4415. * When a scope unit is registered, a new property "Controller"
  4416. may be set. If set to a valid bus name, systemd will send a
  4417. RequestStop() signal to this name when it would like to shut
  4418. down the scope. This may be used to hook manager logic into
  4419. the shutdown logic of scope units. Also, scope units may now
  4420. be put in a special "abandoned" state, in which case the
  4421. manager process which created them takes no further
  4422. responsibilities for it.
  4423. * When reading unit files, systemd will now verify
  4424. the access mode of these files, and warn about certain
  4425. suspicious combinations. This has been added to make it
  4426. easier to track down packaging bugs where unit files are
  4427. marked executable or world-writable.
  4428. * systemd-nspawn gained a new "--setenv=" switch to set
  4429. container-wide environment variables. The similar option in
  4430. systemd-activate was renamed from "--environment=" to
  4431. "--setenv=" for consistency.
  4432. * systemd-nspawn has been updated to create a new kdbus domain
  4433. for each container that is invoked, thus allowing each
  4434. container to have its own set of system and user buses,
  4435. independent of the host.
  4436. * systemd-nspawn gained a new --drop-capability= switch to run
  4437. the container with less capabilities than the default. Both
  4438. --drop-capability= and --capability= now take the special
  4439. string "all" for dropping or keeping all capabilities.
  4440. * systemd-nspawn gained new switches for executing containers
  4441. with specific SELinux labels set.
  4442. * systemd-nspawn gained a new --quiet switch to not generate
  4443. any additional output but the container's own console
  4444. output.
  4445. * systemd-nspawn gained a new --share-system switch to run a
  4446. container without PID namespacing enabled.
  4447. * systemd-nspawn gained a new --register= switch to control
  4448. whether the container is registered with systemd-machined or
  4449. not. This is useful for containers that do not run full
  4450. OS images, but only specific apps.
  4451. * systemd-nspawn gained a new --keep-unit which may be used
  4452. when invoked as the only program from a service unit, and
  4453. results in registration of the unit service itself in
  4454. systemd-machined, instead of a newly opened scope unit.
  4455. * systemd-nspawn gained a new --network-interface= switch for
  4456. moving arbitrary interfaces to the container. The new
  4457. --network-veth switch creates a virtual Ethernet connection
  4458. between host and container. The new --network-bridge=
  4459. switch then allows assigning the host side of this virtual
  4460. Ethernet connection to a bridge device.
  4461. * systemd-nspawn gained a new --personality= switch for
  4462. setting the kernel personality for the container. This is
  4463. useful when running a 32-bit container on a 64-bit host. A
  4464. similar option Personality= is now also available for service
  4465. units to use.
  4466. * logind will now also track a "Desktop" identifier for each
  4467. session which encodes the desktop environment of it. This is
  4468. useful for desktop environments that want to identify
  4469. multiple running sessions of itself easily.
  4470. * A new SELinuxContext= setting for service units has been
  4471. added that allows setting a specific SELinux execution
  4472. context for a service.
  4473. * Most systemd client tools will now honour $SYSTEMD_LESS for
  4474. settings of the "less" pager. By default, these tools will
  4475. override $LESS to allow certain operations to work, such as
  4476. jump-to-the-end. With $SYSTEMD_LESS, it is possible to
  4477. influence this logic.
  4478. * systemd's "seccomp" hook-up has been changed to make use of
  4479. the libseccomp library instead of using its own
  4480. implementation. This has benefits for portability among
  4481. other things.
  4482. * For usage together with SystemCallFilter=, a new
  4483. SystemCallErrorNumber= setting has been introduced that
  4484. allows configuration of a system error number to be returned
  4485. on filtered system calls, instead of immediately killing the
  4486. process. Also, SystemCallArchitectures= has been added to
  4487. limit access to system calls of a particular architecture
  4488. (in order to turn off support for unused secondary
  4489. architectures). There is also a global
  4490. SystemCallArchitectures= setting in system.conf now to turn
  4491. off support for non-native system calls system-wide.
  4492. * systemd requires a kernel with a working name_to_handle_at(),
  4493. please see the kernel config requirements in the README file.
  4494. Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
  4495. Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
  4496. Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
  4497. Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
  4498. Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
  4499. David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
  4500. Elia Pinto, Florian Weimer, George McCollister, Goffredo
  4501. Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
  4502. Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
  4503. Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
  4504. Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
  4505. Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
  4506. Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
  4507. Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
  4508. Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
  4509. Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
  4510. Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
  4511. Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
  4512. Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
  4513. Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
  4514. Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
  4515. Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
  4516. Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
  4517. Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
  4518. — Berlin, 2014-02-20
  4519. CHANGES WITH 208:
  4520. * logind has gained support for facilitating privileged input
  4521. and drm device access for unprivileged clients. This work is
  4522. useful to allow Wayland display servers (and similar
  4523. programs, such as kmscon) to run under the user's ID and
  4524. access input and drm devices which are normally
  4525. protected. When this is used (and the kernel is new enough)
  4526. logind will "mute" IO on the file descriptors passed to
  4527. Wayland as long as it is in the background and "unmute" it
  4528. if it returns into the foreground. This allows secure
  4529. session switching without allowing background sessions to
  4530. eavesdrop on input and display data. This also introduces
  4531. session switching support if VT support is turned off in the
  4532. kernel, and on seats that are not seat0.
  4533. * A new kernel command line option luks.options= is understood
  4534. now which allows specifying LUKS options for usage for LUKS
  4535. encrypted partitions specified with luks.uuid=.
  4536. * tmpfiles.d(5) snippets may now use specifier expansion in
  4537. path names. More specifically %m, %b, %H, %v, are now
  4538. replaced by the local machine id, boot id, hostname, and
  4539. kernel version number.
  4540. * A new tmpfiles.d(5) command "m" has been introduced which
  4541. may be used to change the owner/group/access mode of a file
  4542. or directory if it exists, but do nothing if it does not.
  4543. * This release removes high-level support for the
  4544. MemorySoftLimit= cgroup setting. The underlying kernel
  4545. cgroup attribute memory.soft_limit= is currently badly
  4546. designed and likely to be removed from the kernel API in its
  4547. current form, hence we should not expose it for now.
  4548. * The memory.use_hierarchy cgroup attribute is now enabled for
  4549. all cgroups systemd creates in the memory cgroup
  4550. hierarchy. This option is likely to be come the built-in
  4551. default in the kernel anyway, and the non-hierarchical mode
  4552. never made much sense in the intrinsically hierarchical
  4553. cgroup system.
  4554. * A new field _SYSTEMD_SLICE= is logged along with all journal
  4555. messages containing the slice a message was generated
  4556. from. This is useful to allow easy per-customer filtering of
  4557. logs among other things.
  4558. * systemd-journald will no longer adjust the group of journal
  4559. files it creates to the "systemd-journal" group. Instead we
  4560. rely on the journal directory to be owned by the
  4561. "systemd-journal" group, and its setgid bit set, so that the
  4562. kernel file system layer will automatically enforce that
  4563. journal files inherit this group assignment. The reason for
  4564. this change is that we cannot allow NSS look-ups from
  4565. journald which would be necessary to resolve
  4566. "systemd-journal" to a numeric GID, because this might
  4567. create deadlocks if NSS involves synchronous queries to
  4568. other daemons (such as nscd, or sssd) which in turn are
  4569. logging clients of journald and might block on it, which
  4570. would then dead lock. A tmpfiles.d(5) snippet included in
  4571. systemd will make sure the setgid bit and group are
  4572. properly set on the journal directory if it exists on every
  4573. boot. However, we recommend adjusting it manually after
  4574. upgrades too (or from RPM scriptlets), so that the change is
  4575. not delayed until next reboot.
  4576. * Backlight and random seed files in /var/lib/ have moved into
  4577. the /var/lib/systemd/ directory, in order to centralize all
  4578. systemd generated files in one directory.
  4579. * Boot time performance measurements (as displayed by
  4580. "systemd-analyze" for example) will now read ACPI 5.0 FPDT
  4581. performance information if that's available to determine how
  4582. much time BIOS and boot loader initialization required. With
  4583. a sufficiently new BIOS you hence no longer need to boot
  4584. with Gummiboot to get access to such information.
  4585. Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
  4586. Cristian Rodríguez, Dave Reisner, David Herrmann, David
  4587. Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
  4588. feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
  4589. Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
  4590. Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
  4591. Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
  4592. — Berlin, 2013-10-02
  4593. CHANGES WITH 207:
  4594. * The Restart= option for services now understands a new
  4595. on-watchdog setting, which will restart the service
  4596. automatically if the service stops sending out watchdog keep
  4597. alive messages (as configured with WatchdogSec=).
  4598. * The getty generator (which is responsible for bringing up a
  4599. getty on configured serial consoles) will no longer only
  4600. start a getty on the primary kernel console but on all
  4601. others, too. This makes the order in which console= is
  4602. specified on the kernel command line less important.
  4603. * libsystemd-logind gained a new sd_session_get_vt() call to
  4604. retrieve the VT number of a session.
  4605. * If the option "tries=0" is set for an entry of /etc/crypttab
  4606. its passphrase is queried indefinitely instead of any
  4607. maximum number of tries.
  4608. * If a service with a configure PID file terminates its PID
  4609. file will now be removed automatically if it still exists
  4610. afterwards. This should put an end to stale PID files.
  4611. * systemd-run will now also take relative binary path names
  4612. for execution and no longer insists on absolute paths.
  4613. * InaccessibleDirectories= and ReadOnlyDirectories= now take
  4614. paths that are optionally prefixed with "-" to indicate that
  4615. it should not be considered a failure if they do not exist.
  4616. * journalctl -o (and similar commands) now understands a new
  4617. output mode "short-precise", it is similar to "short" but
  4618. shows timestamps with usec accuracy.
  4619. * The option "discard" (as known from Debian) is now
  4620. synonymous to "allow-discards" in /etc/crypttab. In fact,
  4621. "discard" is preferred now (since it is easier to remember
  4622. and type).
  4623. * Some licensing clean-ups were made, so that more code is now
  4624. LGPL-2.1 licensed than before.
  4625. * A minimal tool to save/restore the display backlight
  4626. brightness across reboots has been added. It will store the
  4627. backlight setting as late as possible at shutdown, and
  4628. restore it as early as possible during reboot.
  4629. * A logic to automatically discover and enable home and swap
  4630. partitions on GPT disks has been added. With this in place
  4631. /etc/fstab becomes optional for many setups as systemd can
  4632. discover certain partitions located on the root disk
  4633. automatically. Home partitions are recognized under their
  4634. GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
  4635. partitions are recognized under their GPT type ID
  4636. 0657fd6da4ab43c484e50933c84b4f4f.
  4637. * systemd will no longer pass any environment from the kernel
  4638. or initrd to system services. If you want to set an
  4639. environment for all services, do so via the kernel command
  4640. line systemd.setenv= assignment.
  4641. * The systemd-sysctl tool no longer natively reads the file
  4642. /etc/sysctl.conf. If desired, the file should be symlinked
  4643. from /etc/sysctl.d/99-sysctl.conf. Apart from providing
  4644. legacy support by a symlink rather than built-in code, it
  4645. also makes the otherwise hidden order of application of the
  4646. different files visible. (Note that this partly reverts to a
  4647. pre-198 application order of sysctl knobs!)
  4648. * The "systemctl set-log-level" and "systemctl dump" commands
  4649. have been moved to systemd-analyze.
  4650. * systemd-run learned the new --remain-after-exit switch,
  4651. which causes the scope unit not to be cleaned up
  4652. automatically after the process terminated.
  4653. * tmpfiles learned a new --exclude-prefix= switch to exclude
  4654. certain paths from operation.
  4655. * journald will now automatically flush all messages to disk
  4656. as soon as a message at the log level CRIT, ALERT or EMERG
  4657. is received.
  4658. Contributions from: Andrew Cook, Brandon Philips, Christian
  4659. Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
  4660. Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
  4661. McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
  4662. Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
  4663. Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
  4664. Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
  4665. Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
  4666. Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
  4667. Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
  4668. Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
  4669. Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
  4670. William Giokas, Zbigniew Jędrzejewski-Szmek
  4671. — Berlin, 2013-09-13
  4672. CHANGES WITH 206:
  4673. * The documentation has been updated to cover the various new
  4674. concepts introduced with 205.
  4675. * Unit files now understand the new %v specifier which
  4676. resolves to the kernel version string as returned by "uname
  4677. -r".
  4678. * systemctl now supports filtering the unit list output by
  4679. load state, active state and sub state, using the new
  4680. --state= parameter.
  4681. * "systemctl status" will now show the results of the
  4682. condition checks (like ConditionPathExists= and similar) of
  4683. the last start attempts of the unit. They are also logged to
  4684. the journal.
  4685. * "journalctl -b" may now be used to look for boot output of a
  4686. specific boot. Try "journalctl -b -1" for the previous boot,
  4687. but the syntax is substantially more powerful.
  4688. * "journalctl --show-cursor" has been added which prints the
  4689. cursor string the last shown log line. This may then be used
  4690. with the new "journalctl --after-cursor=" switch to continue
  4691. browsing logs from that point on.
  4692. * "journalctl --force" may now be used to force regeneration
  4693. of an FSS key.
  4694. * Creation of "dead" device nodes has been moved from udev
  4695. into kmod and tmpfiles. Previously, udev would read the kmod
  4696. databases to pre-generate dead device nodes based on meta
  4697. information contained in kernel modules, so that these would
  4698. be auto-loaded on access rather then at boot. As this
  4699. does not really have much to do with the exposing actual
  4700. kernel devices to userspace this has always been slightly
  4701. alien in the udev codebase. Following the new scheme kmod
  4702. will now generate a runtime snippet for tmpfiles from the
  4703. module meta information and it now is tmpfiles' job to the
  4704. create the nodes. This also allows overriding access and
  4705. other parameters for the nodes using the usual tmpfiles
  4706. facilities. As side effect this allows us to remove the
  4707. CAP_SYS_MKNOD capability bit from udevd entirely.
  4708. * logind's device ACLs may now be applied to these "dead"
  4709. devices nodes too, thus finally allowing managed access to
  4710. devices such as /dev/snd/sequencer without loading the
  4711. backing module right-away.
  4712. * A new RPM macro has been added that may be used to apply
  4713. tmpfiles configuration during package installation.
  4714. * systemd-detect-virt and ConditionVirtualization= now can
  4715. detect User-Mode-Linux machines (UML).
  4716. * journald will now implicitly log the effective capabilities
  4717. set of processes in the message metadata.
  4718. * systemd-cryptsetup has gained support for TrueCrypt volumes.
  4719. * The initrd interface has been simplified (more specifically,
  4720. support for passing performance data via environment
  4721. variables and fsck results via files in /run has been
  4722. removed). These features were non-essential, and are
  4723. nowadays available in a much nicer way by having systemd in
  4724. the initrd serialize its state and have the hosts systemd
  4725. deserialize it again.
  4726. * The udev "keymap" data files and tools to apply keyboard
  4727. specific mappings of scan to key codes, and force-release
  4728. scan code lists have been entirely replaced by a udev
  4729. "keyboard" builtin and a hwdb data file.
  4730. * systemd will now honour the kernel's "quiet" command line
  4731. argument also during late shutdown, resulting in a
  4732. completely silent shutdown when used.
  4733. * There's now an option to control the SO_REUSEPORT socket
  4734. option in .socket units.
  4735. * Instance units will now automatically get a per-template
  4736. subslice of system.slice unless something else is explicitly
  4737. configured. For example, instances of sshd@.service will now
  4738. implicitly be placed in system-sshd.slice rather than
  4739. system.slice as before.
  4740. * Test coverage support may now be enabled at build time.
  4741. Contributions from: Dave Reisner, Frederic Crozat, Harald
  4742. Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
  4743. Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
  4744. Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
  4745. Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
  4746. Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
  4747. Giokas, Zbigniew Jędrzejewski-Szmek
  4748. — Berlin, 2013-07-23
  4749. CHANGES WITH 205:
  4750. * Two new unit types have been introduced:
  4751. Scope units are very similar to service units, however, are
  4752. created out of pre-existing processes — instead of PID 1
  4753. forking off the processes. By using scope units it is
  4754. possible for system services and applications to group their
  4755. own child processes (worker processes) in a powerful way
  4756. which then maybe used to organize them, or kill them
  4757. together, or apply resource limits on them.
  4758. Slice units may be used to partition system resources in an
  4759. hierarchical fashion and then assign other units to them. By
  4760. default there are now three slices: system.slice (for all
  4761. system services), user.slice (for all user sessions),
  4762. machine.slice (for VMs and containers).
  4763. Slices and scopes have been introduced primarily in
  4764. context of the work to move cgroup handling to a
  4765. single-writer scheme, where only PID 1
  4766. creates/removes/manages cgroups.
  4767. * There's a new concept of "transient" units. In contrast to
  4768. normal units these units are created via an API at runtime,
  4769. not from configuration from disk. More specifically this
  4770. means it is now possible to run arbitrary programs as
  4771. independent services, with all execution parameters passed
  4772. in via bus APIs rather than read from disk. Transient units
  4773. make systemd substantially more dynamic then it ever was,
  4774. and useful as a general batch manager.
  4775. * logind has been updated to make use of scope and slice units
  4776. for managing user sessions. As a user logs in he will get
  4777. his own private slice unit, to which all sessions are added
  4778. as scope units. We also added support for automatically
  4779. adding an instance of user@.service for the user into the
  4780. slice. Effectively logind will no longer create cgroup
  4781. hierarchies on its own now, it will defer entirely to PID 1
  4782. for this by means of scope, service and slice units. Since
  4783. user sessions this way become entities managed by PID 1
  4784. the output of "systemctl" is now a lot more comprehensive.
  4785. * A new mini-daemon "systemd-machined" has been added which
  4786. may be used by virtualization managers to register local
  4787. VMs/containers. nspawn has been updated accordingly, and
  4788. libvirt will be updated shortly. machined will collect a bit
  4789. of meta information about the VMs/containers, and assign
  4790. them their own scope unit (see above). The collected
  4791. meta-data is then made available via the "machinectl" tool,
  4792. and exposed in "ps" and similar tools. machined/machinectl
  4793. is compile-time optional.
  4794. * As discussed earlier, the low-level cgroup configuration
  4795. options ControlGroup=, ControlGroupModify=,
  4796. ControlGroupPersistent=, ControlGroupAttribute= have been
  4797. removed. Please use high-level attribute settings instead as
  4798. well as slice units.
  4799. * A new bus call SetUnitProperties() has been added to alter
  4800. various runtime parameters of a unit. This is primarily
  4801. useful to alter cgroup parameters dynamically in a nice way,
  4802. but will be extended later on to make more properties
  4803. modifiable at runtime. systemctl gained a new set-properties
  4804. command that wraps this call.
  4805. * A new tool "systemd-run" has been added which can be used to
  4806. run arbitrary command lines as transient services or scopes,
  4807. while configuring a number of settings via the command
  4808. line. This tool is currently very basic, however already
  4809. very useful. We plan to extend this tool to even allow
  4810. queuing of execution jobs with time triggers from the
  4811. command line, similar in fashion to "at".
  4812. * nspawn will now inform the user explicitly that kernels with
  4813. audit enabled break containers, and suggest the user to turn
  4814. off audit.
  4815. * Support for detecting the IMA and AppArmor security
  4816. frameworks with ConditionSecurity= has been added.
  4817. * journalctl gained a new "-k" switch for showing only kernel
  4818. messages, mimicking dmesg output; in addition to "--user"
  4819. and "--system" switches for showing only user's own logs
  4820. and system logs.
  4821. * systemd-delta can now show information about drop-in
  4822. snippets extending unit files.
  4823. * libsystemd-bus has been substantially updated but is still
  4824. not available as public API.
  4825. * systemd will now look for the "debug" argument on the kernel
  4826. command line and enable debug logging, similar to what
  4827. "systemd.log_level=debug" already did before.
  4828. * "systemctl set-default", "systemctl get-default" has been
  4829. added to configure the symlink, which
  4830. controls what to boot into by default.
  4831. * "systemctl set-log-level" has been added as a convenient
  4832. way to raise and lower systemd logging threshold.
  4833. * "systemd-analyze plot" will now show the time the various
  4834. generators needed for execution, as well as information
  4835. about the unit file loading.
  4836. * libsystemd-journal gained a new sd_journal_open_files() call
  4837. for opening specific journal files. journactl also gained a
  4838. new switch to expose this new functionality. Previously we
  4839. only supported opening all files from a directory, or all
  4840. files from the system, as opening individual files only is
  4841. racy due to journal file rotation.
  4842. * systemd gained the new DefaultEnvironment= setting in
  4843. /etc/systemd/system.conf to set environment variables for
  4844. all services.
  4845. * If a privileged process logs a journal message with the
  4846. OBJECT_PID= field set, then journald will automatically
  4847. augment this with additional OBJECT_UID=, OBJECT_GID=,
  4848. OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
  4849. system services want to log events about specific client
  4850. processes. journactl/systemctl has been updated to make use
  4851. of this information if all log messages regarding a specific
  4852. unit is requested.
  4853. Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
  4854. Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
  4855. Reisner, David Coppa, David King, David Strauss, Eelco
  4856. Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
  4857. Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
  4858. Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
  4859. Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
  4860. Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
  4861. Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
  4862. Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
  4863. Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
  4864. Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
  4865. Łukasz Stelmach, 장동준
  4866. CHANGES WITH 204:
  4867. * The Python bindings gained some minimal support for the APIs
  4868. exposed by libsystemd-logind.
  4869. * ConditionSecurity= gained support for detecting SMACK. Since
  4870. this condition already supports SELinux and AppArmor we only
  4871. miss IMA for this. Patches welcome!
  4872. Contributions from: Karol Lewandowski, Lennart Poettering,
  4873. Zbigniew Jędrzejewski-Szmek
  4874. CHANGES WITH 203:
  4875. * systemd-nspawn will now create /etc/resolv.conf if
  4876. necessary, before bind-mounting the host's file onto it.
  4877. * systemd-nspawn will now store meta information about a
  4878. container on the container's cgroup as extended attribute
  4879. fields, including the root directory.
  4880. * The cgroup hierarchy has been reworked in many ways. All
  4881. objects any of the components systemd creates in the cgroup
  4882. tree are now suffixed. More specifically, user sessions are
  4883. now placed in cgroups suffixed with ".session", users in
  4884. cgroups suffixed with ".user", and nspawn containers in
  4885. cgroups suffixed with ".nspawn". Furthermore, all cgroup
  4886. names are now escaped in a simple scheme to avoid collision
  4887. of userspace object names with kernel filenames. This work
  4888. is preparation for making these objects relocatable in the
  4889. cgroup tree, in order to allow easy resource partitioning of
  4890. these objects without causing naming conflicts.
  4891. * systemctl list-dependencies gained the new switches
  4892. --plain, --reverse, --after and --before.
  4893. * systemd-inhibit now shows the process name of processes that
  4894. have taken an inhibitor lock.
  4895. * nss-myhostname will now also resolve "localhost"
  4896. implicitly. This makes /etc/hosts an optional file and
  4897. nicely handles that on IPv6 ::1 maps to both "localhost" and
  4898. the local hostname.
  4899. * gained a new call
  4900. sd_get_machine_names() to enumerate running containers and
  4901. VMs (currently only supported by very new libvirt and
  4902. nspawn). sd_login_monitor can now be used to watch
  4903. VMs/containers coming and going.
  4904. * .include is not allowed recursively anymore, and only in
  4905. unit files. Usually it is better to use drop-in snippets in
  4906. .d/*.conf anyway, as introduced with systemd 198.
  4907. * systemd-analyze gained a new "critical-chain" command that
  4908. determines the slowest chain of units run during system
  4909. boot-up. It is very useful for tracking down where
  4910. optimizing boot time is the most beneficial.
  4911. * systemd will no longer allow manipulating service paths in
  4912. the name=systemd:/system cgroup tree using ControlGroup= in
  4913. units. (But is still fine with it in all other dirs.)
  4914. * There's a new systemd-nspawn@.service service file that may
  4915. be used to easily run nspawn containers as system
  4916. services. With the container's root directory in
  4917. /var/lib/container/foobar it is now sufficient to run
  4918. "systemctl start systemd-nspawn@foobar.service" to boot it.
  4919. * systemd-cgls gained a new parameter "--machine" to list only
  4920. the processes within a certain container.
  4921. * ConditionSecurity= now can check for "apparmor". We still
  4922. are lacking checks for SMACK and IMA for this condition
  4923. check though. Patches welcome!
  4924. * A new configuration file /etc/systemd/sleep.conf has been
  4925. added that may be used to configure which kernel operation
  4926. systemd is supposed to execute when "suspend", "hibernate"
  4927. or "hybrid-sleep" is requested. This makes the new kernel
  4928. "freeze" state accessible to the user.
  4929. * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
  4930. the passed argument if applicable.
  4931. Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
  4932. Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
  4933. Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
  4934. Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
  4935. MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
  4936. Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
  4937. Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
  4938. Jędrzejewski-Szmek
  4939. CHANGES WITH 202:
  4940. * The output of 'systemctl list-jobs' got some polishing. The
  4941. '--type=' argument may now be passed more than once. A new
  4942. command 'systemctl list-sockets' has been added which shows
  4943. a list of kernel sockets systemd is listening on with the
  4944. socket units they belong to, plus the units these socket
  4945. units activate.
  4946. * The experimental libsystemd-bus library got substantial
  4947. updates to work in conjunction with the (also experimental)
  4948. kdbus kernel project. It works well enough to exchange
  4949. messages with some sophistication. Note that kdbus is not
  4950. ready yet, and the library is mostly an elaborate test case
  4951. for now, and not installable.
  4952. * systemd gained a new unit 'systemd-static-nodes.service'
  4953. that generates static device nodes earlier during boot, and
  4954. can run in conjunction with udev.
  4955. * libsystemd-login gained a new call sd_pid_get_user_unit()
  4956. to retrieve the user systemd unit a process is running
  4957. in. This is useful for systems where systemd is used as
  4958. session manager.
  4959. * systemd-nspawn now places all containers in the new /machine
  4960. top-level cgroup directory in the name=systemd
  4961. hierarchy. libvirt will soon do the same, so that we get a
  4962. uniform separation of /system, /user and /machine for system
  4963. services, user processes and containers/virtual
  4964. machines. This new cgroup hierarchy is also useful to stick
  4965. stable names to specific container instances, which can be
  4966. recognized later this way (this name may be controlled
  4967. via systemd-nspawn's new -M switch). libsystemd-login also
  4968. gained a new call sd_pid_get_machine_name() to retrieve the
  4969. name of the container/VM a specific process belongs to.
  4970. * bootchart can now store its data in the journal.
  4971. * libsystemd-journal gained a new call
  4972. sd_journal_add_conjunction() for AND expressions to the
  4973. matching logic. This can be used to express more complex
  4974. logical expressions.
  4975. * journactl can now take multiple --unit= and --user-unit=
  4976. switches.
  4977. * The cryptsetup logic now understands the "luks.key=" kernel
  4978. command line switch for specifying a file to read the
  4979. decryption key from. Also, if a configured key file is not
  4980. found the tool will now automatically fall back to prompting
  4981. the user.
  4982. * Python systemd.journal module was updated to wrap recently
  4983. added functions from libsystemd-journal. The interface was
  4984. changed to bring the low level interface in s.j._Reader
  4985. closer to the C API, and the high level interface in
  4986. s.j.Reader was updated to wrap and convert all data about
  4987. an entry.
  4988. Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
  4989. Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
  4990. Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
  4991. Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
  4992. Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
  4993. Tom Gundersen, Zbigniew Jędrzejewski-Szmek
  4994. CHANGES WITH 201:
  4995. * journalctl --update-catalog now understands a new --root=
  4996. option to operate on catalogs found in a different root
  4997. directory.
  4998. * During shutdown after systemd has terminated all running
  4999. services a final killing loop kills all remaining left-over
  5000. processes. We will now print the name of these processes
  5001. when we send SIGKILL to them, since this usually indicates a
  5002. problem.
  5003. * If /etc/crypttab refers to password files stored on
  5004. configured mount points automatic dependencies will now be
  5005. generated to ensure the specific mount is established first
  5006. before the key file is attempted to be read.
  5007. * 'systemctl status' will now show information about the
  5008. network sockets a socket unit is listening on.
  5009. * 'systemctl status' will also shown information about any
  5010. drop-in configuration file for units. (Drop-In configuration
  5011. files in this context are files such as
  5012. /etc/systemd/systemd/foobar.service.d/*.conf)
  5013. * systemd-cgtop now optionally shows summed up CPU times of
  5014. cgroups. Press '%' while running cgtop to switch between
  5015. percentage and absolute mode. This is useful to determine
  5016. which cgroups use up the most CPU time over the entire
  5017. runtime of the system. systemd-cgtop has also been updated
  5018. to be 'pipeable' for processing with further shell tools.
  5019. * 'hostnamectl set-hostname' will now allow setting of FQDN
  5020. hostnames.
  5021. * The formatting and parsing of time span values has been
  5022. changed. The parser now understands fractional expressions
  5023. such as "5.5h". The formatter will now output fractional
  5024. expressions for all time spans under 1min, i.e. "5.123456s"
  5025. rather than "5s 123ms 456us". For time spans under 1s
  5026. millisecond values are shown, for those under 1ms
  5027. microsecond values are shown. This should greatly improve
  5028. all time-related output of systemd.
  5029. * libsystemd-login and libsystemd-journal gained new
  5030. functions for querying the poll() events mask and poll()
  5031. timeout value for integration into arbitrary event
  5032. loops.
  5033. * localectl gained the ability to list available X11 keymaps
  5034. (models, layouts, variants, options).
  5035. * 'systemd-analyze dot' gained the ability to filter for
  5036. specific units via shell-style globs, to create smaller,
  5037. more useful graphs. I.e. it is now possible to create simple
  5038. graphs of all the dependencies between only target units, or
  5039. of all units that Avahi has dependencies with.
  5040. Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
  5041. Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
  5042. Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
  5043. Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
  5044. Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
  5045. Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
  5046. Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
  5047. CHANGES WITH 200:
  5048. * The boot-time readahead implementation for rotating media
  5049. will now read the read-ahead data in multiple passes which
  5050. consist of all read requests made in equidistant time
  5051. intervals. This means instead of strictly reading read-ahead
  5052. data in its physical order on disk we now try to find a
  5053. middle ground between physical and access time order.
  5054. * /etc/os-release files gained a new BUILD_ID= field for usage
  5055. on operating systems that provide continuous builds of OS
  5056. images.
  5057. Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
  5058. Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
  5059. William Douglas, Zbigniew Jędrzejewski-Szmek
  5060. CHANGES WITH 199:
  5061. * systemd-python gained an API exposing libsystemd-daemon.
  5062. * The SMACK setup logic gained support for uploading CIPSO
  5063. security policy.
  5064. * Behaviour of PrivateTmp=, ReadWriteDirectories=,
  5065. ReadOnlyDirectories= and InaccessibleDirectories= has
  5066. changed. The private /tmp and /var/tmp directories are now
  5067. shared by all processes of a service (which means
  5068. ExecStartPre= may now leave data in /tmp that ExecStart= of
  5069. the same service can still access). When a service is
  5070. stopped its temporary directories are immediately deleted
  5071. (normal clean-up with tmpfiles is still done in addition to
  5072. this though).
  5073. * By default, systemd will now set a couple of sysctl
  5074. variables in the kernel: the safe sysrq options are turned
  5075. on, IP route verification is turned on, and source routing
  5076. disabled. The recently added hardlink and softlink
  5077. protection of the kernel is turned on. These settings should
  5078. be reasonably safe, and good defaults for all new systems.
  5079. * The predictable network naming logic may now be turned off
  5080. with a new kernel command line switch: net.ifnames=0.
  5081. * A new libsystemd-bus module has been added that implements a
  5082. pretty complete D-Bus client library. For details see:
  5084. * journald will now explicitly flush the journal files to disk
  5085. at the latest 5min after each write. The file will then also
  5086. be marked offline until the next write. This should increase
  5087. reliability in case of a crash. The synchronization delay
  5088. can be configured via SyncIntervalSec= in journald.conf.
  5089. * There's a new unit that can be used
  5090. to pull in specific services when at least one remote file
  5091. system is to be mounted.
  5092. * There are new targets and as
  5093. canonical targets to pull user timer and path units in
  5094. from. This complements with a similar
  5095. purpose for socket units.
  5096. * libudev gained a new call udev_device_set_attribute_value()
  5097. to set sysfs attributes of a device.
  5098. * The udev daemon now sets the default number of worker
  5099. processes executed in parallel based on the number of available
  5100. CPUs instead of the amount of available RAM. This is supposed
  5101. to provide a more reliable default and limit a too aggressive
  5102. parallelism for setups with 1000s of devices connected.
  5103. Contributions from: Auke Kok, Colin Walters, Cristian
  5104. Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
  5105. Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
  5106. Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
  5107. Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
  5108. Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
  5109. Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
  5110. Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
  5111. Zbigniew Jędrzejewski-Szmek
  5112. CHANGES WITH 198:
  5113. * Configuration of unit files may now be extended via drop-in
  5114. files without having to edit/override the unit files
  5115. themselves. More specifically, if the administrator wants to
  5116. change one value for a service file foobar.service he can
  5117. now do so by dropping in a configuration snippet into
  5118. /etc/systemd/system/foobar.service.d/*.conf. The unit logic
  5119. will load all these snippets and apply them on top of the
  5120. main unit configuration file, possibly extending or
  5121. overriding its settings. Using these drop-in snippets is
  5122. generally nicer than the two earlier options for changing
  5123. unit files locally: copying the files from
  5124. /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
  5125. them there; or creating a new file in /etc/systemd/system/
  5126. that incorporates the original one via ".include". Drop-in
  5127. snippets into these .d/ directories can be placed in any
  5128. directory systemd looks for units in, and the usual
  5129. overriding semantics between /usr/lib, /etc and /run apply
  5130. for them too.
  5131. * Most unit file settings which take lists of items can now be
  5132. reset by assigning the empty string to them. For example,
  5133. normally, settings such as Environment=FOO=BAR append a new
  5134. environment variable assignment to the environment block,
  5135. each time they are used. By assigning Environment= the empty
  5136. string the environment block can be reset to empty. This is
  5137. particularly useful with the .d/*.conf drop-in snippets
  5138. mentioned above, since this adds the ability to reset list
  5139. settings from vendor unit files via these drop-ins.
  5140. * systemctl gained a new "list-dependencies" command for
  5141. listing the dependencies of a unit recursively.
  5142. * Inhibitors are now honored and listed by "systemctl
  5143. suspend", "systemctl poweroff" (and similar) too, not only
  5144. GNOME. These commands will also list active sessions by
  5145. other users.
  5146. * Resource limits (as exposed by the various control group
  5147. controllers) can now be controlled dynamically at runtime
  5148. for all units. More specifically, you can now use a command
  5149. like "systemctl set-cgroup-attr foobar.service cpu.shares
  5150. 2000" to alter the CPU shares a specific service gets. These
  5151. settings are stored persistently on disk, and thus allow the
  5152. administrator to easily adjust the resource usage of
  5153. services with a few simple commands. This dynamic resource
  5154. management logic is also available to other programs via the
  5155. bus. Almost any kernel cgroup attribute and controller is
  5156. supported.
  5157. * systemd-vconsole-setup will now copy all font settings to
  5158. all allocated VTs, where it previously applied them only to
  5159. the foreground VT.
  5160. * libsystemd-login gained the new sd_session_get_tty() API
  5161. call.
  5162. * This release drops support for a few legacy or
  5163. distribution-specific LSB facility names when parsing init
  5164. scripts: $x-display-manager, $mail-transfer-agent,
  5165. $mail-transport-agent, $mail-transfer-agent, $smtp,
  5166. $null. Also, the unit backing
  5167. this has been removed. Distributions which want to retain
  5168. compatibility with this should carry the burden for
  5169. supporting this themselves and patch support for these back
  5170. in, if they really need to. Also, the facilities $syslog and
  5171. $local_fs are now ignored, since systemd does not support
  5172. early-boot LSB init scripts anymore, and these facilities
  5173. are implied anyway for normal services. has
  5174. also been removed.
  5175. * There are new bus calls on PID1's Manager object for
  5176. cancelling jobs, and removing snapshot units. Previously,
  5177. both calls were only available on the Job and Snapshot
  5178. objects themselves.
  5179. * systemd-journal-gatewayd gained SSL support.
  5180. * The various "environment" files, such as /etc/locale.conf
  5181. now support continuation lines with a backslash ("\") as
  5182. last character in the line, similarly in style (but different)
  5183. to how this is supported in shells.
  5184. * For normal user processes the _SYSTEMD_USER_UNIT= field is
  5185. now implicitly appended to every log entry logged. systemctl
  5186. has been updated to filter by this field when operating on a
  5187. user systemd instance.
  5188. * nspawn will now implicitly add the CAP_AUDIT_WRITE and
  5189. CAP_AUDIT_CONTROL capabilities to the capabilities set for
  5190. the container. This makes it easier to boot unmodified
  5191. Fedora systems in a container, which however still requires
  5192. audit=0 to be passed on the kernel command line. Auditing in
  5193. kernel and userspace is unfortunately still too broken in
  5194. context of containers, hence we recommend compiling it out
  5195. of the kernel or using audit=0. Hopefully this will be fixed
  5196. one day for good in the kernel.
  5197. * nspawn gained the new --bind= and --bind-ro= parameters to
  5198. bind mount specific directories from the host into the
  5199. container.
  5200. * nspawn will now mount its own devpts file system instance
  5201. into the container, in order not to leak pty devices from
  5202. the host into the container.
  5203. * systemd will now read the firmware boot time performance
  5204. information from the EFI variables, if the used boot loader
  5205. supports this, and takes it into account for boot performance
  5206. analysis via "systemd-analyze". This is currently supported
  5207. only in conjunction with Gummiboot, but could be supported
  5208. by other boot loaders too. For details see:
  5210. * A new generator has been added that automatically mounts the
  5211. EFI System Partition (ESP) to /boot, if that directory
  5212. exists, is empty, and no other file system has been
  5213. configured to be mounted there.
  5214. * logind will now send out PrepareForSleep(false) out
  5215. unconditionally, after coming back from suspend. This may be
  5216. used by applications as asynchronous notification for
  5217. system resume events.
  5218. * "systemctl unlock-sessions" has been added, that allows
  5219. unlocking the screens of all user sessions at once, similar
  5220. to how "systemctl lock-sessions" already locked all users
  5221. sessions. This is backed by a new D-Bus call UnlockSessions().
  5222. * "loginctl seat-status" will now show the master device of a
  5223. seat. (i.e. the device of a seat that needs to be around for
  5224. the seat to be considered available, usually the graphics
  5225. card).
  5226. * tmpfiles gained a new "X" line type, that allows
  5227. configuration of files and directories (with wildcards) that
  5228. shall be excluded from automatic cleanup ("aging").
  5229. * udev default rules set the device node permissions now only
  5230. at "add" events, and do not change them any longer with a
  5231. later "change" event.
  5232. * The log messages for lid events and power/sleep keypresses
  5233. now carry a message ID.
  5234. * We now have a substantially larger unit test suite, but this
  5235. continues to be work in progress.
  5236. * udevadm hwdb gained a new --root= parameter to change the
  5237. root directory to operate relative to.
  5238. * logind will now issue a background sync() request to the kernel
  5239. early at shutdown, so that dirty buffers are flushed to disk early
  5240. instead of at the last moment, in order to optimize shutdown
  5241. times a little.
  5242. * A new bootctl tool has been added that is an interface for
  5243. certain boot loader operations. This is currently a preview
  5244. and is likely to be extended into a small mechanism daemon
  5245. like timedated, localed, hostnamed, and can be used by
  5246. graphical UIs to enumerate available boot options, and
  5247. request boot into firmware operations.
  5248. * systemd-bootchart has been relicensed to LGPLv2.1+ to match
  5249. the rest of the package. It also has been updated to work
  5250. correctly in initrds.
  5251. * Policykit previously has been runtime optional, and is now
  5252. also compile time optional via a configure switch.
  5253. * systemd-analyze has been reimplemented in C. Also "systemctl
  5254. dot" has moved into systemd-analyze.
  5255. * "systemctl status" with no further parameters will now print
  5256. the status of all active or failed units.
  5257. * Operations such as "systemctl start" can now be executed
  5258. with a new mode "--irreversible" which may be used to queue
  5259. operations that cannot accidentally be reversed by a later
  5260. job queuing. This is by default used to make shutdown
  5261. requests more robust.
  5262. * The Python API of systemd now gained a new module for
  5263. reading journal files.
  5264. * A new tool kernel-install has been added that can install
  5265. kernel images according to the Boot Loader Specification:
  5267. * Boot time console output has been improved to provide
  5268. animated boot time output for hanging jobs.
  5269. * A new tool systemd-activate has been added which can be used
  5270. to test socket activation with, directly from the command
  5271. line. This should make it much easier to test and debug
  5272. socket activation in daemons.
  5273. * journalctl gained a new "--reverse" (or -r) option to show
  5274. journal output in reverse order (i.e. newest line first).
  5275. * journalctl gained a new "--pager-end" (or -e) option to jump
  5276. to immediately jump to the end of the journal in the
  5277. pager. This is only supported in conjunction with "less".
  5278. * journalctl gained a new "--user-unit=" option, that works
  5279. similarly to "--unit=" but filters for user units rather than
  5280. system units.
  5281. * A number of unit files to ease adoption of systemd in
  5282. initrds has been added. This moves some minimal logic from
  5283. the various initrd implementations into systemd proper.
  5284. * The journal files are now owned by a new group
  5285. "systemd-journal", which exists specifically to allow access
  5286. to the journal, and nothing else. Previously, we used the
  5287. "adm" group for that, which however possibly covers more
  5288. than just journal/log file access. This new group is now
  5289. already used by systemd-journal-gatewayd to ensure this
  5290. daemon gets access to the journal files and as little else
  5291. as possible. Note that "make install" will also set FS ACLs
  5292. up for /var/log/journal to give "adm" and "wheel" read
  5293. access to it, in addition to "systemd-journal" which owns
  5294. the journal files. We recommend that packaging scripts also
  5295. add read access to "adm" + "wheel" to /var/log/journal, and
  5296. all existing/future journal files. To normal users and
  5297. administrators little changes, however packagers need to
  5298. ensure to create the "systemd-journal" system group at
  5299. package installation time.
  5300. * The systemd-journal-gatewayd now runs as unprivileged user
  5301. systemd-journal-gateway:systemd-journal-gateway. Packaging
  5302. scripts need to create these system user/group at
  5303. installation time.
  5304. * timedated now exposes a new boolean property CanNTP that
  5305. indicates whether a local NTP service is available or not.
  5306. * systemd-detect-virt will now also detect xen PVs
  5307. * The pstore file system is now mounted by default, if it is
  5308. available.
  5309. * In addition to the SELinux and IMA policies we will now also
  5310. load SMACK policies at early boot.
  5311. Contributions from: Adel Gadllah, Aleksander Morgado, Auke
  5312. Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
  5313. Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
  5314. Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
  5315. Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
  5316. Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
  5317. Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
  5318. Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
  5319. Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
  5320. Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
  5321. Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
  5322. Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
  5323. Gundersen, Umut Tezduyar, William Giokas, Zbigniew
  5324. Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
  5325. CHANGES WITH 197:
  5326. * Timer units now support calendar time events in addition to
  5327. monotonic time events. That means you can now trigger a unit
  5328. based on a calendar time specification such as "Thu,Fri
  5329. 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
  5330. or fifth day of any month of the year 2013, given that it is
  5331. a thursday or friday. This brings timer event support
  5332. considerably closer to cron's capabilities. For details on
  5333. the supported calendar time specification language see
  5334. systemd.time(7).
  5335. * udev now supports a number of different naming policies for
  5336. network interfaces for predictable names, and a combination
  5337. of these policies is now the default. Please see this wiki
  5338. document for details:
  5340. * Auke Kok's bootchart implementation has been added to the
  5341. systemd tree. It is an optional component that can graph the
  5342. boot in quite some detail. It is one of the best bootchart
  5343. implementations around and minimal in its code and
  5344. dependencies.
  5345. * nss-myhostname has been integrated into the systemd source
  5346. tree. nss-myhostname guarantees that the local hostname
  5347. always stays resolvable via NSS. It has been a weak
  5348. requirement of systemd-hostnamed since a long time, and
  5349. since its code is actually trivial we decided to just
  5350. include it in systemd's source tree. It can be turned off
  5351. with a configure switch.
  5352. * The read-ahead logic is now capable of properly detecting
  5353. whether a btrfs file system is on SSD or rotating media, in
  5354. order to optimize the read-ahead scheme. Previously, it was
  5355. only capable of detecting this on traditional file systems
  5356. such as ext4.
  5357. * In udev, additional device properties are now read from the
  5358. IAB in addition to the OUI database. Also, Bluetooth company
  5359. identities are attached to the devices as well.
  5360. * In service files %U may be used as specifier that is
  5361. replaced by the configured user name of the service.
  5362. * nspawn may now be invoked without a controlling TTY. This
  5363. makes it suitable for invocation as its own service. This
  5364. may be used to set up a simple containerized server system
  5365. using only core OS tools.
  5366. * systemd and nspawn can now accept socket file descriptors
  5367. when they are started for socket activation. This enables
  5368. implementation of socket activated nspawn
  5369. containers. i.e. think about autospawning an entire OS image
  5370. when the first SSH or HTTP connection is received. We expect
  5371. that similar functionality will also be added to libvirt-lxc
  5372. eventually.
  5373. * journalctl will now suppress ANSI color codes when
  5374. presenting log data.
  5375. * systemctl will no longer show control group information for
  5376. a unit if the control group is empty anyway.
  5377. * logind can now automatically suspend/hibernate/shutdown the
  5378. system on idle.
  5379. * /etc/machine-info and hostnamed now also expose the chassis
  5380. type of the system. This can be used to determine whether
  5381. the local system is a laptop, desktop, handset or
  5382. tablet. This information may either be configured by the
  5383. user/vendor or is automatically determined from ACPI and DMI
  5384. information if possible.
  5385. * A number of PolicyKit actions are now bound together with
  5386. "imply" rules. This should simplify creating UIs because
  5387. many actions will now authenticate similar ones as well.
  5388. * Unit files learnt a new condition ConditionACPower= which
  5389. may be used to conditionalize a unit depending on whether an
  5390. AC power source is connected or not, of whether the system
  5391. is running on battery power.
  5392. * systemctl gained a new "is-failed" verb that may be used in
  5393. shell scripts and suchlike to check whether a specific unit
  5394. is in the "failed" state.
  5395. * The EnvironmentFile= setting in unit files now supports file
  5396. globbing, and can hence be used to easily read a number of
  5397. environment files at once.
  5398. * systemd will no longer detect and recognize specific
  5399. distributions. All distribution-specific #ifdeffery has been
  5400. removed, systemd is now fully generic and
  5401. distribution-agnostic. Effectively, not too much is lost as
  5402. a lot of the code is still accessible via explicit configure
  5403. switches. However, support for some distribution specific
  5404. legacy configuration file formats has been dropped. We
  5405. recommend distributions to simply adopt the configuration
  5406. files everybody else uses now and convert the old
  5407. configuration from packaging scripts. Most distributions
  5408. already did that. If that's not possible or desirable,
  5409. distributions are welcome to forward port the specific
  5410. pieces of code locally from the git history.
  5411. * When logging a message about a unit systemd will now always
  5412. log the unit name in the message meta data.
  5413. * localectl will now also discover system locale data that is
  5414. not stored in locale archives, but directly unpacked.
  5415. * logind will no longer unconditionally use framebuffer
  5416. devices as seat masters, i.e. as devices that are required
  5417. to be existing before a seat is considered preset. Instead,
  5418. it will now look for all devices that are tagged as
  5419. "seat-master" in udev. By default, framebuffer devices will
  5420. be marked as such, but depending on local systems, other
  5421. devices might be marked as well. This may be used to
  5422. integrate graphics cards using closed source drivers (such
  5423. as NVidia ones) more nicely into logind. Note however, that
  5424. we recommend using the open source NVidia drivers instead,
  5425. and no udev rules for the closed-source drivers will be
  5426. shipped from us upstream.
  5427. Contributions from: Adam Williamson, Alessandro Crismani, Auke
  5428. Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
  5429. Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
  5430. Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
  5431. Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
  5432. Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
  5433. Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
  5434. Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
  5435. Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
  5436. Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
  5437. Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
  5438. Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
  5439. Jędrzejewski-Szmek
  5440. CHANGES WITH 196:
  5441. * udev gained support for loading additional device properties
  5442. from an indexed database that is keyed by vendor/product IDs
  5443. and similar device identifiers. For the beginning this
  5444. "hwdb" is populated with data from the well-known PCI and
  5445. USB database, but also includes PNP, ACPI and OID data. In
  5446. the longer run this indexed database shall grow into
  5447. becoming the one central database for non-essential
  5448. userspace device metadata. Previously, data from the PCI/USB
  5449. database was only attached to select devices, since the
  5450. lookup was a relatively expensive operation due to O(n) time
  5451. complexity (with n being the number of entries in the
  5452. database). Since this is now O(1), we decided to add in this
  5453. data for all devices where this is available, by
  5454. default. Note that the indexed database needs to be rebuilt
  5455. when new data files are installed. To achieve this you need
  5456. to update your packaging scripts to invoke "udevadm hwdb
  5457. --update" after installation of hwdb data files. For
  5458. RPM-based distributions we introduced the new
  5459. %udev_hwdb_update macro for this purpose.
  5460. * The Journal gained support for the "Message Catalog", an
  5461. indexed database to link up additional information with
  5462. journal entries. For further details please check:
  5464. The indexed message catalog database also needs to be
  5465. rebuilt after installation of message catalog files. Use
  5466. "journalctl --update-catalog" for this. For RPM-based
  5467. distributions we introduced the %journal_catalog_update
  5468. macro for this purpose.
  5469. * The Python Journal bindings gained support for the standard
  5470. Python logging framework.
  5471. * The Journal API gained new functions for checking whether
  5472. the underlying file system of a journal file is capable of
  5473. properly reporting file change notifications, or whether
  5474. applications that want to reflect journal changes "live"
  5475. need to recheck journal files continuously in appropriate
  5476. time intervals.
  5477. * It is now possible to set the "age" field for tmpfiles
  5478. entries to 0, indicating that files matching this entry
  5479. shall always be removed when the directories are cleaned up.
  5480. * coredumpctl gained a new "gdb" verb which invokes gdb
  5481. right-away on the selected coredump.
  5482. * There's now support for "hybrid sleep" on kernels that
  5483. support this, in addition to "suspend" and "hibernate". Use
  5484. "systemctl hybrid-sleep" to make use of this.
  5485. * logind's HandleSuspendKey= setting (and related settings)
  5486. now gained support for a new "lock" setting to simply
  5487. request the screen lock on all local sessions, instead of
  5488. actually executing a suspend or hibernation.
  5489. * systemd will now mount the EFI variables file system by
  5490. default.
  5491. * Socket units now gained support for configuration of the
  5492. SMACK security label.
  5493. * timedatectl will now output the time of the last and next
  5494. daylight saving change.
  5495. * We dropped support for various legacy and distro-specific
  5496. concepts, such as insserv, early-boot SysV services
  5497. (i.e. those for non-standard runlevels such as 'b' or 'S')
  5498. or ArchLinux /etc/rc.conf support. We recommend the
  5499. distributions who still need support this to either continue
  5500. to maintain the necessary patches downstream, or find a
  5501. different solution. (Talk to us if you have questions!)
  5502. * Various systemd components will now bypass PolicyKit checks
  5503. for root and otherwise handle properly if PolicyKit is not
  5504. found to be around. This should fix most issues for
  5505. PolicyKit-less systems. Quite frankly this should have been
  5506. this way since day one. It is absolutely our intention to
  5507. make systemd work fine on PolicyKit-less systems, and we
  5508. consider it a bug if something does not work as it should if
  5509. PolicyKit is not around.
  5510. * For embedded systems it is now possible to build udev and
  5511. systemd without blkid and/or kmod support.
  5512. * "systemctl switch-root" is now capable of switching root
  5513. more than once. I.e. in addition to transitions from the
  5514. initrd to the host OS it is now possible to transition to
  5515. further OS images from the host. This is useful to implement
  5516. offline updating tools.
  5517. * Various other additions have been made to the RPM macros
  5518. shipped with systemd. Use %udev_rules_update() after
  5519. installing new udev rules files. %_udevhwdbdir,
  5520. %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
  5521. %_sysctldir are now available which resolve to the right
  5522. directories for packages to place various data files in.
  5523. * journalctl gained the new --full switch (in addition to
  5524. --all, to disable ellipsation for long messages.
  5525. Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
  5526. Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
  5527. Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
  5528. Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
  5529. Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
  5530. Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
  5531. Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
  5532. Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
  5533. Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
  5534. CHANGES WITH 195:
  5535. * journalctl gained new --since= and --until= switches to
  5536. filter by time. It also now supports nice filtering for
  5537. units via --unit=/-u.
  5538. * Type=oneshot services may use ExecReload= and do the
  5539. right thing.
  5540. * The journal daemon now supports time-based rotation and
  5541. vacuuming, in addition to the usual disk-space based
  5542. rotation.
  5543. * The journal will now index the available field values for
  5544. each field name. This enables clients to show pretty drop
  5545. downs of available match values when filtering. The bash
  5546. completion of journalctl has been updated
  5547. accordingly. journalctl gained a new switch -F to list all
  5548. values a certain field takes in the journal database.
  5549. * More service events are now written as structured messages
  5550. to the journal, and made recognizable via message IDs.
  5551. * The timedated, localed and hostnamed mini-services which
  5552. previously only provided support for changing time, locale
  5553. and hostname settings from graphical DEs such as GNOME now
  5554. also have a minimal (but very useful) text-based client
  5555. utility each. This is probably the nicest way to changing
  5556. these settings from the command line now, especially since
  5557. it lists available options and is fully integrated with bash
  5558. completion.
  5559. * There's now a new tool "systemd-coredumpctl" to list and
  5560. extract coredumps from the journal.
  5561. * We now install a README each in /var/log/ and
  5562. /etc/rc.d/init.d explaining where the system logs and init
  5563. scripts went. This hopefully should help folks who go to
  5564. that dirs and look into the otherwise now empty void and
  5565. scratch their heads.
  5566. * When user-services are invoked (by systemd --user) the
  5567. $MANAGERPID env var is set to the PID of systemd.
  5568. * SIGRTMIN+24 when sent to a --user instance will now result
  5569. in immediate termination of systemd.
  5570. * gatewayd received numerous feature additions such as a
  5571. "follow" mode, for live syncing and filtering.
  5572. * browse.html now allows filtering and showing detailed
  5573. information on specific entries. Keyboard navigation and
  5574. mouse screen support has been added.
  5575. * gatewayd/journalctl now supports HTML5/JSON
  5576. Server-Sent-Events as output.
  5577. * The SysV init script compatibility logic will now
  5578. heuristically determine whether a script supports the
  5579. "reload" verb, and only then make this available as
  5580. "systemctl reload".
  5581. * "systemctl status --follow" has been removed, use "journalctl
  5582. -u" instead.
  5583. * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
  5584. have been removed since they are hardly useful to be
  5585. configured.
  5586. * And I'd like to take the opportunity to specifically mention
  5587. Zbigniew for his great contributions. Zbigniew, you rock!
  5588. Contributions from: Andrew Eikum, Christian Hesse, Colin
  5589. Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
  5590. Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
  5591. Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
  5592. Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
  5593. Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
  5594. Jędrzejewski-Szmek, Сковорода Никита Андреевич
  5595. CHANGES WITH 194:
  5596. * If /etc/vconsole.conf is non-existent or empty we will no
  5597. longer load any console font or key map at boot by
  5598. default. Instead the kernel defaults will be left
  5599. intact. This is definitely the right thing to do, as no
  5600. configuration should mean no configuration, and hard-coding
  5601. font names that are different on all archs is probably a bad
  5602. idea. Also, the kernel default key map and font should be
  5603. good enough for most cases anyway, and mostly identical to
  5604. the userspace fonts/key maps we previously overloaded them
  5605. with. If distributions want to continue to default to a
  5606. non-kernel font or key map they should ship a default
  5607. /etc/vconsole.conf with the appropriate contents.
  5608. Contributions from: Colin Walters, Daniel J Walsh, Dave
  5609. Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
  5610. Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
  5611. CHANGES WITH 193:
  5612. * journalctl gained a new --cursor= switch to show entries
  5613. starting from the specified location in the journal.
  5614. * We now enforce a size limit on journal entry fields exported
  5615. with "-o json" in journalctl. Fields larger than 4K will be
  5616. assigned null. This can be turned off with --all.
  5617. * An (optional) journal gateway daemon is now available as
  5618. "systemd-journal-gatewayd.service". This service provides
  5619. access to the journal via HTTP and JSON. This functionality
  5620. will be used to implement live log synchronization in both
  5621. pull and push modes, but has various other users too, such
  5622. as easy log access for debugging of embedded devices. Right
  5623. now it is already useful to retrieve the journal via HTTP:
  5624. # systemctl start systemd-journal-gatewayd.service
  5625. # wget http://localhost:19531/entries
  5626. This will download the journal contents in a
  5627. /var/log/messages compatible format. The same as JSON:
  5628. # curl -H"Accept: application/json" http://localhost:19531/entries
  5629. This service is also accessible via a web browser where a
  5630. single static HTML5 app is served that uses the JSON logic
  5631. to enable the user to do some basic browsing of the
  5632. journal. This will be extended later on. Here's an example
  5633. screenshot of this app in its current state:
  5635. Contributions from: Kay Sievers, Lennart Poettering, Robert
  5636. Milasan, Tom Gundersen
  5637. CHANGES WITH 192:
  5638. * The bash completion logic is now available for journalctl
  5639. too.
  5640. * We do not mount the "cpuset" controller anymore together with
  5641. "cpu" and "cpuacct", as "cpuset" groups generally cannot be
  5642. started if no parameters are assigned to it. "cpuset" hence
  5643. broke code that assumed it could create "cpu" groups and
  5644. just start them.
  5645. * journalctl -f will now subscribe to terminal size changes,
  5646. and line break accordingly.
  5647. Contributions from: Dave Reisner, Kay Sievers, Lennart
  5648. Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
  5649. CHANGES WITH 191:
  5650. * nspawn will now create a symlink /etc/localtime in the
  5651. container environment, copying the host's timezone
  5652. setting. Previously this has been done via a bind mount, but
  5653. since symlinks cannot be bind mounted this has now been
  5654. changed to create/update the appropriate symlink.
  5655. * journalctl -n's line number argument is now optional, and
  5656. will default to 10 if omitted.
  5657. * journald will now log the maximum size the journal files may
  5658. take up on disk. This is particularly useful if the default
  5659. built-in logic of determining this parameter from the file
  5660. system size is used. Use "systemctl status
  5661. systemd-journald.service" to see this information.
  5662. * The multi-seat X wrapper tool has been stripped down. As X
  5663. is now capable of enumerating graphics devices via udev in a
  5664. seat-aware way the wrapper is not strictly necessary
  5665. anymore. A stripped down temporary stop-gap is still shipped
  5666. until the upstream display managers have been updated to
  5667. fully support the new X logic. Expect this wrapper to be
  5668. removed entirely in one of the next releases.
  5669. * HandleSleepKey= in logind.conf has been split up into
  5670. HandleSuspendKey= and HandleHibernateKey=. The old setting
  5671. is not available anymore. X11 and the kernel are
  5672. distinguishing between these keys and we should too. This
  5673. also means the inhibition lock for these keys has been split
  5674. into two.
  5675. Contributions from: Dave Airlie, Eelco Dolstra, Lennart
  5676. Poettering, Lukas Nykryn, Václav Pavlín
  5677. CHANGES WITH 190:
  5678. * Whenever a unit changes state we will now log this to the
  5679. journal and show along the unit's own log output in
  5680. "systemctl status".
  5681. * ConditionPathIsMountPoint= can now properly detect bind
  5682. mount points too. (Previously, a bind mount of one file
  5683. system to another place in the same file system could not be
  5684. detected as mount, since they shared struct stat's st_dev
  5685. field.)
  5686. * We will now mount the cgroup controllers cpu, cpuacct,
  5687. cpuset and the controllers net_cls, net_prio together by
  5688. default.
  5689. * nspawn containers will now have a virtualized boot
  5690. ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
  5691. over with a randomized ID at container initialization). This
  5692. has the effect of making "journalctl -b" do the right thing
  5693. in a container.
  5694. * The JSON output journal serialization has been updated not
  5695. to generate "endless" list objects anymore, but rather one
  5696. JSON object per line. This is more in line how most JSON
  5697. parsers expect JSON objects. The new output mode
  5698. "json-pretty" has been added to provide similar output, but
  5699. neatly aligned for readability by humans.
  5700. * We dropped all explicit sync() invocations in the shutdown
  5701. code. The kernel does this implicitly anyway in the kernel
  5702. reboot() syscall. halt(8)'s -n option is now a compatibility
  5703. no-op.
  5704. * We now support virtualized reboot() in containers, as
  5705. supported by newer kernels. We will fall back to exit() if
  5706. CAP_SYS_REBOOT is not available to the container. Also,
  5707. nspawn makes use of this now and will actually reboot the
  5708. container if the containerized OS asks for that.
  5709. * journalctl will only show local log output by default
  5710. now. Use --merge (-m) to show remote log output, too.
  5711. * libsystemd-journal gained the new sd_journal_get_usage()
  5712. call to determine the current disk usage of all journal
  5713. files. This is exposed in the new "journalctl --disk-usage"
  5714. command.
  5715. * journald gained a new configuration setting SplitMode= in
  5716. journald.conf which may be used to control how user journals
  5717. are split off. See journald.conf(5) for details.
  5718. * A new condition type ConditionFileNotEmpty= has been added.
  5719. * tmpfiles' "w" lines now support file globbing, to write
  5720. multiple files at once.
  5721. * We added Python bindings for the journal submission
  5722. APIs. More Python APIs for a number of selected APIs will
  5723. likely follow. Note that we intend to add native bindings
  5724. only for the Python language, as we consider it common
  5725. enough to deserve bindings shipped within systemd. There are
  5726. various projects outside of systemd that provide bindings
  5727. for languages such as PHP or Lua.
  5728. * Many conditions will now resolve specifiers such as %i. In
  5729. addition, PathChanged= and related directives of .path units
  5730. now support specifiers as well.
  5731. * There's now a new RPM macro definition for the system preset
  5732. dir: %_presetdir.
  5733. * journald will now warn if it ca not forward a message to the
  5734. syslog daemon because its socket is full.
  5735. * timedated will no longer write or process /etc/timezone,
  5736. except on Debian. As we do not support late mounted /usr
  5737. anymore /etc/localtime always being a symlink is now safe,
  5738. and hence the information in /etc/timezone is not necessary
  5739. anymore.
  5740. * logind will now always reserve one VT for a text getty (VT6
  5741. by default). Previously if more than 6 X sessions where
  5742. started they took up all the VTs with auto-spawned gettys,
  5743. so that no text gettys were available anymore.
  5744. * udev will now automatically inform the btrfs kernel logic
  5745. about btrfs RAID components showing up. This should make
  5746. simple hotplug based btrfs RAID assembly work.
  5747. * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
  5748. (but not for its children which will stay at the kernel
  5749. default). This should allow setups with a lot more listening
  5750. sockets.
  5751. * systemd will now always pass the configured timezone to the
  5752. kernel at boot. timedated will do the same when the timezone
  5753. is changed.
  5754. * logind's inhibition logic has been updated. By default,
  5755. logind will now handle the lid switch, the power and sleep
  5756. keys all the time, even in graphical sessions. If DEs want
  5757. to handle these events on their own they should take the new
  5758. handle-power-key, handle-sleep-key and handle-lid-switch
  5759. inhibitors during their runtime. A simple way to achieve
  5760. that is to invoke the DE wrapped in an invocation of:
  5761. systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
  5762. * Access to unit operations is now checked via SELinux taking
  5763. the unit file label and client process label into account.
  5764. * systemd will now notify the administrator in the journal
  5765. when he over-mounts a non-empty directory.
  5766. * There are new specifiers that are resolved in unit files,
  5767. for the host name (%H), the machine ID (%m) and the boot ID
  5768. (%b).
  5769. Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
  5770. Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
  5771. Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
  5772. Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
  5773. Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
  5774. Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
  5775. Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
  5776. CHANGES WITH 189:
  5777. * Support for reading structured kernel messages from
  5778. /dev/kmsg has now been added and is enabled by default.
  5779. * Support for reading kernel messages from /proc/kmsg has now
  5780. been removed. If you want kernel messages in the journal
  5781. make sure to run a recent kernel (>= 3.5) that supports
  5782. reading structured messages from /dev/kmsg (see
  5783. above). /proc/kmsg is now exclusive property of classic
  5784. syslog daemons again.
  5785. * The libudev API gained the new
  5786. udev_device_new_from_device_id() call.
  5787. * The logic for file system namespace (ReadOnlyDirectory=,
  5788. ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
  5789. require pivot_root() anymore. This means fewer temporary
  5790. directories are created below /tmp for this feature.
  5791. * nspawn containers will now see and receive all submounts
  5792. made on the host OS below the root file system of the
  5793. container.
  5794. * Forward Secure Sealing is now supported for Journal files,
  5795. which provide cryptographical sealing of journal files so
  5796. that attackers cannot alter log history anymore without this
  5797. being detectable. Lennart will soon post a blog story about
  5798. this explaining it in more detail.
  5799. * There are two new service settings RestartPreventExitStatus=
  5800. and SuccessExitStatus= which allow configuration of exit
  5801. status (exit code or signal) which will be excepted from the
  5802. restart logic, resp. consider successful.
  5803. * journalctl gained the new --verify switch that can be used
  5804. to check the integrity of the structure of journal files and
  5805. (if Forward Secure Sealing is enabled) the contents of
  5806. journal files.
  5807. * nspawn containers will now be run with /dev/stdin, /dev/fd/
  5808. and similar symlinks pre-created. This makes running shells
  5809. as container init process a lot more fun.
  5810. * The fstab support can now handle PARTUUID= and PARTLABEL=
  5811. entries.
  5812. * A new ConditionHost= condition has been added to match
  5813. against the hostname (with globs) and machine ID. This is
  5814. useful for clusters where a single OS image is used to
  5815. provision a large number of hosts which shall run slightly
  5816. different sets of services.
  5817. * Services which hit the restart limit will now be placed in a
  5818. failure state.
  5819. Contributions from: Bertram Poettering, Dave Reisner, Huang
  5820. Han