1 systemd System and Service Manager
5 * journalctl gained the new options -t/--identifier= to match
6 on the syslog identifier (aka "tag"), as well as --utc to
7 show log timestamps in the UTC timezone. journalctl now also
8 accepts -n/--lines=all to disable line capping in a pager.
10 * Services can notify the manager before they start a reload
11 (by sending RELOADING=1) or shutdown (by sending
12 STOPPING=1). This allows the manager to track and show the
13 internal state of daemons and closes a race condition when
14 the process is still running but has closed its D-Bus
17 * Services with Type=oneshot do not have to have any ExecStart
20 * User units are now loaded also from
21 $XDG_RUNTIME_DIR/systemd/user/. This is similar to the
22 /run/systemd/user directory that was already previously
23 supported, but is under the control of the user.
25 * Job timeouts (i.e. time-outs on the time a job that is
26 queued stays in the run queue) can now optionally result in
27 immediate reboot or power-off actions (JobTimeoutAction= and
28 JobTimeoutRebootArgument=). This is useful on ".target"
29 units, to limit the maximum time a target remains
30 undispatched in the run queue, and to trigger an emergency
31 operation in such a case. This is now used by default to
32 turn off the system if boot-up (as defined by everything in
33 basic.target) hangs and does not complete for at least
34 15min. Also, if power-off or reboot hang for at least 30min
35 an immediate power-off/reboot operation is triggered. This
36 functionality is particularly useful to increase reliability
37 on embedded devices, but also on laptops which might
38 accidentally get powered on when carried in a backpack and
39 whose boot stays stuck in a hard disk encryption passphrase
42 * systemd-logind can be configured to also handle lid switch
43 events even when the machine is docked or multiple displays
44 are attached (HandleLidSwitchDocked= option).
46 * A helper binary and a service have been added which can be
47 used to resume from hibernation in the initramfs. A
48 generator will parse the resume= option on the kernel
49 command-line to trigger resume.
51 * A user console daemon systemd-consoled has been
52 added. Currently, it is a preview, and will so far open a
53 single terminal on each session of the user marked as
54 Desktop=systemd-console.
56 * Route metrics can be specified for DHCP routes added by
59 * The SELinux context of socket-activated services can be set
60 from the information provided by the networking stack
61 (SELinuxContextFromNet= option).
63 * Userspace firmware loading support has been removed and
64 the minimum supported kernel version is thus bumped to 3.7.
66 * Timeout for udev workers has been increased from 1 to 3
67 minutes, but a warning will be printed after 1 minute to
68 help diagnose kernel modules that take a long time to load.
70 * Udev rules can now remove tags on devices with TAG-="foobar".
72 * systemd's readahead implementation has been removed. In many
73 circumstances it didn't give expected benefits even for
74 rotational disk drives and was becoming less relevant in the
75 age of SSDs. As none of the developers has been using
76 rotating media anymore, and nobody stepped up to actively
77 maintain this component of systemd it has now been removed.
79 * Swap units can use Discard= to specify discard options.
80 Discard options specified for swaps in /etc/fstab are now
83 * Docker containers are now detected as a separate type of
86 * The Password Agent protocol gained support for queries where
87 the user input is shown, useful e.g. for user names.
88 systemd-ask-password gained a new --echo option to turn that
91 * The default sysctl.d/ snippets will now set:
93 net.core.default_qdisc = fq_codel
95 This selects Fair Queuing Controlled Delay as the default
96 queuing discipline for network interfaces. fq_codel helps
97 fight the network bufferbloat problem. It is believed to be
98 a good default with no tuning required for most workloads.
99 Downstream distributions may override this choice. On 10Gbit
100 servers that do not do forwarding, "fq" may perform better.
101 Systems without a good clocksource should use "pfifo_fast".
103 * If kdbus is enabled during build a new option BusPolicy= is
104 available for service units, that allows locking all service
105 processes into a stricter bus policy, in order to limit
106 access to various bus services, or even hide most of them
107 from the service's view entirely.
109 * networkctl will now show the .network and .link file
110 networkd has applied to a specific interface.
112 * sd-login gained a new API call sd_session_get_desktop() to
113 query which desktop environment has been selected for a
116 * UNIX utmp support is now compile-time optional to support
119 * systemctl gained two new commands "add-wants" and
120 "add-requires" for pulling in units from specific targets
123 * If the word "rescue" is specified on the kernel command line
124 the system will now boot into rescue mode (aka
125 rescue.target), which was previously available only by
126 specifying "1" or "systemd.unit=rescue.target" on the kernel
127 command line. This new kernel command line option nicely
128 mirrors the already existing "emergency" kernel command line
131 * New kernel command line options mount.usr=, mount.usrflags=,
132 mount.usrfstype= have been added that match root=, rootflags=,
133 rootfstype= but allow mounting a specific file system to
136 * The $NOTIFY_SOCKET is now also passed to control processes of
137 services, not only the main process.
139 * This version reenables support for fsck's -l switch. This
140 means at least version v2.25 of util-linux is required for
141 operation, otherwise dead-locks on device nodes may
142 occur. Again: you need to update util-linux to at least
143 v2.25 when updating systemd to v217.
145 * The "multi-seat-x" tool has been removed from systemd, as
146 its functionality has been integrated into X servers 1.16,
147 and the tool is hence redundant. It is recommended to update
148 display managers invoking this tool to simply invoke X
149 directly from now on, again.
151 * Support for the new ALLOW_INTERACTIVE_AUTHORIZATION D-Bus
152 message flag has been added for all of systemd's PolicyKit
153 authenticated method calls has been added. In particular
154 this now allows optional interactive authorization via
155 PolicyKit for many of PID1's privileged operations such as
156 unit file enabling and disabling.
158 * "udevadm hwdb --update" learnt a new switch "--usr" for
159 placing the rebuilt hardware database in /usr instead of
160 /etc. When used only hardware database entries stored in
161 /usr will be used, and any user database entries in /etc are
162 ignored. This functionality is useful for vendors to ship a
163 pre-built database on systems where local configuration is
164 unnecessary or unlikely.
166 * Calendar time specifications in .timer units now also
167 understand the strings "semi-annually", "quarterly" and
168 "minutely" as shortcuts (in addition to the preexisting
169 "anually", "hourly", ...).
171 * systemd-tmpfiles will now correctly create files in /dev
172 at boot which are marked for creation only at boot. It is
173 recommended to always create static device nodes with 'c!'
174 and 'b!', so that they are created only at boot and not
175 overwritten at runtime.
177 Contributions from: Andreas Henriksson, Andrei Borzenkov,
178 Angus Gibson, Ansgar Burchardt, Ben Wolsieffer, Brandon L.
179 Black, Christian Hesse, Cristian Rodríguez, Daniel Buch,
180 Daniele Medri, Daniel Mack, Dan Williams, Dave Reisner, David
181 Herrmann, David Sommerseth, David Strauss, Emil Renner
182 Berthing, Eric Cook, Evangelos Foutras, Filipe Brandenburger,
183 Gustavo Sverzut Barbieri, Hans de Goede, Harald Hoyer, Hristo
184 Venev, Hugo Grostabussiat, Ivan Shapovalov, Jan Janssen, Jan
185 Synacek, Jonathan Liu, Juho Son, Karel Zak, Kay Sievers, Klaus
186 Purer, Koen Kooi, Lennart Poettering, Lukas Nykryn, Lukasz
187 Skalski, Łukasz Stelmach, Mantas Mikulėnas, Marcel Holtmann,
188 Marius Tessmann, Marko Myllynen, Martin Pitt, Michael Biebl,
189 Michael Marineau, Michael Olbrich, Michael Scherer, Michal
190 Schmidt, Michal Sekletar, Miroslav Lichvar, Patrik Flykt,
191 Philippe De Swert, Piotr Drąg, Rahul Sundaram, Richard
192 Weinberger, Robert Milasan, Ronny Chevalier, Ruben Kerkhof,
193 Santiago Vila, Sergey Ptashnick, Simon McVittie, Sjoerd
194 Simons, Stefan Brüns, Steven Allen, Steven Noonan, Susant
195 Sahani, Sylvain Plantefève, Thomas Hindoe Paaboel Andersen,
196 Timofey Titovets, Tobias Hunger, Tom Gundersen, Torstein
197 Husebø, Umut Tezduyar Lindskog, WaLyong Cho, Zbigniew
200 -- Berlin, 2014-10-28
204 * timedated no longer reads NTP implementation unit names from
205 /usr/lib/systemd/ntp-units.d/*.list. Alternative NTP
206 implementations should add a
208 Conflicts=systemd-timesyncd.service
210 to their unit files to take over and replace systemd's NTP
211 default functionality.
213 * systemd-sysusers gained a new line type "r" for configuring
214 which UID/GID ranges to allocate system users/groups
215 from. Lines of type "u" may now add an additional column
216 that specifies the home directory for the system user to be
217 created. Also, systemd-sysusers may now optionally read user
218 information from STDIN instead of a file. This is useful for
219 invoking it from RPM preinst scriptlets that need to create
220 users before the first RPM file is installed since these
221 files might need to be owned by them. A new
222 %sysusers_create_inline RPM macro has been introduced to do
223 just that. systemd-sysusers now updates the shadow files as
224 well as the user/group databases, which should enhance
225 compatibility with certain tools like grpck.
227 * A number of bus APIs of PID 1 now optionally consult
228 PolicyKit to permit access for otherwise unprivileged
229 clients under certain conditions. Note that this currently
230 doesn't support interactive authentication yet, but this is
231 expected to be added eventually, too.
233 * /etc/machine-info now has new fields for configuring the
234 deployment environment of the machine, as well as the
235 location of the machine. hostnamectl has been updated with
236 new command to update these fields.
238 * systemd-timesyncd has been updated to automatically acquire
239 NTP server information from systemd-networkd, which might
240 have been discovered via DHCP.
242 * systemd-resolved now includes a caching DNS stub resolver
243 and a complete LLMNR name resolution implementation. A new
244 NSS module "nss-resolve" has been added which can be used
245 instead of glibc's own "nss-dns" to resolve hostnames via
246 systemd-resolved. Hostnames, addresses and arbitrary RRs may
247 be resolved via systemd-resolved D-Bus APIs. In contrast to
248 the glibc internal resolver systemd-resolved is aware of
249 multi-homed system, and keeps DNS server and caches separate
250 and per-interface. Queries are sent simultaneously on all
251 interfaces that have DNS servers configured, in order to
252 properly handle VPNs and local LANs which might resolve
253 separate sets of domain names. systemd-resolved may acquire
254 DNS server information from systemd-networkd automatically,
255 which in turn might have discovered them via DHCP. A tool
256 "systemd-resolve-host" has been added that may be used to
257 query the DNS logic in resolved. systemd-resolved implements
258 IDNA and automatically uses IDNA or UTF-8 encoding depending
259 on whether classic DNS or LLMNR is used as transport. In the
260 next releases we intend to add a DNSSEC and mDNS/DNS-SD
261 implementation to systemd-resolved.
263 * A new NSS module nss-mymachines has been added, that
264 automatically resolves the names of all local registered
265 containers to their respective IP addresses.
267 * A new client tool "networkctl" for systemd-networkd has been
268 added. It currently is entirely passive and will query
269 networking configuration from udev, rtnetlink and networkd,
270 and present it to the user in a very friendly
271 way. Eventually, we hope to extend it to become a full
272 control utility for networkd.
274 * .socket units gained a new DeferAcceptSec= setting that
275 controls the kernels' TCP_DEFER_ACCEPT sockopt for
276 TCP. Similar, support for controlling TCP keep-alive
277 settings has been added (KeepAliveTimeSec=,
278 KeepAliveIntervalSec=, KeepAliveProbes=). Also, support for
279 turning off Nagle's algorithm on TCP has been added
282 * logind learned a new session type "web", for use in projects
283 like Cockpit which register web clients as PAM sessions.
285 * timer units with at least one OnCalendar= setting will now
286 be started only after timer-sync.target has been
287 reached. This way they will not elapse before the system
288 clock has been corrected by a local NTP client or
289 similar. This is particular useful on RTC-less embedded
290 machines, that come up with an invalid system clock.
292 * systemd-nspawn's --network-veth= switch should now result in
293 stable MAC addresses for both the outer and the inner side
296 * systemd-nspawn gained a new --volatile= switch for running
297 container instances with /etc or /var unpopulated.
299 * The kdbus client code has been updated to use the new Linux
300 3.17 memfd subsystem instead of the old kdbus-specific one.
302 * systemd-networkd's DHCP client and server now support
303 FORCERENEW. There are also new configuration options to
304 configure the vendor client identifier and broadcast mode
307 * systemd will no longer inform the kernel about the current
308 timezone, as this is necessarily incorrect and racy as the
309 kernel has no understanding of DST and similar
310 concepts. This hence means FAT timestamps will be always
311 considered UTC, similar to what Android is already
312 doing. Also, when the RTC is configured to the local time
313 (rather than UTC) systemd will never synchronize back to it,
314 as this might confuse Windows at a later boot.
316 * systemd-analyze gained a new command "verify" for offline
317 validation of unit files.
319 * systemd-networkd gained support for a couple of additional
320 settings for bonding networking setups. Also, the metric for
321 statically configured routes may now be configured. For
322 network interfaces where this is appropriate the peer IP
323 address may now be configured.
325 * systemd-networkd's DHCP client will no longer request
326 broadcasting by default, as this tripped up some networks.
327 For hardware where broadcast is required the feature should
328 be switched back on using RequestBroadcast=yes.
330 * systemd-networkd will now set up IPv4LL addresses (when
331 enabled) even if DHCP is configured successfully.
333 * udev will now default to respect network device names given
334 by the kernel when the kernel indicates that these are
335 predictable. This behavior can be tweaked by changing
336 NamePolicy= in the relevant .link file.
338 * A new library systemd-terminal has been added that
339 implements full TTY stream parsing and rendering. This
340 library is supposed to be used later on for implementing a
341 full userspace VT subsystem, replacing the current kernel
344 * A new tool systemd-journal-upload has been added to push
345 journal data to a remote system running
346 systemd-journal-remote.
348 * journald will no longer forward all local data to another
349 running syslog daemon. This change has been made because
350 rsyslog (which appears to be the most commonly used syslog
351 implementation these days) no longer makes use of this, and
352 instead pulls the data out of the journal on its own. Since
353 forwarding the messages to a non-existent syslog server is
354 more expensive than we assumed we have now turned this
355 off. If you run a syslog server that is not a recent rsyslog
356 version, you have to turn this option on again
357 (ForwardToSyslog= in journald.conf).
359 * journald now optionally supports the LZ4 compressor for
360 larger journal fields. This compressor should perform much
361 better than XZ which was the previous default.
363 * machinectl now shows the IP addresses of local containers,
364 if it knows them, plus the interface name of the container.
366 * A new tool "systemd-escape" has been added that makes it
367 easy to escape strings to build unit names and similar.
369 * sd_notify() messages may now include a new ERRNO= field
370 which is parsed and collected by systemd and shown among the
371 "systemctl status" output for a service.
373 * A new component "systemd-firstboot" has been added that
374 queries the most basic systemd information (timezone,
375 hostname, root password) interactively on first
376 boot. Alternatively it may also be used to provision these
377 things offline on OS images installed into directories.
379 * The default sysctl.d/ snippets will now set
381 net.ipv4.conf.default.promote_secondaries=1
383 This has the benefit of no flushing secondary IP addresses
384 when primary addresses are removed.
386 Contributions from: Ansgar Burchardt, Bastien Nocera, Colin
387 Walters, Dan Dedrick, Daniel Buch, Daniel Korostil, Daniel
388 Mack, Dan Williams, Dave Reisner, David Herrmann, Denis
389 Kenzior, Eelco Dolstra, Eric Cook, Hannes Reinecke, Harald
390 Hoyer, Hong Shick Pak, Hui Wang, Jean-André Santoni, Jóhann
391 B. Guðmundsson, Jon Severinsson, Karel Zak, Kay Sievers, Kevin
392 Wells, Lennart Poettering, Lukas Nykryn, Mantas Mikulėnas,
393 Marc-Antoine Perennou, Martin Pitt, Michael Biebl, Michael
394 Marineau, Michael Olbrich, Michal Schmidt, Michal Sekletar,
395 Miguel Angel Ajo, Mike Gilbert, Olivier Brunel, Robert
396 Schiele, Ronny Chevalier, Simon McVittie, Sjoerd Simons, Stef
397 Walter, Steven Noonan, Susant Sahani, Tanu Kaskinen, Thomas
398 Blume, Thomas Hindoe Paaboel Andersen, Timofey Titovets,
399 Tobias Geerinckx-Rice, Tomasz Torcz, Tom Gundersen, Umut
400 Tezduyar Lindskog, Zbigniew Jędrzejewski-Szmek
402 -- Berlin, 2014-08-19
406 * A new tool systemd-sysusers has been added. This tool
407 creates system users and groups in /etc/passwd and
408 /etc/group, based on static declarative system user/group
409 definitions in /usr/lib/sysusers.d/. This is useful to
410 enable factory resets and volatile systems that boot up with
411 an empty /etc directory, and thus need system users and
412 groups created during early boot. systemd now also ships
413 with two default sysusers.d/ files for the most basic
414 users and groups systemd and the core operating system
417 * A new tmpfiles snippet has been added that rebuilds the
418 essential files in /etc on boot, should they be missing.
420 * A directive for ensuring automatic clean-up of
421 /var/cache/man/ has been removed from the default
422 configuration. This line should now be shipped by the man
423 implementation. The necessary change has been made to the
424 man-db implementation. Note that you need to update your man
425 implementation to one that ships this line, otherwise no
426 automatic clean-up of /var/cache/man will take place.
428 * A new condition ConditionNeedsUpdate= has been added that
429 may conditionalize services to only run when /etc or /var
430 are "older" than the vendor operating system resources in
431 /usr. This is useful for reconstructing or updating /etc
432 after an offline update of /usr or a factory reset, on the
433 next reboot. Services that want to run once after such an
434 update or reset should use this condition and order
435 themselves before the new systemd-update-done.service, which
436 will mark the two directories as fully updated. A number of
437 service files have been added making use of this, to rebuild
438 the udev hardware database, the journald message catalog and
439 dynamic loader cache (ldconfig). The systemd-sysusers tool
440 described above also makes use of this now. With this in
441 place it is now possible to start up a minimal operating
442 system with /etc empty cleanly. For more information on the
443 concepts involved see this recent blog story:
445 http://0pointer.de/blog/projects/stateless.html
447 * A new system group "input" has been introduced, and all
448 input device nodes get this group assigned. This is useful
449 for system-level software to get access to input devices. It
450 complements what is already done for "audio" and "video".
452 * systemd-networkd learnt minimal DHCPv4 server support in
453 addition to the existing DHCPv4 client support. It also
454 learnt DHCPv6 client and IPv6 Router Solicitation client
455 support. The DHCPv4 client gained support for static routes
456 passed in from the server. Note that the [DHCPv4] section
457 known in older systemd-networkd versions has been renamed to
458 [DHCP] and is now also used by the DHCPv6 client. Existing
459 .network files using settings of this section should be
460 updated, though compatibility is maintained. Optionally, the
461 client hostname may now be sent to the DHCP server.
463 * networkd gained support for vxlan virtual networks as well
464 as tun/tap and dummy devices.
466 * networkd gained support for automatic allocation of address
467 ranges for interfaces from a system-wide pool of
468 addresses. This is useful for dynamically managing a large
469 number of interfaces with a single network configuration
470 file. In particular this is useful to easily assign
471 appropriate IP addresses to the veth links of a large number
474 * RPM macros for processing sysusers, sysctl and binfmt
475 drop-in snippets at package installation time have been
478 * The /etc/os-release file should now be placed in
479 /usr/lib/os-release. The old location is automatically
480 created as symlink. /usr/lib is the more appropriate
481 location of this file, since it shall actually describe the
482 vendor operating system shipped in /usr, and not the
483 configuration stored in /etc.
485 * .mount units gained a new boolean SloppyOptions= setting
486 that maps to mount(8)'s -s option which enables permissive
487 parsing of unknown mount options.
489 * tmpfiles learnt a new "L+" directive which creates a symlink
490 but (unlike "L") deletes a pre-existing file first, should
491 it already exist and not already be the correct
492 symlink. Similar, "b+", "c+" and "p+" directives have been
493 added as well, which create block and character devices, as
494 well as fifos in the filesystem, possibly removing any
495 pre-existing files of different types.
497 * For tmpfiles' "L", "L+", "C" and "C+" directives the final
498 'argument' field (which so far specified the source to
499 symlink/copy the files from) is now optional. If omitted the
500 same file os copied from /usr/share/factory/ suffixed by the
501 full destination path. This is useful for populating /etc
502 with essential files, by copying them from vendor defaults
503 shipped in /usr/share/factory/etc.
505 * A new command "systemctl preset-all" has been added that
506 applies the service preset settings to all installed unit
507 files. A new switch --preset-mode= has been added that
508 controls whether only enable or only disable operations
511 * A new command "systemctl is-system-running" has been added
512 that allows checking the overall state of the system, for
513 example whether it is fully up and running.
515 * When the system boots up with an empty /etc, the equivalent
516 to "systemctl preset-all" is executed during early boot, to
517 make sure all default services are enabled after a factory
520 * systemd now contains a minimal preset file that enables the
521 most basic services systemd ships by default.
523 * Unit files' [Install] section gained a new DefaultInstance=
524 field for defining the default instance to create if a
525 template unit is enabled with no instance specified.
527 * A new passive target cryptsetup-pre.target has been added
528 that may be used by services that need to make they run and
529 finish before the first LUKS cryptographic device is set up.
531 * The /dev/loop-control and /dev/btrfs-control device nodes
532 are now owned by the "disk" group by default, opening up
533 access to this group.
535 * systemd-coredump will now automatically generate a
536 stack trace of all core dumps taking place on the system,
537 based on elfutils' libdw library. This stack trace is logged
540 * systemd-coredump may now optionally store coredumps directly
541 on disk (in /var/lib/systemd/coredump, possibly compressed),
542 instead of storing them unconditionally in the journal. This
543 mode is the new default. A new configuration file
544 /etc/systemd/coredump.conf has been added to configure this
545 and other parameters of systemd-coredump.
547 * coredumpctl gained a new "info" verb to show details about a
548 specific coredump. A new switch "-1" has also been added
549 that makes sure to only show information about the most
550 recent entry instead of all entries. Also, as the tool is
551 generally useful now the "systemd-" prefix of the binary
552 name has been removed. Distributions that want to maintain
553 compatibility with the old name should add a symlink from
554 the old name to the new name.
556 * journald's SplitMode= now defaults to "uid". This makes sure
557 that unprivileged users can access their own coredumps with
558 coredumpctl without restrictions.
560 * New kernel command line options "systemd.wants=" (for
561 pulling an additional unit during boot), "systemd.mask="
562 (for masking a specific unit for the boot), and
563 "systemd.debug-shell" (for enabling the debug shell on tty9)
564 have been added. This is implemented in the new generator
565 "systemd-debug-generator".
567 * systemd-nspawn will now by default filter a couple of
568 syscalls for containers, among them those required for
569 kernel module loading, direct x86 IO port access, swap
570 management, and kexec. Most importantly though
571 open_by_handle_at() is now prohibited for containers,
572 closing a hole similar to a recently discussed vulnerability
573 in docker regarding access to files on file hierarchies the
574 container should normally not have access to. Note that for
575 nspawn we generally make no security claims anyway (and
576 this is explicitly documented in the man page), so this is
577 just a fix for one of the most obvious problems.
579 * A new man page file-hierarchy(7) has been added that
580 contains a minimized, modernized version of the file system
581 layout systemd expects, similar in style to the FHS
582 specification or hier(5). A new tool systemd-path(1) has
583 been added to query many of these paths for the local
586 * Automatic time-based clean-up of $XDG_RUNTIME_DIR is no
587 longer done. Since the directory now has a per-user size
588 limit, and is cleaned on logout this appears unnecessary,
589 in particular since this now brings the lifecycle of this
590 directory closer in line with how IPC objects are handled.
592 * systemd.pc now exports a number of additional directories,
593 including $libdir (which is useful to identify the library
594 path for the primary architecture of the system), and a
595 couple of drop-in directories.
597 * udev's predictable network interface names now use the dev_port
598 sysfs attribute, introduced in linux 3.15 instead of dev_id to
599 distinguish between ports of the same PCI function. dev_id should
600 only be used for ports using the same HW address, hence the need
603 * machined has been updated to export the OS version of a
604 container (read from /etc/os-release and
605 /usr/lib/os-release) on the bus. This is now shown in
606 "machinectl status" for a machine.
608 * A new service setting RestartForceExitStatus= has been
609 added. If configured to a set of exit signals or process
610 return values, the service will be restarted when the main
611 daemon process exits with any of them, regardless of the
614 * systemctl's -H switch for connecting to remote systemd
615 machines has been extended so that it may be used to
616 directly connect to a specific container on the
617 host. "systemctl -H root@foobar:waldi" will now connect as
618 user "root" to host "foobar", and then proceed directly to
619 the container named "waldi". Note that currently you have to
620 authenticate as user "root" for this to work, as entering
621 containers is a privileged operation.
623 Contributions from: Andreas Henriksson, Benjamin Steinwender,
624 Carl Schaefer, Christian Hesse, Colin Ian King, Cristian
625 Rodríguez, Daniel Mack, Dave Reisner, David Herrmann, Eugene
626 Yakubovich, Filipe Brandenburger, Frederic Crozat, Hristo
627 Venev, Jan Engelhardt, Jonathan Boulle, Kay Sievers, Lennart
628 Poettering, Luke Shumaker, Mantas Mikulėnas, Marc-Antoine
629 Perennou, Marcel Holtmann, Michael Marineau, Michael Olbrich,
630 Michał Bartoszkiewicz, Michal Sekletar, Patrik Flykt, Ronan Le
631 Martret, Ronny Chevalier, Ruediger Oertel, Steven Noonan,
632 Susant Sahani, Thadeu Lima de Souza Cascardo, Thomas Hindoe
633 Paaboel Andersen, Tom Gundersen, Tom Hirst, Umut Tezduyar
634 Lindskog, Uoti Urpala, Zbigniew Jędrzejewski-Szmek
636 -- Berlin, 2014-07-03
640 * As an experimental feature, udev now tries to lock the
641 disk device node (flock(LOCK_SH|LOCK_NB)) while it
642 executes events for the disk or any of its partitions.
643 Applications like partitioning programs can lock the
644 disk device node (flock(LOCK_EX)) and claim temporary
645 device ownership that way; udev will entirely skip all event
646 handling for this disk and its partitions. If the disk
647 was opened for writing, the close will trigger a partition
648 table rescan in udev's "watch" facility, and if needed
649 synthesize "change" events for the disk and all its partitions.
650 This is now unconditionally enabled, and if it turns out to
651 cause major problems, we might turn it on only for specific
652 devices, or might need to disable it entirely. Device Mapper
653 devices are excluded from this logic.
655 * We temporarily dropped the "-l" switch for fsck invocations,
656 since they collide with the flock() logic above. util-linux
657 upstream has been changed already to avoid this conflict,
658 and we will readd "-l" as soon as util-linux with this
659 change has been released.
661 * The dependency on libattr has been removed. Since a long
662 time, the extended attribute calls have moved to glibc, and
663 libattr is thus unnecessary.
665 * Virtualization detection works without priviliges now. This
666 means the systemd-detect-virt binary no longer requires
667 CAP_SYS_PTRACE file capabilities, and our daemons can run
668 with fewer privileges.
670 * systemd-networkd now runs under its own "systemd-network"
671 user. It retains the CAP_NET_ADMIN, CAP_NET_BIND_SERVICE,
672 CAP_NET_BROADCAST, CAP_NET_RAW capabilities though, but
673 loses the ability to write to files owned by root this way.
675 * Similar, systemd-resolved now runs under its own
676 "systemd-resolve" user with no capabilities remaining.
678 * Similar, systemd-bus-proxyd now runs under its own
679 "systemd-bus-proxy" user with only CAP_IPC_OWNER remaining.
681 * systemd-networkd gained support for setting up "veth"
682 virtual ethernet devices for container connectivity, as well
683 as GRE and VTI tunnels.
685 * systemd-networkd will no longer automatically attempt to
686 manually load kernel modules necessary for certain tunnel
687 transports. Instead, it is assumed the kernel loads them
688 automatically when required. This only works correctly on
689 very new kernels. On older kernels, please consider adding
690 the kernel modules to /etc/modules-load.d/ as a work-around.
692 * The resolv.conf file systemd-resolved generates has been
693 moved to /run/systemd/resolve/. If you have a symlink from
694 /etc/resolv.conf, it might be necessary to correct it.
696 * Two new service settings, ProtectHome= and ProtectSystem=,
697 have been added. When enabled, they will make the user data
698 (such as /home) inaccessible or read-only and the system
699 (such as /usr) read-only, for specific services. This allows
700 very light-weight per-service sandboxing to avoid
701 modifications of user data or system files from
702 services. These two new switches have been enabled for all
703 of systemd's long-running services, where appropriate.
705 * Socket units gained new SocketUser= and SocketGroup=
706 settings to set the owner user and group of AF_UNIX sockets
707 and FIFOs in the file system.
709 * Socket units gained a new RemoveOnStop= setting. If enabled,
710 all FIFOS and sockets in the file system will be removed
711 when the specific socket unit is stopped.
713 * Socket units gained a new Symlinks= setting. It takes a list
714 of symlinks to create to file system sockets or FIFOs
715 created by the specific Unix sockets. This is useful to
716 manage symlinks to socket nodes with the same life-cycle as
719 * The /dev/log socket and /dev/initctl FIFO have been moved to
720 /run, and have been replaced by symlinks. This allows
721 connecting to these facilities even if PrivateDevices=yes is
722 used for a service (which makes /dev/log itself unavailable,
723 but /run is left). This also has the benefit of ensuring
724 that /dev only contains device nodes, directories and
725 symlinks, and nothing else.
727 * sd-daemon gained two new calls sd_pid_notify() and
728 sd_pid_notifyf(). They are similar to sd_notify() and
729 sd_notifyf(), but allow overriding of the source PID of
730 notification messages if permissions permit this. This is
731 useful to send notify messages on behalf of a different
732 process (for example, the parent process). The
733 systemd-notify tool has been updated to make use of this
734 when sending messages (so that notification messages now
735 originate from the shell script invoking systemd-notify and
736 not the systemd-notify process itself. This should minimize
737 a race where systemd fails to associate notification
738 messages to services when the originating process already
741 * A new "on-abnormal" setting for Restart= has been added. If
742 set, it will result in automatic restarts on all "abnormal"
743 reasons for a process to exit, which includes unclean
744 signals, core dumps, timeouts and watchdog timeouts, but
745 does not include clean and unclean exit codes or clean
746 signals. Restart=on-abnormal is an alternative for
747 Restart=on-failure for services that shall be able to
748 terminate and avoid restarts on certain errors, by
749 indicating so with an unclean exit code. Restart=on-failure
750 or Restart=on-abnormal is now the recommended setting for
751 all long-running services.
753 * If the InaccessibleDirectories= service setting points to a
754 mount point (or if there are any submounts contained within
755 it), it is now attempted to completely unmount it, to make
756 the file systems truly unavailable for the respective
759 * The ReadOnlyDirectories= service setting and
760 systemd-nspawn's --read-only parameter are now recursively
761 applied to all submounts, too.
763 * Mount units may now be created transiently via the bus APIs.
765 * The support for SysV and LSB init scripts has been removed
766 from the systemd daemon itself. Instead, it is now
767 implemented as a generator that creates native systemd units
768 from these scripts when needed. This enables us to remove a
769 substantial amount of legacy code from PID 1, following the
770 fact that many distributions only ship a very small number
771 of LSB/SysV init scripts nowadays.
773 * Priviliged Xen (dom0) domains are not considered
774 virtualization anymore by the virtualization detection
775 logic. After all, they generally have unrestricted access to
776 the hardware and usually are used to manage the unprivileged
779 * systemd-tmpfiles gained a new "C" line type, for copying
780 files or entire directories.
782 * systemd-tmpfiles "m" lines are now fully equivalent to "z"
783 lines. So far, they have been non-globbing versions of the
784 latter, and have thus been redundant. In future, it is
785 recommended to only use "z". "m" has hence been removed
786 from the documentation, even though it stays supported.
788 * A tmpfiles snippet to recreate the most basic structure in
789 /var has been added. This is enough to create the /var/run →
790 /run symlink and create a couple of structural
791 directories. This allows systems to boot up with an empty or
792 volatile /var. Of course, while with this change, the core OS
793 now is capable with dealing with a volatile /var, not all
794 user services are ready for it. However, we hope that sooner
795 or later, many service daemons will be changed upstream so
796 that they are able to automatically create their necessary
797 directories in /var at boot, should they be missing. This is
798 the first step to allow state-less systems that only require
799 the vendor image for /usr to boot.
801 * systemd-nspawn has gained a new --tmpfs= switch to mount an
802 empty tmpfs instance to a specific directory. This is
803 particularly useful for making use of the automatic
804 reconstruction of /var (see above), by passing --tmpfs=/var.
806 * Access modes specified in tmpfiles snippets may now be
807 prefixed with "~", which indicates that they shall be masked
808 by whether the existing file or directory is currently
809 writable, readable or executable at all. Also, if specified,
810 the sgid/suid/sticky bits will be masked for all
813 * A new passive target unit "network-pre.target" has been
814 added which is useful for services that shall run before any
815 network is configured, for example firewall scripts.
817 * The "floppy" group that previously owned the /dev/fd*
818 devices is no longer used. The "disk" group is now used
819 instead. Distributions should probably deprecate usage of
822 Contributions from: Camilo Aguilar, Christian Hesse, Colin Ian
823 King, Cristian Rodríguez, Daniel Buch, Dave Reisner, David
824 Strauss, Denis Tikhomirov, John, Jonathan Liu, Kay Sievers,
825 Lennart Poettering, Mantas Mikulėnas, Mark Eichin, Ronny
826 Chevalier, Susant Sahani, Thomas Blume, Thomas Hindoe Paaboel
827 Andersen, Tom Gundersen, Umut Tezduyar Lindskog, Zbigniew
830 -- Berlin, 2014-06-11
834 * A new "systemd-timesyncd" daemon has been added for
835 synchronizing the system clock across the network. It
836 implements an SNTP client. In contrast to NTP
837 implementations such as chrony or the NTP reference server,
838 this only implements a client side, and does not bother with
839 the full NTP complexity, focusing only on querying time from
840 one remote server and synchronizing the local clock to
841 it. Unless you intend to serve NTP to networked clients or
842 want to connect to local hardware clocks, this simple NTP
843 client should be more than appropriate for most
844 installations. The daemon runs with minimal privileges, and
845 has been hooked up with networkd to only operate when
846 network connectivity is available. The daemon saves the
847 current clock to disk every time a new NTP sync has been
848 acquired, and uses this to possibly correct the system clock
849 early at bootup, in order to accommodate for systems that
850 lack an RTC such as the Raspberry Pi and embedded devices,
851 and to make sure that time monotonically progresses on these
852 systems, even if it is not always correct. To make use of
853 this daemon, a new system user and group "systemd-timesync"
854 needs to be created on installation of systemd.
856 * The queue "seqnum" interface of libudev has been disabled, as
857 it was generally incompatible with device namespacing as
858 sequence numbers of devices go "missing" if the devices are
859 part of a different namespace.
861 * "systemctl list-timers" and "systemctl list-sockets" gained
862 a --recursive switch for showing units of these types also
863 for all local containers, similar in style to the already
864 supported --recursive switch for "systemctl list-units".
866 * A new RebootArgument= setting has been added for service
867 units, which may be used to specify a kernel reboot argument
868 to use when triggering reboots with StartLimitAction=.
870 * A new FailureAction= setting has been added for service
871 units which may be used to specify an operation to trigger
872 when a service fails. This works similarly to
873 StartLimitAction=, but unlike it, controls what is done
874 immediately rather than only after several attempts to
875 restart the service in question.
877 * hostnamed got updated to also expose the kernel name,
878 release, and version on the bus. This is useful for
879 executing commands like hostnamectl with the -H switch.
880 systemd-analyze makes use of this to properly display
881 details when running non-locally.
883 * The bootchart tool can now show cgroup information in the
886 * The CFS CPU quota cgroup attribute is now exposed for
887 services. The new CPUQuota= switch has been added for this
888 which takes a percentage value. Setting this will have the
889 result that a service may never get more CPU time than the
890 specified percentage, even if the machine is otherwise idle.
892 * systemd-networkd learned IPIP and SIT tunnel support.
894 * LSB init scripts exposing a dependency on $network will now
895 get a dependency on network-online.target rather than simply
896 network.target. This should bring LSB handling closer to
897 what it was on SysV systems.
899 * A new fsck.repair= kernel option has been added to control
900 how fsck shall deal with unclean file systems at boot.
902 * The (.ini) configuration file parser will now silently
903 ignore sections whose name begins with "X-". This may be
904 used to maintain application-specific extension sections in unit
907 * machined gained a new API to query the IP addresses of
908 registered containers. "machinectl status" has been updated
909 to show these addresses in its output.
911 * A new call sd_uid_get_display() has been added to the
912 sd-login APIs for querying the "primary" session of a
913 user. The "primary" session of the user is elected from the
914 user's sessions and generally a graphical session is
915 preferred over a text one.
917 * A minimal systemd-resolved daemon has been added. It
918 currently simply acts as a companion to systemd-networkd and
919 manages resolv.conf based on per-interface DNS
920 configuration, possibly supplied via DHCP. In the long run
921 we hope to extend this into a local DNSSEC enabled DNS and
924 * The systemd-networkd-wait-online tool is now enabled by
925 default. It will delay network-online.target until a network
926 connection has been configured. The tool primarily integrates
927 with networkd, but will also make a best effort to make sense
928 of network configuration performed in some other way.
930 * Two new service options StartupCPUShares= and
931 StartupBlockIOWeight= have been added that work similarly to
932 CPUShares= and BlockIOWeight= however only apply during
933 system startup. This is useful to prioritize certain services
934 differently during bootup than during normal runtime.
936 * hostnamed has been changed to prefer the statically
937 configured hostname in /etc/hostname (unless set to
938 'localhost' or empty) over any dynamic one supplied by
939 dhcp. With this change, the rules for picking the hostname
940 match more closely the rules of other configuration settings
941 where the local administrator's configuration in /etc always
942 overrides any other settings.
944 Contributions fron: Ali H. Caliskan, Alison Chaiken, Bas van
945 den Berg, Brandon Philips, Cristian Rodríguez, Daniel Buch,
946 Dan Kilman, Dave Reisner, David Härdeman, David Herrmann,
947 David Strauss, Dimitris Spingos, Djalal Harouni, Eelco
948 Dolstra, Evan Nemerson, Florian Albrechtskirchinger, Greg
949 Kroah-Hartman, Harald Hoyer, Holger Hans Peter Freyther, Jan
950 Engelhardt, Jani Nikula, Jason St. John, Jeffrey Clark,
951 Jonathan Boulle, Kay Sievers, Lennart Poettering, Lukas
952 Nykryn, Lukasz Skalski, Łukasz Stelmach, Mantas Mikulėnas,
953 Marcel Holtmann, Martin Pitt, Matthew Monaco, Michael
954 Marineau, Michael Olbrich, Michal Sekletar, Mike Gilbert, Nis
955 Martensen, Patrik Flykt, Philip Lorenz, poma, Ray Strode,
956 Reyad Attiyat, Robert Milasan, Scott Thrasher, Stef Walter,
957 Steven Siloti, Susant Sahani, Tanu Kaskinen, Thomas Bächler,
958 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar
959 Lindskog, WaLyong Cho, Will Woods, Zbigniew
962 -- Beijing, 2014-05-28
966 * When restoring the screen brightness at boot, stay away from
967 the darkest setting or from the lowest 5% of the available
968 range, depending on which is the larger value of both. This
969 should effectively protect the user from rebooting into a
970 black screen, should the brightness have been set to minimum
973 * sd-login gained a new sd_machine_get_class() call to
974 determine the class ("vm" or "container") of a machine
975 registered with machined.
977 * sd-login gained new calls
978 sd_peer_get_{session,owner_uid,unit,user_unit,slice,machine_name}(),
979 to query the identity of the peer of a local AF_UNIX
980 connection. They operate similarly to their sd_pid_get_xyz()
983 * PID 1 will now maintain a system-wide system state engine
984 with the states "starting", "running", "degraded",
985 "maintenance", "stopping". These states are bound to system
986 startup, normal runtime, runtime with at least one failed
987 service, rescue/emergency mode and system shutdown. This
988 state is shown in the "systemctl status" output when no unit
989 name is passed. It is useful to determine system state, in
990 particularly when doing so for many systems or containers at
993 * A new command "list-machines" has been added to "systemctl"
994 that lists all local OS containers and shows their system
995 state (see above), if systemd runs inside of them.
997 * systemctl gained a new "-r" switch to recursively enumerate
998 units on all local containers, when used with the
999 "list-unit" command (which is the default one that is
1000 executed when no parameters are specified).
1002 * The GPT automatic partition discovery logic will now honour
1003 two GPT partition flags: one may be set on a partition to
1004 cause it to be mounted read-only, and the other may be set
1005 on a partition to ignore it during automatic discovery.
1007 * Two new GPT type UUIDs have been added for automatic root
1008 partition discovery, for 32-bit and 64-bit ARM. This is not
1009 particularly useful for discovering the root directory on
1010 these architectures during bare-metal boots (since UEFI is
1011 not common there), but still very useful to allow booting of
1012 ARM disk images in nspawn with the -i option.
1014 * MAC addresses of interfaces created with nspawn's
1015 --network-interface= switch will now be generated from the
1016 machine name, and thus be stable between multiple invocations
1019 * logind will now automatically remove all IPC objects owned
1020 by a user if she or he fully logs out. This makes sure that
1021 users who are logged out cannot continue to consume IPC
1022 resources. This covers SysV memory, semaphores and message
1023 queues as well as POSIX shared memory and message
1024 queues. Traditionally, SysV and POSIX IPC had no life-cycle
1025 limits. With this functionality, that is corrected. This may
1026 be turned off by using the RemoveIPC= switch of logind.conf.
1028 * The systemd-machine-id-setup and tmpfiles tools gained a
1029 --root= switch to operate on a specific root directory,
1032 * journald can now forward logged messages to the TTYs of all
1033 logged in users ("wall"). This is the default for all
1034 emergency messages now.
1036 * A new tool systemd-journal-remote has been added to stream
1037 journal log messages across the network.
1039 * /sys/fs/cgroup/ is now mounted read-only after all cgroup
1040 controller trees are mounted into it. Note that the
1041 directories mounted beneath it are not read-only. This is a
1042 security measure and is particularly useful because glibc
1043 actually includes a search logic to pick any tmpfs it can
1044 find to implement shm_open() if /dev/shm is not available
1045 (which it might very well be in namespaced setups).
1047 * machinectl gained a new "poweroff" command to cleanly power
1048 down a local OS container.
1050 * The PrivateDevices= unit file setting will now also drop the
1051 CAP_MKNOD capability from the capability bound set, and
1052 imply DevicePolicy=closed.
1054 * PrivateDevices=, PrivateNetwork= and PrivateTmp= is now used
1055 comprehensively on all long-running systemd services where
1056 this is appropriate.
1058 * systemd-udevd will now run in a disassociated mount
1059 namespace. To mount directories from udev rules, make sure to
1060 pull in mount units via SYSTEMD_WANTS properties.
1062 * The kdbus support gained support for uploading policy into
1063 the kernel. sd-bus gained support for creating "monitoring"
1064 connections that can eavesdrop into all bus communication
1065 for debugging purposes.
1067 * Timestamps may now be specified in seconds since the UNIX
1068 epoch Jan 1st, 1970 by specifying "@" followed by the value
1071 * Native tcpwrap support in systemd has been removed. tcpwrap
1072 is old code, not really maintained anymore and has serious
1073 shortcomings, and better options such as firewalls
1074 exist. For setups that require tcpwrap usage, please
1075 consider invoking your socket-activated service via tcpd,
1076 like on traditional inetd.
1078 * A new system.conf configuration option
1079 DefaultTimerAccuracySec= has been added that controls the
1080 default AccuracySec= setting of .timer units.
1082 * Timer units gained a new WakeSystem= switch. If enabled,
1083 timers configured this way will cause the system to resume
1084 from system suspend (if the system supports that, which most
1087 * Timer units gained a new Persistent= switch. If enabled,
1088 timers configured this way will save to disk when they have
1089 been last triggered. This information is then used on next
1090 reboot to possible execute overdue timer events, that
1091 could not take place because the system was powered off.
1092 This enables simple anacron-like behaviour for timer units.
1094 * systemctl's "list-timers" will now also list the time a
1095 timer unit was last triggered in addition to the next time
1096 it will be triggered.
1098 * systemd-networkd will now assign predictable IPv4LL
1099 addresses to its local interfaces.
1101 Contributions from: Brandon Philips, Daniel Buch, Daniel Mack,
1102 Dave Reisner, David Herrmann, Gerd Hoffmann, Greg
1103 Kroah-Hartman, Hendrik Brueckner, Jason St. John, Josh
1104 Triplett, Kay Sievers, Lennart Poettering, Marc-Antoine
1105 Perennou, Michael Marineau, Michael Olbrich, Miklos Vajna,
1106 Patrik Flykt, poma, Sebastian Thorarensen, Thomas Bächler,
1107 Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom Gundersen,
1108 Umut Tezduyar Lindskog, Wieland Hoffmann, Zbigniew
1111 -- Berlin, 2014-03-25
1115 * A new unit file setting RestrictAddressFamilies= has been
1116 added to restrict which socket address families unit
1117 processes gain access to. This takes address family names
1118 like "AF_INET" or "AF_UNIX", and is useful to minimize the
1119 attack surface of services via exotic protocol stacks. This
1120 is built on seccomp system call filters.
1122 * Two new unit file settings RuntimeDirectory= and
1123 RuntimeDirectoryMode= have been added that may be used to
1124 manage a per-daemon runtime directories below /run. This is
1125 an alternative for setting up directory permissions with
1126 tmpfiles snippets, and has the advantage that the runtime
1127 directory's lifetime is bound to the daemon runtime and that
1128 the daemon starts up with an empty directory each time. This
1129 is particularly useful when writing services that drop
1130 privileges using the User= or Group= setting.
1132 * The DeviceAllow= unit setting now supports globbing for
1133 matching against device group names.
1135 * The systemd configuration file system.conf gained new
1136 settings DefaultCPUAccounting=, DefaultBlockIOAccounting=,
1137 DefaultMemoryAccounting= to globally turn on/off accounting
1138 for specific resources (cgroups) for all units. These
1139 settings may still be overridden individually in each unit
1142 * systemd-gpt-auto-generator is now able to discover /srv and
1143 root partitions in addition to /home and swap partitions. It
1144 also supports LUKS-encrypted partitions now. With this in
1145 place, automatic discovery of partitions to mount following
1146 the Discoverable Partitions Specification
1147 (http://www.freedesktop.org/wiki/Specifications/DiscoverablePartitionsSpec)
1148 is now a lot more complete. This allows booting without
1149 /etc/fstab and without root= on the kernel command line on
1150 systems prepared appropriately.
1152 * systemd-nspawn gained a new --image= switch which allows
1153 booting up disk images and Linux installations on any block
1154 device that follow the Discoverable Partitions Specification
1155 (see above). This means that installations made with
1156 appropriately updated installers may now be started and
1157 deployed using container managers, completely
1158 unmodified. (We hope that libvirt-lxc will add support for
1159 this feature soon, too.)
1161 * systemd-nspawn gained a new --network-macvlan= setting to
1162 set up a private macvlan interface for the
1163 container. Similarly, systemd-networkd gained a new
1164 Kind=macvlan setting in .netdev files.
1166 * systemd-networkd now supports configuring local addresses
1169 * A new tool systemd-network-wait-online has been added to
1170 synchronously wait for network connectivity using
1173 * The sd-bus.h bus API gained a new sd_bus_track object for
1174 tracking the life-cycle of bus peers. Note that sd-bus.h is
1175 still not a public API though (unless you specify
1176 --enable-kdbus on the configure command line, which however
1177 voids your warranty and you get no API stability guarantee).
1179 * The $XDG_RUNTIME_DIR runtime directories for each user are
1180 now individual tmpfs instances, which has the benefit of
1181 introducing separate pools for each user, with individual
1182 size limits, and thus making sure that unprivileged clients
1183 can no longer negatively impact the system or other users by
1184 filling up their $XDG_RUNTIME_DIR. A new logind.conf setting
1185 RuntimeDirectorySize= has been introduced that allows
1186 controlling the default size limit for all users. It
1187 defaults to 10% of the available physical memory. This is no
1188 replacement for quotas on tmpfs though (which the kernel
1189 still does not support), as /dev/shm and /tmp are still
1190 shared resources used by both the system and unprivileged
1193 * logind will now automatically turn off automatic suspending
1194 on laptop lid close when more than one display is
1195 connected. This was previously expected to be implemented
1196 individually in desktop environments (such as GNOME),
1197 however has been added to logind now, in order to fix a
1198 boot-time race where a desktop environment might not have
1199 been started yet and thus not been able to take an inhibitor
1200 lock at the time where logind already suspends the system
1201 due to a closed lid.
1203 * logind will now wait at least 30s after each system
1204 suspend/resume cycle, and 3min after system boot before
1205 suspending the system due to a closed laptop lid. This
1206 should give USB docking stations and similar enough time to
1207 be probed and configured after system resume and boot in
1208 order to then act as suspend blocker.
1210 * systemd-run gained a new --property= setting which allows
1211 initialization of resource control properties (and others)
1212 for the created scope or service unit. Example: "systemd-run
1213 --property=BlockIOWeight=10 updatedb" may be used to run
1214 updatedb at a low block IO scheduling weight.
1216 * systemd-run's --uid=, --gid=, --setenv=, --setenv= switches
1217 now also work in --scope mode.
1219 * When systemd is compiled with kdbus support, basic support
1220 for enforced policies is now in place. (Note that enabling
1221 kdbus still voids your warranty and no API compatibility
1224 Contributions from: Andrey Borzenkov, Ansgar Burchardt, Armin
1225 K., Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
1226 Harald Hoyer, Henrik Grindal Bakken, Jasper St. Pierre, Kay
1227 Sievers, Kieran Clancy, Lennart Poettering, Lukas Nykryn,
1228 Mantas Mikulėnas, Marcel Holtmann, Mark Oteiza, Martin Pitt,
1229 Mike Gilbert, Peter Rajnoha, poma, Samuli Suominen, Stef
1230 Walter, Susant Sahani, Tero Roponen, Thomas Andersen, Thomas
1231 Bächler, Thomas Hindoe Paaboel Andersen, Tomasz Torcz, Tom
1232 Gundersen, Umut Tezduyar Lindskog, Uoti Urpala, Zachary Cook,
1233 Zbigniew Jędrzejewski-Szmek
1235 -- Berlin, 2014-03-12
1239 * systemd will now relabel /dev after loading the SMACK policy
1240 according to SMACK rules.
1242 * A new unit file option AppArmorProfile= has been added to
1243 set the AppArmor profile for the processes of a unit.
1245 * A new condition check ConditionArchitecture= has been added
1246 to conditionalize units based on the system architecture, as
1247 reported by uname()'s "machine" field.
1249 * systemd-networkd now supports matching on the system
1250 virtualization, architecture, kernel command line, host name
1253 * logind is now a lot more aggressive when suspending the
1254 machine due to a closed laptop lid. Instead of acting only
1255 on the lid close action, it will continuously watch the lid
1256 status and act on it. This is useful for laptops where the
1257 power button is on the outside of the chassis so that it can
1258 be reached without opening the lid (such as the Lenovo
1259 Yoga). On those machines, logind will now immediately
1260 re-suspend the machine if the power button has been
1261 accidentally pressed while the laptop was suspended and in a
1262 backpack or similar.
1264 * logind will now watch SW_DOCK switches and inhibit reaction
1265 to the lid switch if it is pressed. This means that logind
1266 will not suspend the machine anymore if the lid is closed
1267 and the system is docked, if the laptop supports SW_DOCK
1268 notifications via the input layer. Note that ACPI docking
1269 stations do not generate this currently. Also note that this
1270 logic is usually not fully sufficient and Desktop
1271 Environments should take a lid switch inhibitor lock when an
1272 external display is connected, as systemd will not watch
1275 * nspawn will now make use of the devices cgroup controller by
1276 default, and only permit creation of and access to the usual
1277 API device nodes like /dev/null or /dev/random, as well as
1278 access to (but not creation of) the pty devices.
1280 * We will now ship a default .network file for
1281 systemd-networkd that automatically configures DHCP for
1282 network interfaces created by nspawn's --network-veth or
1283 --network-bridge= switches.
1285 * systemd will now understand the usual M, K, G, T suffixes
1286 according to SI conventions (i.e. to the base 1000) when
1287 referring to throughput and hardware metrics. It will stay
1288 with IEC conventions (i.e. to the base 1024) for software
1289 metrics, according to what is customary according to
1290 Wikipedia. We explicitly document which base applies for
1291 each configuration option.
1293 * The DeviceAllow= setting in unit files now supports a syntax
1294 to whitelist an entire group of devices node majors at once,
1295 based on the /proc/devices listing. For example, with the
1296 string "char-pts", it is now possible to whitelist all
1297 current and future pseudo-TTYs at once.
1299 * sd-event learned a new "post" event source. Event sources of
1300 this type are triggered by the dispatching of any event
1301 source of a type that is not "post". This is useful for
1302 implementing clean-up and check event sources that are
1303 triggered by other work being done in the program.
1305 * systemd-networkd is no longer statically enabled, but uses
1306 the usual [Install] sections so that it can be
1307 enabled/disabled using systemctl. It still is enabled by
1310 * When creating a veth interface pair with systemd-nspawn, the
1311 host side will now be prefixed with "vb-" if
1312 --network-bridge= is used, and with "ve-" if --network-veth
1313 is used. This way, it is easy to distinguish these cases on
1314 the host, for example to apply different configuration to
1315 them with systemd-networkd.
1317 * The compatibility libraries for libsystemd-journal.so,
1318 libsystem-id128.so, libsystemd-login.so and
1319 libsystemd-daemon.so do not make use of IFUNC
1320 anymore. Instead, we now build libsystemd.so multiple times
1321 under these alternative names. This means that the footprint
1322 is drastically increased, but given that these are
1323 transitional compatibility libraries, this should not matter
1324 much. This change has been made necessary to support the ARM
1325 platform for these compatibility libraries, as the ARM
1326 toolchain is not really at the same level as the toolchain
1327 for other architectures like x86 and does not support
1328 IFUNC. Please make sure to use --enable-compat-libs only
1329 during a transitional period!
1331 Contributions from: Andreas Fuchs, Armin K., Colin Walters,
1332 Daniel Mack, Dave Reisner, David Herrmann, Djalal Harouni,
1333 Holger Schurig, Jason A. Donenfeld, Jason St. John, Jasper
1334 St. Pierre, Kay Sievers, Lennart Poettering, Łukasz Stelmach,
1335 Marcel Holtmann, Michael Scherer, Michal Sekletar, Mike
1336 Gilbert, Samuli Suominen, Thomas Bächler, Thomas Hindoe
1337 Paaboel Andersen, Tom Gundersen, Umut Tezduyar Lindskog,
1338 Zbigniew Jędrzejewski-Szmek
1340 -- Berlin, 2014-02-24
1344 * A new component "systemd-networkd" has been added that can
1345 be used to configure local network interfaces statically or
1346 via DHCP. It is capable of bringing up bridges, VLANs, and
1347 bonding. Currently, no hook-ups for interactive network
1348 configuration are provided. Use this for your initrd,
1349 container, embedded, or server setup if you need a simple,
1350 yet powerful, network configuration solution. This
1351 configuration subsystem is quite nifty, as it allows wildcard
1352 hotplug matching in interfaces. For example, with a single
1353 configuration snippet, you can configure that all Ethernet
1354 interfaces showing up are automatically added to a bridge,
1355 or similar. It supports link-sensing and more.
1357 * A new tool "systemd-socket-proxyd" has been added which can
1358 act as a bidirectional proxy for TCP sockets. This is
1359 useful for adding socket activation support to services that
1360 do not actually support socket activation, including virtual
1361 machines and the like.
1363 * Add a new tool to save/restore rfkill state on
1366 * Save/restore state of keyboard backlights in addition to
1367 display backlights on shutdown/boot.
1369 * udev learned a new SECLABEL{} construct to label device
1370 nodes with a specific security label when they appear. For
1371 now, only SECLABEL{selinux} is supported, but the syntax is
1372 prepared for additional security frameworks.
1374 * udev gained a new scheme to configure link-level attributes
1375 from files in /etc/systemd/network/*.link. These files can
1376 match against MAC address, device path, driver name and type,
1377 and will apply attributes like the naming policy, link speed,
1378 MTU, duplex settings, Wake-on-LAN settings, MAC address, MAC
1379 address assignment policy (randomized, ...).
1381 * The configuration of network interface naming rules for
1382 "permanent interface names" has changed: a new NamePolicy=
1383 setting in the [Link] section of .link files determines the
1384 priority of possible naming schemes (onboard, slot, mac,
1385 path). The default value of this setting is determined by
1386 /usr/lib/net/links/99-default.link. Old
1387 80-net-name-slot.rules udev configuration file has been
1388 removed, so local configuration overriding this file should
1389 be adapated to override 99-default.link instead.
1391 * When the User= switch is used in a unit file, also
1392 initialize $SHELL= based on the user database entry.
1394 * systemd no longer depends on libdbus. All communication is
1395 now done with sd-bus, systemd's low-level bus library
1398 * kdbus support has been added to PID 1 itself. When kdbus is
1399 enabled, this causes PID 1 to set up the system bus and
1400 enable support for a new ".busname" unit type that
1401 encapsulates bus name activation on kdbus. It works a little
1402 bit like ".socket" units, except for bus names. A new
1403 generator has been added that converts classic dbus1 service
1404 activation files automatically into native systemd .busname
1407 * sd-bus: add a light-weight vtable implementation that allows
1408 defining objects on the bus with a simple static const
1409 vtable array of its methods, signals and properties.
1411 * systemd will not generate or install static dbus
1412 introspection data anymore to /usr/share/dbus-1/interfaces,
1413 as the precise format of these files is unclear, and
1414 nothing makes use of it.
1416 * A proxy daemon is now provided to proxy clients connecting
1417 via classic D-Bus AF_UNIX sockets to kdbus, to provide full
1418 compatibility with classic D-Bus.
1420 * A bus driver implementation has been added that supports the
1421 classic D-Bus bus driver calls on kdbus, also for
1422 compatibility purposes.
1424 * A new API "sd-event.h" has been added that implements a
1425 minimal event loop API built around epoll. It provides a
1426 couple of features that direct epoll usage is lacking:
1427 prioritization of events, scales to large numbers of timer
1428 events, per-event timer slack (accuracy), system-wide
1429 coalescing of timer events, exit handlers, watchdog
1430 supervision support using systemd's sd_notify() API, child
1433 * A new API "sd-rntl.h" has been added that provides an API
1434 around the route netlink interface of the kernel, similar in
1435 style to "sd-bus.h".
1437 * A new API "sd-dhcp-client.h" has been added that provides a
1438 small DHCPv4 client-side implementation. This is used by
1441 * There is a new kernel command line option
1442 "systemd.restore_state=0|1". When set to "0", none of the
1443 systemd tools will restore saved runtime state to hardware
1444 devices. More specifically, the rfkill and backlight states
1447 * The FsckPassNo= compatibility option in mount/service units
1448 has been removed. The fstab generator will now add the
1449 necessary dependencies automatically, and does not require
1450 PID1's support for that anymore.
1452 * journalctl gained a new switch, --list-boots, that lists
1453 recent boots with their times and boot IDs.
1455 * The various tools like systemctl, loginctl, timedatectl,
1456 busctl, systemd-run, ... have gained a new switch "-M" to
1457 connect to a specific, local OS container (as direct
1458 connection, without requiring SSH). This works on any
1459 container that is registered with machined, such as those
1460 created by libvirt-lxc or nspawn.
1462 * systemd-run and systemd-analyze also gained support for "-H"
1463 to connect to remote hosts via SSH. This is particularly
1464 useful for systemd-run because it enables queuing of jobs
1465 onto remote systems.
1467 * machinectl gained a new command "login" to open a getty
1468 login in any local container. This works with any container
1469 that is registered with machined (such as those created by
1470 libvirt-lxc or nspawn), and which runs systemd inside.
1472 * machinectl gained a new "reboot" command that may be used to
1473 trigger a reboot on a specific container that is registered
1474 with machined. This works on any container that runs an init
1475 system of some kind.
1477 * systemctl gained a new "list-timers" command to print a nice
1478 listing of installed timer units with the times they elapse
1481 * Alternative reboot() parameters may now be specified on the
1482 "systemctl reboot" command line and are passed to the
1483 reboot() system call.
1485 * systemctl gained a new --job-mode= switch to configure the
1486 mode to queue a job with. This is a more generic version of
1487 --fail, --irreversible, and --ignore-dependencies, which are
1488 still available but not advertised anymore.
1490 * /etc/systemd/system.conf gained new settings to configure
1491 various default timeouts of units, as well as the default
1492 start limit interval and burst. These may still be overridden
1495 * PID1 will now export on the bus profile data of the security
1496 policy upload process (such as the SELinux policy upload to
1499 * journald: when forwarding logs to the console, include
1500 timestamps (following the setting in
1501 /sys/module/printk/parameters/time).
1503 * OnCalendar= in timer units now understands the special
1504 strings "yearly" and "annually". (Both are equivalent)
1506 * The accuracy of timer units is now configurable with the new
1507 AccuracySec= setting. It defaults to 1min.
1509 * A new dependency type JoinsNamespaceOf= has been added that
1510 allows running two services within the same /tmp and network
1511 namespace, if PrivateNetwork= or PrivateTmp= are used.
1513 * A new command "cat" has been added to systemctl. It outputs
1514 the original unit file of a unit, and concatenates the
1515 contents of additional "drop-in" unit file snippets, so that
1516 the full configuration is shown.
1518 * systemctl now supports globbing on the various "list-xyz"
1519 commands, like "list-units" or "list-sockets", as well as on
1520 those commands which take multiple unit names.
1522 * journalctl's --unit= switch gained support for globbing.
1524 * All systemd daemons now make use of the watchdog logic so
1525 that systemd automatically notices when they hang.
1527 * If the $container_ttys environment variable is set,
1528 getty-generator will automatically spawn a getty for each
1529 listed tty. This is useful for container managers to request
1530 login gettys to be spawned on as many ttys as needed.
1532 * %h, %s, %U specifier support is not available anymore when
1533 used in unit files for PID 1. This is because NSS calls are
1534 not safe from PID 1. They stay available for --user
1535 instances of systemd, and as special case for the root user.
1537 * loginctl gained a new "--no-legend" switch to turn off output
1540 * The "sd-login.h" API gained three new calls:
1541 sd_session_is_remote(), sd_session_get_remote_user(),
1542 sd_session_get_remote_host() to query information about
1545 * The udev hardware database now also carries vendor/product
1546 information of SDIO devices.
1548 * The "sd-daemon.h" API gained a new sd_watchdog_enabled() to
1549 determine whether watchdog notifications are requested by
1552 * Socket-activated per-connection services now include a
1553 short description of the connection parameters in the
1556 * tmpfiles gained a new "--boot" option. When this is not used,
1557 only lines where the command character is not suffixed with
1558 "!" are executed. When this option is specified, those
1559 options are executed too. This partitions tmpfiles
1560 directives into those that can be safely executed at any
1561 time, and those which should be run only at boot (for
1562 example, a line that creates /run/nologin).
1564 * A new API "sd-resolve.h" has been added which provides a simple
1565 asynchronous wrapper around glibc NSS host name resolution
1566 calls, such as getaddrinfo(). In contrast to glibc's
1567 getaddrinfo_a(), it does not use signals. In contrast to most
1568 other asynchronous name resolution libraries, this one does
1569 not reimplement DNS, but reuses NSS, so that alternate
1570 host name resolution systems continue to work, such as mDNS,
1571 LDAP, etc. This API is based on libasyncns, but it has been
1572 cleaned up for inclusion in systemd.
1574 * The APIs "sd-journal.h", "sd-login.h", "sd-id128.h",
1575 "sd-daemon.h" are no longer found in individual libraries
1576 libsystemd-journal.so, libsystemd-login.so,
1577 libsystemd-id128.so, libsystemd-daemon.so. Instead, we have
1578 merged them into a single library, libsystemd.so, which
1579 provides all symbols. The reason for this is cyclic
1580 dependencies, as these libraries tend to use each other's
1581 symbols. So far, we have managed to workaround that by linking
1582 a copy of a good part of our code into each of these
1583 libraries again and again, which, however, makes certain
1584 things hard to do, like sharing static variables. Also, it
1585 substantially increases footprint. With this change, there
1586 is only one library for the basic APIs systemd
1587 provides. Also, "sd-bus.h", "sd-memfd.h", "sd-event.h",
1588 "sd-rtnl.h", "sd-resolve.h", "sd-utf8.h" are found in this
1589 library as well, however are subject to the --enable-kdbus
1590 switch (see below). Note that "sd-dhcp-client.h" is not part
1591 of this library (this is because it only consumes, never
1592 provides, services of/to other APIs). To make the transition
1593 easy from the separate libraries to the unified one, we
1594 provide the --enable-compat-libs compile-time switch which
1595 will generate stub libraries that are compatible with the
1596 old ones but redirect all calls to the new one.
1598 * All of the kdbus logic and the new APIs "sd-bus.h",
1599 "sd-memfd.h", "sd-event.h", "sd-rtnl.h", "sd-resolve.h",
1600 and "sd-utf8.h" are compile-time optional via the
1601 "--enable-kdbus" switch, and they are not compiled in by
1602 default. To make use of kdbus, you have to explicitly enable
1603 the switch. Note however, that neither the kernel nor the
1604 userspace API for all of this is considered stable yet. We
1605 want to maintain the freedom to still change the APIs for
1606 now. By specifying this build-time switch, you acknowledge
1607 that you are aware of the instability of the current
1610 * Also, note that while kdbus is pretty much complete,
1611 it lacks one thing: proper policy support. This means you
1612 can build a fully working system with all features; however,
1613 it will be highly insecure. Policy support will be added in
1614 one of the next releases, at the same time that we will
1615 declare the APIs stable.
1617 * When the kernel command-line argument "kdbus" is specified,
1618 systemd will automatically load the kdbus.ko kernel module. At
1619 this stage of development, it is only useful for testing kdbus
1620 and should not be used in production. Note: if "--enable-kdbus"
1621 is specified, and the kdbus.ko kernel module is available, and
1622 "kdbus" is added to the kernel command line, the entire system
1623 runs with kdbus instead of dbus-daemon, with the above mentioned
1624 problem of missing the system policy enforcement. Also a future
1625 version of kdbus.ko or a newer systemd will not be compatible with
1626 each other, and will unlikely be able to boot the machine if only
1627 one of them is updated.
1629 * systemctl gained a new "import-environment" command which
1630 uploads the caller's environment (or parts thereof) into the
1631 service manager so that it is inherited by services started
1632 by the manager. This is useful to upload variables like
1633 $DISPLAY into the user service manager.
1635 * A new PrivateDevices= switch has been added to service units
1636 which allows running a service with a namespaced /dev
1637 directory that does not contain any device nodes for
1638 physical devices. More specifically, it only includes devices
1639 such as /dev/null, /dev/urandom, and /dev/zero which are API
1642 * logind has been extended to support behaviour like VT
1643 switching on seats that do not support a VT. This makes
1644 multi-session available on seats that are not the first seat
1645 (seat0), and on systems where kernel support for VTs has
1646 been disabled at compile-time.
1648 * If a process holds a delay lock for system sleep or shutdown
1649 and fails to release it in time, we will now log its
1650 identity. This makes it easier to identify processes that
1651 cause slow suspends or power-offs.
1653 * When parsing /etc/crypttab, support for a new key-slot=
1654 option as supported by Debian is added. It allows indicating
1655 which LUKS slot to use on disk, speeding up key loading.
1657 * The sd_journald_sendv() API call has been checked and
1658 officially declared to be async-signal-safe so that it may
1659 be invoked from signal handlers for logging purposes.
1661 * Boot-time status output is now enabled automatically after a
1662 short timeout if boot does not progress, in order to give
1663 the user an indication what she or he is waiting for.
1665 * The boot-time output has been improved to show how much time
1666 remains until jobs expire.
1668 * The KillMode= switch in service units gained a new possible
1669 value "mixed". If set, and the unit is shut down, then the
1670 initial SIGTERM signal is sent only to the main daemon
1671 process, while the following SIGKILL signal is sent to
1672 all remaining processes of the service.
1674 * When a scope unit is registered, a new property "Controller"
1675 may be set. If set to a valid bus name, systemd will send a
1676 RequestStop() signal to this name when it would like to shut
1677 down the scope. This may be used to hook manager logic into
1678 the shutdown logic of scope units. Also, scope units may now
1679 be put in a special "abandoned" state, in which case the
1680 manager process which created them takes no further
1681 responsibilities for it.
1683 * When reading unit files, systemd will now verify
1684 the access mode of these files, and warn about certain
1685 suspicious combinations. This has been added to make it
1686 easier to track down packaging bugs where unit files are
1687 marked executable or world-writable.
1689 * systemd-nspawn gained a new "--setenv=" switch to set
1690 container-wide environment variables. The similar option in
1691 systemd-activate was renamed from "--environment=" to
1692 "--setenv=" for consistency.
1694 * systemd-nspawn has been updated to create a new kdbus domain
1695 for each container that is invoked, thus allowing each
1696 container to have its own set of system and user buses,
1697 independent of the host.
1699 * systemd-nspawn gained a new --drop-capability= switch to run
1700 the container with less capabilities than the default. Both
1701 --drop-capability= and --capability= now take the special
1702 string "all" for dropping or keeping all capabilities.
1704 * systemd-nspawn gained new switches for executing containers
1705 with specific SELinux labels set.
1707 * systemd-nspawn gained a new --quiet switch to not generate
1708 any additional output but the container's own console
1711 * systemd-nspawn gained a new --share-system switch to run a
1712 container without PID namespacing enabled.
1714 * systemd-nspawn gained a new --register= switch to control
1715 whether the container is registered with systemd-machined or
1716 not. This is useful for containers that do not run full
1717 OS images, but only specific apps.
1719 * systemd-nspawn gained a new --keep-unit which may be used
1720 when invoked as the only program from a service unit, and
1721 results in registration of the unit service itself in
1722 systemd-machined, instead of a newly opened scope unit.
1724 * systemd-nspawn gained a new --network-interface= switch for
1725 moving arbitrary interfaces to the container. The new
1726 --network-veth switch creates a virtual Ethernet connection
1727 between host and container. The new --network-bridge=
1728 switch then allows assigning the host side of this virtual
1729 Ethernet connection to a bridge device.
1731 * systemd-nspawn gained a new --personality= switch for
1732 setting the kernel personality for the container. This is
1733 useful when running a 32-bit container on a 64-bit host. A
1734 similar option Personality= is now also available for service
1737 * logind will now also track a "Desktop" identifier for each
1738 session which encodes the desktop environment of it. This is
1739 useful for desktop environments that want to identify
1740 multiple running sessions of itself easily.
1742 * A new SELinuxContext= setting for service units has been
1743 added that allows setting a specific SELinux execution
1744 context for a service.
1746 * Most systemd client tools will now honour $SYSTEMD_LESS for
1747 settings of the "less" pager. By default, these tools will
1748 override $LESS to allow certain operations to work, such as
1749 jump-to-the-end. With $SYSTEMD_LESS, it is possible to
1750 influence this logic.
1752 * systemd's "seccomp" hook-up has been changed to make use of
1753 the libseccomp library instead of using its own
1754 implementation. This has benefits for portability among
1757 * For usage together with SystemCallFilter=, a new
1758 SystemCallErrorNumber= setting has been introduced that
1759 allows configuration of a system error number to be returned
1760 on filtered system calls, instead of immediately killing the
1761 process. Also, SystemCallArchitectures= has been added to
1762 limit access to system calls of a particular architecture
1763 (in order to turn off support for unused secondary
1764 architectures). There is also a global
1765 SystemCallArchitectures= setting in system.conf now to turn
1766 off support for non-native system calls system-wide.
1768 * systemd requires a kernel with a working name_to_handle_at(),
1769 please see the kernel config requirements in the README file.
1771 Contributions from: Adam Williamson, Alex Jia, Anatol Pomozov,
1772 Ansgar Burchardt, AppleBloom, Auke Kok, Bastien Nocera,
1773 Chengwei Yang, Christian Seiler, Colin Guthrie, Colin Walters,
1774 Cristian Rodríguez, Daniel Buch, Daniele Medri, Daniel J
1775 Walsh, Daniel Mack, Dan McGee, Dave Reisner, David Coppa,
1776 David Herrmann, David Strauss, Djalal Harouni, Dmitry Pisklov,
1777 Elia Pinto, Florian Weimer, George McCollister, Goffredo
1778 Baroncelli, Greg Kroah-Hartman, Hendrik Brueckner, Igor
1779 Zhbanov, Jan Engelhardt, Jan Janssen, Jason A. Donenfeld,
1780 Jason St. John, Jasper St. Pierre, Jóhann B. Guðmundsson, Jose
1781 Ignacio Naranjo, Karel Zak, Kay Sievers, Kristian Høgsberg,
1782 Lennart Poettering, Lubomir Rintel, Lukas Nykryn, Lukasz
1783 Skalski, Łukasz Stelmach, Luke Shumaker, Mantas Mikulėnas,
1784 Marc-Antoine Perennou, Marcel Holtmann, Marcos Felipe Rasia de
1785 Mello, Marko Myllynen, Martin Pitt, Matthew Monaco, Michael
1786 Marineau, Michael Scherer, Michał Górny, Michal Sekletar,
1787 Michele Curti, Oleksii Shevchuk, Olivier Brunel, Patrik Flykt,
1788 Pavel Holica, Raudi, Richard Marko, Ronny Chevalier, Sébastien
1789 Luttringer, Sergey Ptashnick, Shawn Landden, Simon Peeters,
1790 Stefan Beller, Susant Sahani, Sylvain Plantefeve, Sylvia Else,
1791 Tero Roponen, Thomas Bächler, Thomas Hindoe Paaboel Andersen,
1792 Tom Gundersen, Umut Tezduyar Lindskog, Unai Uribarri, Václav
1793 Pavlín, Vincent Batts, WaLyong Cho, William Giokas, Yang
1794 Zhiyong, Yin Kangkai, Yuxuan Shui, Zbigniew Jędrzejewski-Szmek
1796 -- Berlin, 2014-02-20
1800 * logind has gained support for facilitating privileged input
1801 and drm device access for unprivileged clients. This work is
1802 useful to allow Wayland display servers (and similar
1803 programs, such as kmscon) to run under the user's ID and
1804 access input and drm devices which are normally
1805 protected. When this is used (and the kernel is new enough)
1806 logind will "mute" IO on the file descriptors passed to
1807 Wayland as long as it is in the background and "unmute" it
1808 if it returns into the foreground. This allows secure
1809 session switching without allowing background sessions to
1810 eavesdrop on input and display data. This also introduces
1811 session switching support if VT support is turned off in the
1812 kernel, and on seats that are not seat0.
1814 * A new kernel command line option luks.options= is understood
1815 now which allows specifying LUKS options for usage for LUKS
1816 encrypted partitions specified with luks.uuid=.
1818 * tmpfiles.d(5) snippets may now use specifier expansion in
1819 path names. More specifically %m, %b, %H, %v, are now
1820 replaced by the local machine id, boot id, hostname, and
1821 kernel version number.
1823 * A new tmpfiles.d(5) command "m" has been introduced which
1824 may be used to change the owner/group/access mode of a file
1825 or directory if it exists, but do nothing if it does not.
1827 * This release removes high-level support for the
1828 MemorySoftLimit= cgroup setting. The underlying kernel
1829 cgroup attribute memory.soft_limit= is currently badly
1830 designed and likely to be removed from the kernel API in its
1831 current form, hence we should not expose it for now.
1833 * The memory.use_hierarchy cgroup attribute is now enabled for
1834 all cgroups systemd creates in the memory cgroup
1835 hierarchy. This option is likely to be come the built-in
1836 default in the kernel anyway, and the non-hierarchial mode
1837 never made much sense in the intrinsically hierarchial
1840 * A new field _SYSTEMD_SLICE= is logged along with all journal
1841 messages containing the slice a message was generated
1842 from. This is useful to allow easy per-customer filtering of
1843 logs among other things.
1845 * systemd-journald will no longer adjust the group of journal
1846 files it creates to the "systemd-journal" group. Instead we
1847 rely on the journal directory to be owned by the
1848 "systemd-journal" group, and its setgid bit set, so that the
1849 kernel file system layer will automatically enforce that
1850 journal files inherit this group assignment. The reason for
1851 this change is that we cannot allow NSS look-ups from
1852 journald which would be necessary to resolve
1853 "systemd-journal" to a numeric GID, because this might
1854 create deadlocks if NSS involves synchronous queries to
1855 other daemons (such as nscd, or sssd) which in turn are
1856 logging clients of journald and might block on it, which
1857 would then dead lock. A tmpfiles.d(5) snippet included in
1858 systemd will make sure the setgid bit and group are
1859 properly set on the journal directory if it exists on every
1860 boot. However, we recommend adjusting it manually after
1861 upgrades too (or from RPM scriptlets), so that the change is
1862 not delayed until next reboot.
1864 * Backlight and random seed files in /var/lib/ have moved into
1865 the /var/lib/systemd/ directory, in order to centralize all
1866 systemd generated files in one directory.
1868 * Boot time performance measurements (as displayed by
1869 "systemd-analyze" for example) will now read ACPI 5.0 FPDT
1870 performance information if that's available to determine how
1871 much time BIOS and boot loader initialization required. With
1872 a sufficiently new BIOS you hence no longer need to boot
1873 with Gummiboot to get access to such information.
1875 Contributions from: Andrey Borzenkov, Chen Jie, Colin Walters,
1876 Cristian Rodríguez, Dave Reisner, David Herrmann, David
1877 Mackey, David Strauss, Eelco Dolstra, Evan Callicoat, Gao
1878 feng, Harald Hoyer, Jimmie Tauriainen, Kay Sievers, Lennart
1879 Poettering, Lukas Nykryn, Mantas Mikulėnas, Martin Pitt,
1880 Michael Scherer, Michał Górny, Mike Gilbert, Patrick McCarty,
1881 Sebastian Ott, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
1883 -- Berlin, 2013-10-02
1887 * The Restart= option for services now understands a new
1888 on-watchdog setting, which will restart the service
1889 automatically if the service stops sending out watchdog keep
1890 alive messages (as configured with WatchdogSec=).
1892 * The getty generator (which is responsible for bringing up a
1893 getty on configured serial consoles) will no longer only
1894 start a getty on the primary kernel console but on all
1895 others, too. This makes the order in which console= is
1896 specified on the kernel command line less important.
1898 * libsystemd-logind gained a new sd_session_get_vt() call to
1899 retrieve the VT number of a session.
1901 * If the option "tries=0" is set for an entry of /etc/crypttab
1902 its passphrase is queried indefinitely instead of any
1903 maximum number of tries.
1905 * If a service with a configure PID file terminates its PID
1906 file will now be removed automatically if it still exists
1907 afterwards. This should put an end to stale PID files.
1909 * systemd-run will now also take relative binary path names
1910 for execution and no longer insists on absolute paths.
1912 * InaccessibleDirectories= and ReadOnlyDirectories= now take
1913 paths that are optionally prefixed with "-" to indicate that
1914 it should not be considered a failure if they do not exist.
1916 * journalctl -o (and similar commands) now understands a new
1917 output mode "short-precise", it is similar to "short" but
1918 shows timestamps with usec accuracy.
1920 * The option "discard" (as known from Debian) is now
1921 synonymous to "allow-discards" in /etc/crypttab. In fact,
1922 "discard" is preferred now (since it is easier to remember
1925 * Some licensing clean-ups were made, so that more code is now
1926 LGPL-2.1 licensed than before.
1928 * A minimal tool to save/restore the display backlight
1929 brightness across reboots has been added. It will store the
1930 backlight setting as late as possible at shutdown, and
1931 restore it as early as possible during reboot.
1933 * A logic to automatically discover and enable home and swap
1934 partitions on GPT disks has been added. With this in place
1935 /etc/fstab becomes optional for many setups as systemd can
1936 discover certain partitions located on the root disk
1937 automatically. Home partitions are recognized under their
1938 GPT type ID 933ac7e12eb44f13b8440e14e2aef915. Swap
1939 partitions are recognized under their GPT type ID
1940 0657fd6da4ab43c484e50933c84b4f4f.
1942 * systemd will no longer pass any environment from the kernel
1943 or initrd to system services. If you want to set an
1944 environment for all services, do so via the kernel command
1945 line systemd.setenv= assignment.
1947 * The systemd-sysctl tool no longer natively reads the file
1948 /etc/sysctl.conf. If desired, the file should be symlinked
1949 from /etc/sysctl.d/99-sysctl.conf. Apart from providing
1950 legacy support by a symlink rather than built-in code, it
1951 also makes the otherwise hidden order of application of the
1952 different files visible. (Note that this partly reverts to a
1953 pre-198 application order of sysctl knobs!)
1955 * The "systemctl set-log-level" and "systemctl dump" commands
1956 have been moved to systemd-analyze.
1958 * systemd-run learned the new --remain-after-exit switch,
1959 which causes the scope unit not to be cleaned up
1960 automatically after the process terminated.
1962 * tmpfiles learned a new --exclude-prefix= switch to exclude
1963 certain paths from operation.
1965 * journald will now automatically flush all messages to disk
1966 as soon as a message of the log priorities CRIT, ALERT or
1969 Contributions from: Andrew Cook, Brandon Philips, Christian
1970 Hesse, Christoph Junghans, Colin Walters, Daniel Schaal,
1971 Daniel Wallace, Dave Reisner, David Herrmann, Gao feng, George
1972 McCollister, Giovanni Campagna, Hannes Reinecke, Harald Hoyer,
1973 Herczeg Zsolt, Holger Hans Peter Freyther, Jan Engelhardt,
1974 Jesper Larsen, Kay Sievers, Khem Raj, Lennart Poettering,
1975 Lukas Nykryn, Maciej Wereski, Mantas Mikulėnas, Marcel
1976 Holtmann, Martin Pitt, Michael Biebl, Michael Marineau,
1977 Michael Scherer, Michael Stapelberg, Michal Sekletar, Michał
1978 Górny, Olivier Brunel, Ondrej Balaz, Ronny Chevalier, Shawn
1979 Landden, Steven Hiscocks, Thomas Bächler, Thomas Hindoe
1980 Paaboel Andersen, Tom Gundersen, Umut Tezduyar, WANG Chao,
1981 William Giokas, Zbigniew Jędrzejewski-Szmek
1983 -- Berlin, 2013-09-13
1987 * The documentation has been updated to cover the various new
1988 concepts introduced with 205.
1990 * Unit files now understand the new %v specifier which
1991 resolves to the kernel version string as returned by "uname
1994 * systemctl now supports filtering the unit list output by
1995 load state, active state and sub state, using the new
1998 * "systemctl status" will now show the results of the
1999 condition checks (like ConditionPathExists= and similar) of
2000 the last start attempts of the unit. They are also logged to
2003 * "journalctl -b" may now be used to look for boot output of a
2004 specific boot. Try "journalctl -b -1" for the previous boot,
2005 but the syntax is substantially more powerful.
2007 * "journalctl --show-cursor" has been added which prints the
2008 cursor string the last shown log line. This may then be used
2009 with the new "journalctl --after-cursor=" switch to continue
2010 browsing logs from that point on.
2012 * "journalctl --force" may now be used to force regeneration
2015 * Creation of "dead" device nodes has been moved from udev
2016 into kmod and tmpfiles. Previously, udev would read the kmod
2017 databases to pre-generate dead device nodes based on meta
2018 information contained in kernel modules, so that these would
2019 be auto-loaded on access rather then at boot. As this
2020 does not really have much to do with the exposing actual
2021 kernel devices to userspace this has always been slightly
2022 alien in the udev codebase. Following the new scheme kmod
2023 will now generate a runtime snippet for tmpfiles from the
2024 module meta information and it now is tmpfiles' job to the
2025 create the nodes. This also allows overriding access and
2026 other parameters for the nodes using the usual tmpfiles
2027 facilities. As side effect this allows us to remove the
2028 CAP_SYS_MKNOD capability bit from udevd entirely.
2030 * logind's device ACLs may now be applied to these "dead"
2031 devices nodes too, thus finally allowing managed access to
2032 devices such as /dev/snd/sequencer whithout loading the
2033 backing module right-away.
2035 * A new RPM macro has been added that may be used to apply
2036 tmpfiles configuration during package installation.
2038 * systemd-detect-virt and ConditionVirtualization= now can
2039 detect User-Mode-Linux machines (UML).
2041 * journald will now implicitly log the effective capabilities
2042 set of processes in the message metadata.
2044 * systemd-cryptsetup has gained support for TrueCrypt volumes.
2046 * The initrd interface has been simplified (more specifically,
2047 support for passing performance data via environment
2048 variables and fsck results via files in /run has been
2049 removed). These features were non-essential, and are
2050 nowadays available in a much nicer way by having systemd in
2051 the initrd serialize its state and have the hosts systemd
2052 deserialize it again.
2054 * The udev "keymap" data files and tools to apply keyboard
2055 specific mappings of scan to key codes, and force-release
2056 scan code lists have been entirely replaced by a udev
2057 "keyboard" builtin and a hwdb data file.
2059 * systemd will now honour the kernel's "quiet" command line
2060 argument also during late shutdown, resulting in a
2061 completely silent shutdown when used.
2063 * There's now an option to control the SO_REUSEPORT socket
2064 option in .socket units.
2066 * Instance units will now automatically get a per-template
2067 subslice of system.slice unless something else is explicitly
2068 configured. For example, instances of sshd@.service will now
2069 implicitly be placed in system-sshd.slice rather than
2070 system.slice as before.
2072 * Test coverage support may now be enabled at build time.
2074 Contributions from: Dave Reisner, Frederic Crozat, Harald
2075 Hoyer, Holger Hans Peter Freyther, Jan Engelhardt, Jan
2076 Janssen, Jason St. John, Jesper Larsen, Kay Sievers, Lennart
2077 Poettering, Lukas Nykryn, Maciej Wereski, Martin Pitt, Michael
2078 Olbrich, Ramkumar Ramachandra, Ross Lagerwall, Shawn Landden,
2079 Thomas H.P. Andersen, Tom Gundersen, Tomasz Torcz, William
2080 Giokas, Zbigniew Jędrzejewski-Szmek
2082 -- Berlin, 2013-07-23
2086 * Two new unit types have been introduced:
2088 Scope units are very similar to service units, however, are
2089 created out of pre-existing processes -- instead of PID 1
2090 forking off the processes. By using scope units it is
2091 possible for system services and applications to group their
2092 own child processes (worker processes) in a powerful way
2093 which then maybe used to organize them, or kill them
2094 together, or apply resource limits on them.
2096 Slice units may be used to partition system resources in an
2097 hierarchial fashion and then assign other units to them. By
2098 default there are now three slices: system.slice (for all
2099 system services), user.slice (for all user sessions),
2100 machine.slice (for VMs and containers).
2102 Slices and scopes have been introduced primarily in
2103 context of the work to move cgroup handling to a
2104 single-writer scheme, where only PID 1
2105 creates/removes/manages cgroups.
2107 * There's a new concept of "transient" units. In contrast to
2108 normal units these units are created via an API at runtime,
2109 not from configuration from disk. More specifically this
2110 means it is now possible to run arbitrary programs as
2111 independent services, with all execution parameters passed
2112 in via bus APIs rather than read from disk. Transient units
2113 make systemd substantially more dynamic then it ever was,
2114 and useful as a general batch manager.
2116 * logind has been updated to make use of scope and slice units
2117 for managing user sessions. As a user logs in he will get
2118 his own private slice unit, to which all sessions are added
2119 as scope units. We also added support for automatically
2120 adding an instance of user@.service for the user into the
2121 slice. Effectively logind will no longer create cgroup
2122 hierarchies on its own now, it will defer entirely to PID 1
2123 for this by means of scope, service and slice units. Since
2124 user sessions this way become entities managed by PID 1
2125 the output of "systemctl" is now a lot more comprehensive.
2127 * A new mini-daemon "systemd-machined" has been added which
2128 may be used by virtualization managers to register local
2129 VMs/containers. nspawn has been updated accordingly, and
2130 libvirt will be updated shortly. machined will collect a bit
2131 of meta information about the VMs/containers, and assign
2132 them their own scope unit (see above). The collected
2133 meta-data is then made available via the "machinectl" tool,
2134 and exposed in "ps" and similar tools. machined/machinectl
2135 is compile-time optional.
2137 * As discussed earlier, the low-level cgroup configuration
2138 options ControlGroup=, ControlGroupModify=,
2139 ControlGroupPersistent=, ControlGroupAttribute= have been
2140 removed. Please use high-level attribute settings instead as
2141 well as slice units.
2143 * A new bus call SetUnitProperties() has been added to alter
2144 various runtime parameters of a unit. This is primarily
2145 useful to alter cgroup parameters dynamically in a nice way,
2146 but will be extended later on to make more properties
2147 modifiable at runtime. systemctl gained a new set-properties
2148 command that wraps this call.
2150 * A new tool "systemd-run" has been added which can be used to
2151 run arbitrary command lines as transient services or scopes,
2152 while configuring a number of settings via the command
2153 line. This tool is currently very basic, however already
2154 very useful. We plan to extend this tool to even allow
2155 queuing of execution jobs with time triggers from the
2156 command line, similar in fashion to "at".
2158 * nspawn will now inform the user explicitly that kernels with
2159 audit enabled break containers, and suggest the user to turn
2162 * Support for detecting the IMA and AppArmor security
2163 frameworks with ConditionSecurity= has been added.
2165 * journalctl gained a new "-k" switch for showing only kernel
2166 messages, mimicking dmesg output; in addition to "--user"
2167 and "--system" switches for showing only user's own logs
2170 * systemd-delta can now show information about drop-in
2171 snippets extending unit files.
2173 * libsystemd-bus has been substantially updated but is still
2174 not available as public API.
2176 * systemd will now look for the "debug" argument on the kernel
2177 command line and enable debug logging, similar to what
2178 "systemd.log_level=debug" already did before.
2180 * "systemctl set-default", "systemctl get-default" has been
2181 added to configure the default.target symlink, which
2182 controls what to boot into by default.
2184 * "systemctl set-log-level" has been added as a convenient
2185 way to raise and lower systemd logging threshold.
2187 * "systemd-analyze plot" will now show the time the various
2188 generators needed for execution, as well as information
2189 about the unit file loading.
2191 * libsystemd-journal gained a new sd_journal_open_files() call
2192 for opening specific journal files. journactl also gained a
2193 new switch to expose this new functionality. Previously we
2194 only supported opening all files from a directory, or all
2195 files from the system, as opening individual files only is
2196 racy due to journal file rotation.
2198 * systemd gained the new DefaultEnvironment= setting in
2199 /etc/systemd/system.conf to set environment variables for
2202 * If a privileged process logs a journal message with the
2203 OBJECT_PID= field set, then journald will automatically
2204 augment this with additional OBJECT_UID=, OBJECT_GID=,
2205 OBJECT_COMM=, OBJECT_EXE=, ... fields. This is useful if
2206 system services want to log events about specific client
2207 processes. journactl/systemctl has been updated to make use
2208 of this information if all log messages regarding a specific
2211 Contributions from: Auke Kok, Chengwei Yang, Colin Walters,
2212 Cristian Rodríguez, Daniel Albers, Daniel Wallace, Dave
2213 Reisner, David Coppa, David King, David Strauss, Eelco
2214 Dolstra, Gabriel de Perthuis, Harald Hoyer, Jan Alexander
2215 Steffens, Jan Engelhardt, Jan Janssen, Jason St. John, Johan
2216 Heikkilä, Karel Zak, Karol Lewandowski, Kay Sievers, Lennart
2217 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marius Vollmer,
2218 Martin Pitt, Michael Biebl, Michael Olbrich, Michael Tremer,
2219 Michal Schmidt, Michał Bartoszkiewicz, Nirbheek Chauhan,
2220 Pierre Neidhardt, Ross Burton, Ross Lagerwall, Sean McGovern,
2221 Thomas Hindoe Paaboel Andersen, Tom Gundersen, Umut Tezduyar,
2222 Václav Pavlín, Zachary Cook, Zbigniew Jędrzejewski-Szmek,
2223 Łukasz Stelmach, 장동준
2227 * The Python bindings gained some minimal support for the APIs
2228 exposed by libsystemd-logind.
2230 * ConditionSecurity= gained support for detecting SMACK. Since
2231 this condition already supports SELinux and AppArmor we only
2232 miss IMA for this. Patches welcome!
2234 Contributions from: Karol Lewandowski, Lennart Poettering,
2235 Zbigniew Jędrzejewski-Szmek
2239 * systemd-nspawn will now create /etc/resolv.conf if
2240 necessary, before bind-mounting the host's file onto it.
2242 * systemd-nspawn will now store meta information about a
2243 container on the container's cgroup as extended attribute
2244 fields, including the root directory.
2246 * The cgroup hierarchy has been reworked in many ways. All
2247 objects any of the components systemd creates in the cgroup
2248 tree are now suffixed. More specifically, user sessions are
2249 now placed in cgroups suffixed with ".session", users in
2250 cgroups suffixed with ".user", and nspawn containers in
2251 cgroups suffixed with ".nspawn". Furthermore, all cgroup
2252 names are now escaped in a simple scheme to avoid collision
2253 of userspace object names with kernel filenames. This work
2254 is preparation for making these objects relocatable in the
2255 cgroup tree, in order to allow easy resource partitioning of
2256 these objects without causing naming conflicts.
2258 * systemctl list-dependencies gained the new switches
2259 --plain, --reverse, --after and --before.
2261 * systemd-inhibit now shows the process name of processes that
2262 have taken an inhibitor lock.
2264 * nss-myhostname will now also resolve "localhost"
2265 implicitly. This makes /etc/hosts an optional file and
2266 nicely handles that on IPv6 ::1 maps to both "localhost" and
2269 * libsystemd-logind.so gained a new call
2270 sd_get_machine_names() to enumerate running containers and
2271 VMs (currently only supported by very new libvirt and
2272 nspawn). sd_login_monitor can now be used to watch
2273 VMs/containers coming and going.
2275 * .include is not allowed recursively anymore, and only in
2276 unit files. Usually it is better to use drop-in snippets in
2277 .d/*.conf anyway, as introduced with systemd 198.
2279 * systemd-analyze gained a new "critical-chain" command that
2280 determines the slowest chain of units run during system
2281 boot-up. It is very useful for tracking down where
2282 optimizing boot time is the most beneficial.
2284 * systemd will no longer allow manipulating service paths in
2285 the name=systemd:/system cgroup tree using ControlGroup= in
2286 units. (But is still fine with it in all other dirs.)
2288 * There's a new systemd-nspawn@.service service file that may
2289 be used to easily run nspawn containers as system
2290 services. With the container's root directory in
2291 /var/lib/container/foobar it is now sufficient to run
2292 "systemctl start systemd-nspawn@foobar.service" to boot it.
2294 * systemd-cgls gained a new parameter "--machine" to list only
2295 the processes within a certain container.
2297 * ConditionSecurity= now can check for "apparmor". We still
2298 are lacking checks for SMACK and IMA for this condition
2299 check though. Patches welcome!
2301 * A new configuration file /etc/systemd/sleep.conf has been
2302 added that may be used to configure which kernel operation
2303 systemd is supposed to execute when "suspend", "hibernate"
2304 or "hybrid-sleep" is requested. This makes the new kernel
2305 "freeze" state accessible to the user.
2307 * ENV{SYSTEMD_WANTS} in udev rules will now implicitly escape
2308 the passed argument if applicable.
2310 Contributions from: Auke Kok, Colin Guthrie, Colin Walters,
2311 Cristian Rodríguez, Daniel Buch, Daniel Wallace, Dave Reisner,
2312 Evangelos Foutras, Greg Kroah-Hartman, Harald Hoyer, Josh
2313 Triplett, Kay Sievers, Lennart Poettering, Lukas Nykryn,
2314 MUNEDA Takahiro, Mantas Mikulėnas, Mirco Tischler, Nathaniel
2315 Chen, Nirbheek Chauhan, Ronny Chevalier, Ross Lagerwall, Tom
2316 Gundersen, Umut Tezduyar, Ville Skyttä, Zbigniew
2321 * The output of 'systemctl list-jobs' got some polishing. The
2322 '--type=' argument may now be passed more than once. A new
2323 command 'systemctl list-sockets' has been added which shows
2324 a list of kernel sockets systemd is listening on with the
2325 socket units they belong to, plus the units these socket
2328 * The experimental libsystemd-bus library got substantial
2329 updates to work in conjunction with the (also experimental)
2330 kdbus kernel project. It works well enough to exchange
2331 messages with some sophistication. Note that kdbus is not
2332 ready yet, and the library is mostly an elaborate test case
2333 for now, and not installable.
2335 * systemd gained a new unit 'systemd-static-nodes.service'
2336 that generates static device nodes earlier during boot, and
2337 can run in conjunction with udev.
2339 * libsystemd-login gained a new call sd_pid_get_user_unit()
2340 to retrieve the user systemd unit a process is running
2341 in. This is useful for systems where systemd is used as
2344 * systemd-nspawn now places all containers in the new /machine
2345 top-level cgroup directory in the name=systemd
2346 hierarchy. libvirt will soon do the same, so that we get a
2347 uniform separation of /system, /user and /machine for system
2348 services, user processes and containers/virtual
2349 machines. This new cgroup hierarchy is also useful to stick
2350 stable names to specific container instances, which can be
2351 recognized later this way (this name may be controlled
2352 via systemd-nspawn's new -M switch). libsystemd-login also
2353 gained a new call sd_pid_get_machine_name() to retrieve the
2354 name of the container/VM a specific process belongs to.
2356 * bootchart can now store its data in the journal.
2358 * libsystemd-journal gained a new call
2359 sd_journal_add_conjunction() for AND expressions to the
2360 matching logic. This can be used to express more complex
2361 logical expressions.
2363 * journactl can now take multiple --unit= and --user-unit=
2366 * The cryptsetup logic now understands the "luks.key=" kernel
2367 command line switch for specifying a file to read the
2368 decryption key from. Also, if a configured key file is not
2369 found the tool will now automatically fall back to prompting
2372 * Python systemd.journal module was updated to wrap recently
2373 added functions from libsystemd-journal. The interface was
2374 changed to bring the low level interface in s.j._Reader
2375 closer to the C API, and the high level interface in
2376 s.j.Reader was updated to wrap and convert all data about
2379 Contributions from: Anatol Pomozov, Auke Kok, Harald Hoyer,
2380 Henrik Grindal Bakken, Josh Triplett, Kay Sievers, Lennart
2381 Poettering, Lukas Nykryn, Mantas Mikulėnas Marius Vollmer,
2382 Martin Jansa, Martin Pitt, Michael Biebl, Michal Schmidt,
2383 Mirco Tischler, Pali Rohar, Simon Peeters, Steven Hiscocks,
2384 Tom Gundersen, Zbigniew Jędrzejewski-Szmek
2388 * journalctl --update-catalog now understands a new --root=
2389 option to operate on catalogs found in a different root
2392 * During shutdown after systemd has terminated all running
2393 services a final killing loop kills all remaining left-over
2394 processes. We will now print the name of these processes
2395 when we send SIGKILL to them, since this usually indicates a
2398 * If /etc/crypttab refers to password files stored on
2399 configured mount points automatic dependencies will now be
2400 generated to ensure the specific mount is established first
2401 before the key file is attempted to be read.
2403 * 'systemctl status' will now show information about the
2404 network sockets a socket unit is listening on.
2406 * 'systemctl status' will also shown information about any
2407 drop-in configuration file for units. (Drop-In configuration
2408 files in this context are files such as
2409 /etc/systemd/systemd/foobar.service.d/*.conf)
2411 * systemd-cgtop now optionally shows summed up CPU times of
2412 cgroups. Press '%' while running cgtop to switch between
2413 percentage and absolute mode. This is useful to determine
2414 which cgroups use up the most CPU time over the entire
2415 runtime of the system. systemd-cgtop has also been updated
2416 to be 'pipeable' for processing with further shell tools.
2418 * 'hostnamectl set-hostname' will now allow setting of FQDN
2421 * The formatting and parsing of time span values has been
2422 changed. The parser now understands fractional expressions
2423 such as "5.5h". The formatter will now output fractional
2424 expressions for all time spans under 1min, i.e. "5.123456s"
2425 rather than "5s 123ms 456us". For time spans under 1s
2426 millisecond values are shown, for those under 1ms
2427 microsecond values are shown. This should greatly improve
2428 all time-related output of systemd.
2430 * libsystemd-login and libsystemd-journal gained new
2431 functions for querying the poll() events mask and poll()
2432 timeout value for integration into arbitrary event
2435 * localectl gained the ability to list available X11 keymaps
2436 (models, layouts, variants, options).
2438 * 'systemd-analyze dot' gained the ability to filter for
2439 specific units via shell-style globs, to create smaller,
2440 more useful graphs. I.e. it is now possible to create simple
2441 graphs of all the dependencies between only target units, or
2442 of all units that Avahi has dependencies with.
2444 Contributions from: Cristian Rodríguez, Dr. Tilmann Bubeck,
2445 Harald Hoyer, Holger Hans Peter Freyther, Kay Sievers, Kelly
2446 Anderson, Koen Kooi, Lennart Poettering, Maksim Melnikau,
2447 Marc-Antoine Perennou, Marius Vollmer, Martin Pitt, Michal
2448 Schmidt, Oleksii Shevchuk, Ronny Chevalier, Simon McVittie,
2449 Steven Hiscocks, Thomas Weißschuh, Umut Tezduyar, Václav
2450 Pavlín, Zbigniew Jędrzejewski-Szmek, Łukasz Stelmach
2454 * The boot-time readahead implementation for rotating media
2455 will now read the read-ahead data in multiple passes which
2456 consist of all read requests made in equidistant time
2457 intervals. This means instead of strictly reading read-ahead
2458 data in its physical order on disk we now try to find a
2459 middle ground between physical and access time order.
2461 * /etc/os-release files gained a new BUILD_ID= field for usage
2462 on operating systems that provide continuous builds of OS
2465 Contributions from: Auke Kok, Eelco Dolstra, Kay Sievers,
2466 Lennart Poettering, Lukas Nykryn, Martin Pitt, Václav Pavlín
2467 William Douglas, Zbigniew Jędrzejewski-Szmek
2471 * systemd-python gained an API exposing libsystemd-daemon.
2473 * The SMACK setup logic gained support for uploading CIPSO
2476 * Behaviour of PrivateTmp=, ReadWriteDirectories=,
2477 ReadOnlyDirectories= and InaccessibleDirectories= has
2478 changed. The private /tmp and /var/tmp directories are now
2479 shared by all processes of a service (which means
2480 ExecStartPre= may now leave data in /tmp that ExecStart= of
2481 the same service can still access). When a service is
2482 stopped its temporary directories are immediately deleted
2483 (normal clean-up with tmpfiles is still done in addition to
2486 * By default, systemd will now set a couple of sysctl
2487 variables in the kernel: the safe sysrq options are turned
2488 on, IP route verification is turned on, and source routing
2489 disabled. The recently added hardlink and softlink
2490 protection of the kernel is turned on. These settings should
2491 be reasonably safe, and good defaults for all new systems.
2493 * The predictable network naming logic may now be turned off
2494 with a new kernel command line switch: net.ifnames=0.
2496 * A new libsystemd-bus module has been added that implements a
2497 pretty complete D-Bus client library. For details see:
2499 http://lists.freedesktop.org/archives/systemd-devel/2013-March/009797.html
2501 * journald will now explicitly flush the journal files to disk
2502 at the latest 5min after each write. The file will then also
2503 be marked offline until the next write. This should increase
2504 reliability in case of a crash. The synchronization delay
2505 can be configured via SyncIntervalSec= in journald.conf.
2507 * There's a new remote-fs-setup.target unit that can be used
2508 to pull in specific services when at least one remote file
2509 system is to be mounted.
2511 * There are new targets timers.target and paths.target as
2512 canonical targets to pull user timer and path units in
2513 from. This complements sockets.target with a similar
2514 purpose for socket units.
2516 * libudev gained a new call udev_device_set_attribute_value()
2517 to set sysfs attributes of a device.
2519 * The udev daemon now sets the default number of worker
2520 processes executed in parallel based on the number of available
2521 CPUs instead of the amount of available RAM. This is supposed
2522 to provide a more reliable default and limit a too aggressive
2523 paralellism for setups with 1000s of devices connected.
2525 Contributions from: Auke Kok, Colin Walters, Cristian
2526 Rodríguez, Daniel Buch, Dave Reisner, Frederic Crozat, Hannes
2527 Reinecke, Harald Hoyer, Jan Alexander Steffens, Jan
2528 Engelhardt, Josh Triplett, Kay Sievers, Lennart Poettering,
2529 Mantas Mikulėnas, Martin Pitt, Mathieu Bridon, Michael Biebl,
2530 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nathaniel Chen,
2531 Oleksii Shevchuk, Ozan Çağlayan, Thomas Hindoe Paaboel
2532 Andersen, Tollef Fog Heen, Tom Gundersen, Umut Tezduyar,
2533 Zbigniew Jędrzejewski-Szmek
2537 * Configuration of unit files may now be extended via drop-in
2538 files without having to edit/override the unit files
2539 themselves. More specifically, if the administrator wants to
2540 change one value for a service file foobar.service he can
2541 now do so by dropping in a configuration snippet into
2542 /etc/systemd/system/foobar.service.d/*.conf. The unit logic
2543 will load all these snippets and apply them on top of the
2544 main unit configuration file, possibly extending or
2545 overriding its settings. Using these drop-in snippets is
2546 generally nicer than the two earlier options for changing
2547 unit files locally: copying the files from
2548 /usr/lib/systemd/system/ to /etc/systemd/system/ and editing
2549 them there; or creating a new file in /etc/systemd/system/
2550 that incorporates the original one via ".include". Drop-in
2551 snippets into these .d/ directories can be placed in any
2552 directory systemd looks for units in, and the usual
2553 overriding semantics between /usr/lib, /etc and /run apply
2556 * Most unit file settings which take lists of items can now be
2557 reset by assigning the empty string to them. For example,
2558 normally, settings such as Environment=FOO=BAR append a new
2559 environment variable assignment to the environment block,
2560 each time they are used. By assigning Environment= the empty
2561 string the environment block can be reset to empty. This is
2562 particularly useful with the .d/*.conf drop-in snippets
2563 mentioned above, since this adds the ability to reset list
2564 settings from vendor unit files via these drop-ins.
2566 * systemctl gained a new "list-dependencies" command for
2567 listing the dependencies of a unit recursively.
2569 * Inhibitors are now honored and listed by "systemctl
2570 suspend", "systemctl poweroff" (and similar) too, not only
2571 GNOME. These commands will also list active sessions by
2574 * Resource limits (as exposed by the various control group
2575 controllers) can now be controlled dynamically at runtime
2576 for all units. More specifically, you can now use a command
2577 like "systemctl set-cgroup-attr foobar.service cpu.shares
2578 2000" to alter the CPU shares a specific service gets. These
2579 settings are stored persistently on disk, and thus allow the
2580 administrator to easily adjust the resource usage of
2581 services with a few simple commands. This dynamic resource
2582 management logic is also available to other programs via the
2583 bus. Almost any kernel cgroup attribute and controller is
2586 * systemd-vconsole-setup will now copy all font settings to
2587 all allocated VTs, where it previously applied them only to
2590 * libsystemd-login gained the new sd_session_get_tty() API
2593 * This release drops support for a few legacy or
2594 distribution-specific LSB facility names when parsing init
2595 scripts: $x-display-manager, $mail-transfer-agent,
2596 $mail-transport-agent, $mail-transfer-agent, $smtp,
2597 $null. Also, the mail-transfer-agent.target unit backing
2598 this has been removed. Distributions which want to retain
2599 compatibility with this should carry the burden for
2600 supporting this themselves and patch support for these back
2601 in, if they really need to. Also, the facilities $syslog and
2602 $local_fs are now ignored, since systemd does not support
2603 early-boot LSB init scripts anymore, and these facilities
2604 are implied anyway for normal services. syslog.target has
2607 * There are new bus calls on PID1's Manager object for
2608 cancelling jobs, and removing snapshot units. Previously,
2609 both calls were only available on the Job and Snapshot
2612 * systemd-journal-gatewayd gained SSL support.
2614 * The various "environment" files, such as /etc/locale.conf
2615 now support continuation lines with a backslash ("\") as
2616 last character in the line, similarly in style (but different)
2617 to how this is supported in shells.
2619 * For normal user processes the _SYSTEMD_USER_UNIT= field is
2620 now implicitly appended to every log entry logged. systemctl
2621 has been updated to filter by this field when operating on a
2622 user systemd instance.
2624 * nspawn will now implicitly add the CAP_AUDIT_WRITE and
2625 CAP_AUDIT_CONTROL capabilities to the capabilities set for
2626 the container. This makes it easier to boot unmodified
2627 Fedora systems in a container, which however still requires
2628 audit=0 to be passed on the kernel command line. Auditing in
2629 kernel and userspace is unfortunately still too broken in
2630 context of containers, hence we recommend compiling it out
2631 of the kernel or using audit=0. Hopefully this will be fixed
2632 one day for good in the kernel.
2634 * nspawn gained the new --bind= and --bind-ro= parameters to
2635 bind mount specific directories from the host into the
2638 * nspawn will now mount its own devpts file system instance
2639 into the container, in order not to leak pty devices from
2640 the host into the container.
2642 * systemd will now read the firmware boot time performance
2643 information from the EFI variables, if the used boot loader
2644 supports this, and takes it into account for boot performance
2645 analysis via "systemd-analyze". This is currently supported
2646 only in conjunction with Gummiboot, but could be supported
2647 by other boot loaders too. For details see:
2649 http://www.freedesktop.org/wiki/Software/systemd/BootLoaderInterface
2651 * A new generator has been added that automatically mounts the
2652 EFI System Partition (ESP) to /boot, if that directory
2653 exists, is empty, and no other file system has been
2654 configured to be mounted there.
2656 * logind will now send out PrepareForSleep(false) out
2657 unconditionally, after coming back from suspend. This may be
2658 used by applications as asynchronous notification for
2659 system resume events.
2661 * "systemctl unlock-sessions" has been added, that allows
2662 unlocking the screens of all user sessions at once, similar
2663 to how "systemctl lock-sessions" already locked all users
2664 sessions. This is backed by a new D-Bus call UnlockSessions().
2666 * "loginctl seat-status" will now show the master device of a
2667 seat. (i.e. the device of a seat that needs to be around for
2668 the seat to be considered available, usually the graphics
2671 * tmpfiles gained a new "X" line type, that allows
2672 configuration of files and directories (with wildcards) that
2673 shall be excluded from automatic cleanup ("aging").
2675 * udev default rules set the device node permissions now only
2676 at "add" events, and do not change them any longer with a
2677 later "change" event.
2679 * The log messages for lid events and power/sleep keypresses
2680 now carry a message ID.
2682 * We now have a substantially larger unit test suite, but this
2683 continues to be work in progress.
2685 * udevadm hwdb gained a new --root= parameter to change the
2686 root directory to operate relative to.
2688 * logind will now issue a background sync() request to the kernel
2689 early at shutdown, so that dirty buffers are flushed to disk early
2690 instead of at the last moment, in order to optimize shutdown
2693 * A new bootctl tool has been added that is an interface for
2694 certain boot loader operations. This is currently a preview
2695 and is likely to be extended into a small mechanism daemon
2696 like timedated, localed, hostnamed, and can be used by
2697 graphical UIs to enumerate available boot options, and
2698 request boot into firmware operations.
2700 * systemd-bootchart has been relicensed to LGPLv2.1+ to match
2701 the rest of the package. It also has been updated to work
2702 correctly in initrds.
2704 * Policykit previously has been runtime optional, and is now
2705 also compile time optional via a configure switch.
2707 * systemd-analyze has been reimplemented in C. Also "systemctl
2708 dot" has moved into systemd-analyze.
2710 * "systemctl status" with no further parameters will now print
2711 the status of all active or failed units.
2713 * Operations such as "systemctl start" can now be executed
2714 with a new mode "--irreversible" which may be used to queue
2715 operations that cannot accidentally be reversed by a later
2716 job queuing. This is by default used to make shutdown
2717 requests more robust.
2719 * The Python API of systemd now gained a new module for
2720 reading journal files.
2722 * A new tool kernel-install has been added that can install
2723 kernel images according to the Boot Loader Specification:
2725 http://www.freedesktop.org/wiki/Specifications/BootLoaderSpec
2727 * Boot time console output has been improved to provide
2728 animated boot time output for hanging jobs.
2730 * A new tool systemd-activate has been added which can be used
2731 to test socket activation with, directly from the command
2732 line. This should make it much easier to test and debug
2733 socket activation in daemons.
2735 * journalctl gained a new "--reverse" (or -r) option to show
2736 journal output in reverse order (i.e. newest line first).
2738 * journalctl gained a new "--pager-end" (or -e) option to jump
2739 to immediately jump to the end of the journal in the
2740 pager. This is only supported in conjunction with "less".
2742 * journalctl gained a new "--user-unit=" option, that works
2743 similarly to "--unit=" but filters for user units rather than
2746 * A number of unit files to ease adoption of systemd in
2747 initrds has been added. This moves some minimal logic from
2748 the various initrd implementations into systemd proper.
2750 * The journal files are now owned by a new group
2751 "systemd-journal", which exists specifically to allow access
2752 to the journal, and nothing else. Previously, we used the
2753 "adm" group for that, which however possibly covers more
2754 than just journal/log file access. This new group is now
2755 already used by systemd-journal-gatewayd to ensure this
2756 daemon gets access to the journal files and as little else
2757 as possible. Note that "make install" will also set FS ACLs
2758 up for /var/log/journal to give "adm" and "wheel" read
2759 access to it, in addition to "systemd-journal" which owns
2760 the journal files. We recommend that packaging scripts also
2761 add read access to "adm" + "wheel" to /var/log/journal, and
2762 all existing/future journal files. To normal users and
2763 administrators little changes, however packagers need to
2764 ensure to create the "systemd-journal" system group at
2765 package installation time.
2767 * The systemd-journal-gatewayd now runs as unprivileged user
2768 systemd-journal-gateway:systemd-journal-gateway. Packaging
2769 scripts need to create these system user/group at
2772 * timedated now exposes a new boolean property CanNTP that
2773 indicates whether a local NTP service is available or not.
2775 * systemd-detect-virt will now also detect xen PVs
2777 * The pstore file system is now mounted by default, if it is
2780 * In addition to the SELinux and IMA policies we will now also
2781 load SMACK policies at early boot.
2783 Contributions from: Adel Gadllah, Aleksander Morgado, Auke
2784 Kok, Ayan George, Bastien Nocera, Colin Walters, Daniel Buch,
2785 Daniel Wallace, Dave Reisner, David Herrmann, David Strauss,
2786 Eelco Dolstra, Enrico Scholz, Frederic Crozat, Harald Hoyer,
2787 Jan Janssen, Jonathan Callen, Kay Sievers, Lennart Poettering,
2788 Lukas Nykryn, Mantas Mikulėnas, Marc-Antoine Perennou, Martin
2789 Pitt, Mauro Dreissig, Max F. Albrecht, Michael Biebl, Michael
2790 Olbrich, Michal Schmidt, Michal Sekletar, Michal Vyskocil,
2791 Michał Bartoszkiewicz, Mirco Tischler, Nathaniel Chen, Nestor
2792 Ovroy, Oleksii Shevchuk, Paul W. Frields, Piotr Drąg, Rob
2793 Clark, Ryan Lortie, Simon McVittie, Simon Peeters, Steven
2794 Hiscocks, Thomas Hindoe Paaboel Andersen, Tollef Fog Heen, Tom
2795 Gundersen, Umut Tezduyar, William Giokas, Zbigniew
2796 Jędrzejewski-Szmek, Zeeshan Ali (Khattak)
2800 * Timer units now support calendar time events in addition to
2801 monotonic time events. That means you can now trigger a unit
2802 based on a calendar time specification such as "Thu,Fri
2803 2013-*-1,5 11:12:13" which refers to 11:12:13 of the first
2804 or fifth day of any month of the year 2013, given that it is
2805 a thursday or friday. This brings timer event support
2806 considerably closer to cron's capabilities. For details on
2807 the supported calendar time specification language see
2810 * udev now supports a number of different naming policies for
2811 network interfaces for predictable names, and a combination
2812 of these policies is now the default. Please see this wiki
2813 document for details:
2815 http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames
2817 * Auke Kok's bootchart implementation has been added to the
2818 systemd tree. It is an optional component that can graph the
2819 boot in quite some detail. It is one of the best bootchart
2820 implementations around and minimal in its code and
2823 * nss-myhostname has been integrated into the systemd source
2824 tree. nss-myhostname guarantees that the local hostname
2825 always stays resolvable via NSS. It has been a weak
2826 requirement of systemd-hostnamed since a long time, and
2827 since its code is actually trivial we decided to just
2828 include it in systemd's source tree. It can be turned off
2829 with a configure switch.
2831 * The read-ahead logic is now capable of properly detecting
2832 whether a btrfs file system is on SSD or rotating media, in
2833 order to optimize the read-ahead scheme. Previously, it was
2834 only capable of detecting this on traditional file systems
2837 * In udev, additional device properties are now read from the
2838 IAB in addition to the OUI database. Also, Bluetooth company
2839 identities are attached to the devices as well.
2841 * In service files %U may be used as specifier that is
2842 replaced by the configured user name of the service.
2844 * nspawn may now be invoked without a controlling TTY. This
2845 makes it suitable for invocation as its own service. This
2846 may be used to set up a simple containerized server system
2847 using only core OS tools.
2849 * systemd and nspawn can now accept socket file descriptors
2850 when they are started for socket activation. This enables
2851 implementation of socket activated nspawn
2852 containers. i.e. think about autospawning an entire OS image
2853 when the first SSH or HTTP connection is received. We expect
2854 that similar functionality will also be added to libvirt-lxc
2857 * journalctl will now suppress ANSI color codes when
2858 presenting log data.
2860 * systemctl will no longer show control group information for
2861 a unit if a the control group is empty anyway.
2863 * logind can now automatically suspend/hibernate/shutdown the
2866 * /etc/machine-info and hostnamed now also expose the chassis
2867 type of the system. This can be used to determine whether
2868 the local system is a laptop, desktop, handset or
2869 tablet. This information may either be configured by the
2870 user/vendor or is automatically determined from ACPI and DMI
2871 information if possible.
2873 * A number of PolicyKit actions are now bound together with
2874 "imply" rules. This should simplify creating UIs because
2875 many actions will now authenticate similar ones as well.
2877 * Unit files learnt a new condition ConditionACPower= which
2878 may be used to conditionalize a unit depending on whether an
2879 AC power source is connected or not, of whether the system
2880 is running on battery power.
2882 * systemctl gained a new "is-failed" verb that may be used in
2883 shell scripts and suchlike to check whether a specific unit
2884 is in the "failed" state.
2886 * The EnvironmentFile= setting in unit files now supports file
2887 globbing, and can hence be used to easily read a number of
2888 environment files at once.
2890 * systemd will no longer detect and recognize specific
2891 distributions. All distribution-specific #ifdeffery has been
2892 removed, systemd is now fully generic and
2893 distribution-agnostic. Effectively, not too much is lost as
2894 a lot of the code is still accessible via explicit configure
2895 switches. However, support for some distribution specific
2896 legacy configuration file formats has been dropped. We
2897 recommend distributions to simply adopt the configuration
2898 files everybody else uses now and convert the old
2899 configuration from packaging scripts. Most distributions
2900 already did that. If that's not possible or desirable,
2901 distributions are welcome to forward port the specific
2902 pieces of code locally from the git history.
2904 * When logging a message about a unit systemd will now always
2905 log the unit name in the message meta data.
2907 * localectl will now also discover system locale data that is
2908 not stored in locale archives, but directly unpacked.
2910 * logind will no longer unconditionally use framebuffer
2911 devices as seat masters, i.e. as devices that are required
2912 to be existing before a seat is considered preset. Instead,
2913 it will now look for all devices that are tagged as
2914 "seat-master" in udev. By default framebuffer devices will
2915 be marked as such, but depending on local systems other
2916 devices might be marked as well. This may be used to
2917 integrate graphics cards using closed source drivers (such
2918 as NVidia ones) more nicely into logind. Note however, that
2919 we recommend using the open source NVidia drivers instead,
2920 and no udev rules for the closed-source drivers will be
2921 shipped from us upstream.
2923 Contributions from: Adam Williamson, Alessandro Crismani, Auke
2924 Kok, Colin Walters, Daniel Wallace, Dave Reisner, David
2925 Herrmann, David Strauss, Dimitrios Apostolou, Eelco Dolstra,
2926 Eric Benoit, Giovanni Campagna, Hannes Reinecke, Henrik
2927 Grindal Bakken, Hermann Gausterer, Kay Sievers, Lennart
2928 Poettering, Lukas Nykryn, Mantas Mikulėnas, Marcel Holtmann,
2929 Martin Pitt, Matthew Monaco, Michael Biebl, Michael Terry,
2930 Michal Schmidt, Michal Sekletar, Michał Bartoszkiewicz, Oleg
2931 Samarin, Pekka Lundstrom, Philip Nilsson, Ramkumar
2932 Ramachandra, Richard Yao, Robert Millan, Sami Kerola, Shawn
2933 Landden, Thomas Hindoe Paaboel Andersen, Thomas Jarosch,
2934 Tollef Fog Heen, Tom Gundersen, Umut Tezduyar, Zbigniew
2939 * udev gained support for loading additional device properties
2940 from an indexed database that is keyed by vendor/product IDs
2941 and similar device identifiers. For the beginning this
2942 "hwdb" is populated with data from the well-known PCI and
2943 USB database, but also includes PNP, ACPI and OID data. In
2944 the longer run this indexed database shall grow into
2945 becoming the one central database for non-essential
2946 userspace device metadata. Previously, data from the PCI/USB
2947 database was only attached to select devices, since the
2948 lookup was a relatively expensive operation due to O(n) time
2949 complexity (with n being the number of entries in the
2950 database). Since this is now O(1), we decided to add in this
2951 data for all devices where this is available, by
2952 default. Note that the indexed database needs to be rebuilt
2953 when new data files are installed. To achieve this you need
2954 to update your packaging scripts to invoke "udevadm hwdb
2955 --update" after installation of hwdb data files. For
2956 RPM-based distributions we introduced the new
2957 %udev_hwdb_update macro for this purpose.
2959 * The Journal gained support for the "Message Catalog", an
2960 indexed database to link up additional information with
2961 journal entries. For further details please check:
2963 http://www.freedesktop.org/wiki/Software/systemd/catalog
2965 The indexed message catalog database also needs to be
2966 rebuilt after installation of message catalog files. Use
2967 "journalctl --update-catalog" for this. For RPM-based
2968 distributions we introduced the %journal_catalog_update
2969 macro for this purpose.
2971 * The Python Journal bindings gained support for the standard
2972 Python logging framework.
2974 * The Journal API gained new functions for checking whether
2975 the underlying file system of a journal file is capable of
2976 properly reporting file change notifications, or whether
2977 applications that want to reflect journal changes "live"
2978 need to recheck journal files continuously in appropriate
2981 * It is now possible to set the "age" field for tmpfiles
2982 entries to 0, indicating that files matching this entry
2983 shall always be removed when the directories are cleaned up.
2985 * coredumpctl gained a new "gdb" verb which invokes gdb
2986 right-away on the selected coredump.
2988 * There's now support for "hybrid sleep" on kernels that
2989 support this, in addition to "suspend" and "hibernate". Use
2990 "systemctl hybrid-sleep" to make use of this.
2992 * logind's HandleSuspendKey= setting (and related settings)
2993 now gained support for a new "lock" setting to simply
2994 request the screen lock on all local sessions, instead of
2995 actually executing a suspend or hibernation.
2997 * systemd will now mount the EFI variables file system by
3000 * Socket units now gained support for configuration of the
3001 SMACK security label.
3003 * timedatectl will now output the time of the last and next
3004 daylight saving change.
3006 * We dropped support for various legacy and distro-specific
3007 concepts, such as insserv, early-boot SysV services
3008 (i.e. those for non-standard runlevels such as 'b' or 'S')
3009 or ArchLinux /etc/rc.conf support. We recommend the
3010 distributions who still need support this to either continue
3011 to maintain the necessary patches downstream, or find a
3012 different solution. (Talk to us if you have questions!)
3014 * Various systemd components will now bypass PolicyKit checks
3015 for root and otherwise handle properly if PolicyKit is not
3016 found to be around. This should fix most issues for
3017 PolicyKit-less systems. Quite frankly this should have been
3018 this way since day one. It is absolutely our intention to
3019 make systemd work fine on PolicyKit-less systems, and we
3020 consider it a bug if something does not work as it should if
3021 PolicyKit is not around.
3023 * For embedded systems it is now possible to build udev and
3024 systemd without blkid and/or kmod support.
3026 * "systemctl switch-root" is now capable of switching root
3027 more than once. I.e. in addition to transitions from the
3028 initrd to the host OS it is now possible to transition to
3029 further OS images from the host. This is useful to implement
3030 offline updating tools.
3032 * Various other additions have been made to the RPM macros
3033 shipped with systemd. Use %udev_rules_update() after
3034 installing new udev rules files. %_udevhwdbdir,
3035 %_udevrulesdir, %_journalcatalogdir, %_tmpfilesdir,
3036 %_sysctldir are now available which resolve to the right
3037 directories for packages to place various data files in.
3039 * journalctl gained the new --full switch (in addition to
3040 --all, to disable ellipsation for long messages.
3042 Contributions from: Anders Olofsson, Auke Kok, Ben Boeckel,
3043 Colin Walters, Cosimo Cecchi, Daniel Wallace, Dave Reisner,
3044 Eelco Dolstra, Holger Hans Peter Freyther, Kay Sievers,
3045 Chun-Yi Lee, Lekensteyn, Lennart Poettering, Mantas Mikulėnas,
3046 Marti Raudsepp, Martin Pitt, Mauro Dreissig, Michael Biebl,
3047 Michal Schmidt, Michal Sekletar, Miklos Vajna, Nis Martensen,
3048 Oleksii Shevchuk, Olivier Brunel, Ramkumar Ramachandra, Thomas
3049 Bächler, Thomas Hindoe Paaboel Andersen, Tom Gundersen, Tony
3050 Camuso, Umut Tezduyar, Zbigniew Jędrzejewski-Szmek
3054 * journalctl gained new --since= and --until= switches to
3055 filter by time. It also now supports nice filtering for
3056 units via --unit=/-u.
3058 * Type=oneshot services may use ExecReload= and do the
3061 * The journal daemon now supports time-based rotation and
3062 vacuuming, in addition to the usual disk-space based
3065 * The journal will now index the available field values for
3066 each field name. This enables clients to show pretty drop
3067 downs of available match values when filtering. The bash
3068 completion of journalctl has been updated
3069 accordingly. journalctl gained a new switch -F to list all
3070 values a certain field takes in the journal database.
3072 * More service events are now written as structured messages
3073 to the journal, and made recognizable via message IDs.
3075 * The timedated, localed and hostnamed mini-services which
3076 previously only provided support for changing time, locale
3077 and hostname settings from graphical DEs such as GNOME now
3078 also have a minimal (but very useful) text-based client
3079 utility each. This is probably the nicest way to changing
3080 these settings from the command line now, especially since
3081 it lists available options and is fully integrated with bash
3084 * There's now a new tool "systemd-coredumpctl" to list and
3085 extract coredumps from the journal.
3087 * We now install a README each in /var/log/ and
3088 /etc/rc.d/init.d explaining where the system logs and init
3089 scripts went. This hopefully should help folks who go to
3090 that dirs and look into the otherwise now empty void and
3091 scratch their heads.
3093 * When user-services are invoked (by systemd --user) the
3094 $MANAGERPID env var is set to the PID of systemd.
3096 * SIGRTMIN+24 when sent to a --user instance will now result
3097 in immediate termination of systemd.
3099 * gatewayd received numerous feature additions such as a
3100 "follow" mode, for live syncing and filtering.
3102 * browse.html now allows filtering and showing detailed
3103 information on specific entries. Keyboard navigation and
3104 mouse screen support has been added.
3106 * gatewayd/journalctl now supports HTML5/JSON
3107 Server-Sent-Events as output.
3109 * The SysV init script compatibility logic will now
3110 heuristically determine whether a script supports the
3111 "reload" verb, and only then make this available as
3114 * "systemctl status --follow" has been removed, use "journalctl
3117 * journald.conf's RuntimeMinSize=, PersistentMinSize= settings
3118 have been removed since they are hardly useful to be
3121 * And I'd like to take the opportunity to specifically mention
3122 Zbigniew for his great contributions. Zbigniew, you rock!
3124 Contributions from: Andrew Eikum, Christian Hesse, Colin
3125 Guthrie, Daniel J Walsh, Dave Reisner, Eelco Dolstra, Ferenc
3126 Wágner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Mantas
3127 Mikulėnas, Martin Mikkelsen, Martin Pitt, Michael Olbrich,
3128 Michael Stapelberg, Michal Schmidt, Sebastian Ott, Thomas
3129 Bächler, Umut Tezduyar, Will Woods, Wulf C. Krueger, Zbigniew
3130 Jędrzejewski-Szmek, Сковорода Никита Андреевич
3134 * If /etc/vconsole.conf is non-existent or empty we will no
3135 longer load any console font or key map at boot by
3136 default. Instead the kernel defaults will be left
3137 intact. This is definitely the right thing to do, as no
3138 configuration should mean no configuration, and hard-coding
3139 font names that are different on all archs is probably a bad
3140 idea. Also, the kernel default key map and font should be
3141 good enough for most cases anyway, and mostly identical to
3142 the userspace fonts/key maps we previously overloaded them
3143 with. If distributions want to continue to default to a
3144 non-kernel font or key map they should ship a default
3145 /etc/vconsole.conf with the appropriate contents.
3147 Contributions from: Colin Walters, Daniel J Walsh, Dave
3148 Reisner, Kay Sievers, Lennart Poettering, Lukas Nykryn, Tollef
3149 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek
3153 * journalctl gained a new --cursor= switch to show entries
3154 starting from the specified location in the journal.
3156 * We now enforce a size limit on journal entry fields exported
3157 with "-o json" in journalctl. Fields larger than 4K will be
3158 assigned null. This can be turned off with --all.
3160 * An (optional) journal gateway daemon is now available as
3161 "systemd-journal-gatewayd.service". This service provides
3162 access to the journal via HTTP and JSON. This functionality
3163 will be used to implement live log synchronization in both
3164 pull and push modes, but has various other users too, such
3165 as easy log access for debugging of embedded devices. Right
3166 now it is already useful to retrieve the journal via HTTP:
3168 # systemctl start systemd-journal-gatewayd.service
3169 # wget http://localhost:19531/entries
3171 This will download the journal contents in a
3172 /var/log/messages compatible format. The same as JSON:
3174 # curl -H"Accept: application/json" http://localhost:19531/entries
3176 This service is also accessible via a web browser where a
3177 single static HTML5 app is served that uses the JSON logic
3178 to enable the user to do some basic browsing of the
3179 journal. This will be extended later on. Here's an example
3180 screenshot of this app in its current state:
3182 http://0pointer.de/public/journal-gatewayd
3184 Contributions from: Kay Sievers, Lennart Poettering, Robert
3185 Milasan, Tom Gundersen
3189 * The bash completion logic is now available for journalctl
3192 * We do not mount the "cpuset" controller anymore together with
3193 "cpu" and "cpuacct", as "cpuset" groups generally cannot be
3194 started if no parameters are assigned to it. "cpuset" hence
3195 broke code that assumed it it could create "cpu" groups and
3198 * journalctl -f will now subscribe to terminal size changes,
3199 and line break accordingly.
3201 Contributions from: Dave Reisner, Kay Sievers, Lennart
3202 Poettering, Lukas Nykrynm, Mirco Tischler, Václav Pavlín
3206 * nspawn will now create a symlink /etc/localtime in the
3207 container environment, copying the host's timezone
3208 setting. Previously this has been done via a bind mount, but
3209 since symlinks cannot be bind mounted this has now been
3210 changed to create/update the appropriate symlink.
3212 * journalctl -n's line number argument is now optional, and
3213 will default to 10 if omitted.
3215 * journald will now log the maximum size the journal files may
3216 take up on disk. This is particularly useful if the default
3217 built-in logic of determining this parameter from the file
3218 system size is used. Use "systemctl status
3219 systemd-journald.service" to see this information.
3221 * The multi-seat X wrapper tool has been stripped down. As X
3222 is now capable of enumerating graphics devices via udev in a
3223 seat-aware way the wrapper is not strictly necessary
3224 anymore. A stripped down temporary stop-gap is still shipped
3225 until the upstream display managers have been updated to
3226 fully support the new X logic. Expect this wrapper to be
3227 removed entirely in one of the next releases.
3229 * HandleSleepKey= in logind.conf has been split up into
3230 HandleSuspendKey= and HandleHibernateKey=. The old setting
3231 is not available anymore. X11 and the kernel are
3232 distuingishing between these keys and we should too. This
3233 also means the inhibition lock for these keys has been split
3236 Contributions from: Dave Airlie, Eelco Dolstra, Lennart
3237 Poettering, Lukas Nykryn, Václav Pavlín
3241 * Whenever a unit changes state we will now log this to the
3242 journal and show along the unit's own log output in
3245 * ConditionPathIsMountPoint= can now properly detect bind
3246 mount points too. (Previously, a bind mount of one file
3247 system to another place in the same file system could not be
3248 detected as mount, since they shared struct stat's st_dev
3251 * We will now mount the cgroup controllers cpu, cpuacct,
3252 cpuset and the controllers net_cls, net_prio together by
3255 * nspawn containers will now have a virtualized boot
3256 ID. (i.e. /proc/sys/kernel/random/boot_id is now mounted
3257 over with a randomized ID at container initialization). This
3258 has the effect of making "journalctl -b" do the right thing
3261 * The JSON output journal serialization has been updated not
3262 to generate "endless" list objects anymore, but rather one
3263 JSON object per line. This is more in line how most JSON
3264 parsers expect JSON objects. The new output mode
3265 "json-pretty" has been added to provide similar output, but
3266 neatly aligned for readability by humans.
3268 * We dropped all explicit sync() invocations in the shutdown
3269 code. The kernel does this implicitly anyway in the kernel
3270 reboot() syscall. halt(8)'s -n option is now a compatibility
3273 * We now support virtualized reboot() in containers, as
3274 supported by newer kernels. We will fall back to exit() if
3275 CAP_SYS_REBOOT is not available to the container. Also,
3276 nspawn makes use of this now and will actually reboot the
3277 container if the containerized OS asks for that.
3279 * journalctl will only show local log output by default
3280 now. Use --merge (-m) to show remote log output, too.
3282 * libsystemd-journal gained the new sd_journal_get_usage()
3283 call to determine the current disk usage of all journal
3284 files. This is exposed in the new "journalctl --disk-usage"
3287 * journald gained a new configuration setting SplitMode= in
3288 journald.conf which may be used to control how user journals
3289 are split off. See journald.conf(5) for details.
3291 * A new condition type ConditionFileNotEmpty= has been added.
3293 * tmpfiles' "w" lines now support file globbing, to write
3294 multiple files at once.
3296 * We added Python bindings for the journal submission
3297 APIs. More Python APIs for a number of selected APIs will
3298 likely follow. Note that we intend to add native bindings
3299 only for the Python language, as we consider it common
3300 enough to deserve bindings shipped within systemd. There are
3301 various projects outside of systemd that provide bindings
3302 for languages such as PHP or Lua.
3304 * Many conditions will now resolve specifiers such as %i. In
3305 addition, PathChanged= and related directives of .path units
3306 now support specifiers as well.
3308 * There's now a new RPM macro definition for the system preset
3311 * journald will now warn if it ca not forward a message to the
3312 syslog daemon because its socket is full.
3314 * timedated will no longer write or process /etc/timezone,
3315 except on Debian. As we do not support late mounted /usr
3316 anymore /etc/localtime always being a symlink is now safe,
3317 and hence the information in /etc/timezone is not necessary
3320 * logind will now always reserve one VT for a text getty (VT6
3321 by default). Previously if more than 6 X sessions where
3322 started they took up all the VTs with auto-spawned gettys,
3323 so that no text gettys were available anymore.
3325 * udev will now automatically inform the btrfs kernel logic
3326 about btrfs RAID components showing up. This should make
3327 simple hotplug based btrfs RAID assembly work.
3329 * PID 1 will now increase its RLIMIT_NOFILE to 64K by default
3330 (but not for its children which will stay at the kernel
3331 default). This should allow setups with a lot more listening
3334 * systemd will now always pass the configured timezone to the
3335 kernel at boot. timedated will do the same when the timezone
3338 * logind's inhibition logic has been updated. By default,
3339 logind will now handle the lid switch, the power and sleep
3340 keys all the time, even in graphical sessions. If DEs want
3341 to handle these events on their own they should take the new
3342 handle-power-key, handle-sleep-key and handle-lid-switch
3343 inhibitors during their runtime. A simple way to achiveve
3344 that is to invoke the DE wrapped in an invocation of:
3346 systemd-inhibit --what=handle-power-key:handle-sleep-key:handle-lid-switch ...
3348 * Access to unit operations is now checked via SELinux taking
3349 the unit file label and client process label into account.
3351 * systemd will now notify the administrator in the journal
3352 when he over-mounts a non-empty directory.
3354 * There are new specifiers that are resolved in unit files,
3355 for the host name (%H), the machine ID (%m) and the boot ID
3358 Contributions from: Allin Cottrell, Auke Kok, Brandon Philips,
3359 Colin Guthrie, Colin Walters, Daniel J Walsh, Dave Reisner,
3360 Eelco Dolstra, Jan Engelhardt, Kay Sievers, Lennart
3361 Poettering, Lucas De Marchi, Lukas Nykryn, Mantas Mikulėnas,
3362 Martin Pitt, Matthias Clasen, Michael Olbrich, Pierre Schmitz,
3363 Shawn Landden, Thomas Hindoe Paaboel Andersen, Tom Gundersen,
3364 Václav Pavlín, Yin Kangkai, Zbigniew Jędrzejewski-Szmek
3368 * Support for reading structured kernel messages from
3369 /dev/kmsg has now been added and is enabled by default.
3371 * Support for reading kernel messages from /proc/kmsg has now
3372 been removed. If you want kernel messages in the journal
3373 make sure to run a recent kernel (>= 3.5) that supports
3374 reading structured messages from /dev/kmsg (see
3375 above). /proc/kmsg is now exclusive property of classic
3376 syslog daemons again.
3378 * The libudev API gained the new
3379 udev_device_new_from_device_id() call.
3381 * The logic for file system namespace (ReadOnlyDirectory=,
3382 ReadWriteDirectoy=, PrivateTmp=) has been reworked not to
3383 require pivot_root() anymore. This means fewer temporary
3384 directories are created below /tmp for this feature.
3386 * nspawn containers will now see and receive all submounts
3387 made on the host OS below the root file system of the
3390 * Forward Secure Sealing is now supported for Journal files,
3391 which provide cryptographical sealing of journal files so
3392 that attackers cannot alter log history anymore without this
3393 being detectable. Lennart will soon post a blog story about
3394 this explaining it in more detail.
3396 * There are two new service settings RestartPreventExitStatus=
3397 and SuccessExitStatus= which allow configuration of exit
3398 status (exit code or signal) which will be excepted from the
3399 restart logic, resp. consider successful.
3401 * journalctl gained the new --verify switch that can be used
3402 to check the integrity of the structure of journal files and
3403 (if Forward Secure Sealing is enabled) the contents of
3406 * nspawn containers will now be run with /dev/stdin, /dev/fd/
3407 and similar symlinks pre-created. This makes running shells
3408 as container init process a lot more fun.
3410 * The fstab support can now handle PARTUUID= and PARTLABEL=
3413 * A new ConditionHost= condition has been added to match
3414 against the hostname (with globs) and machine ID. This is
3415 useful for clusters where a single OS image is used to
3416 provision a large number of hosts which shall run slightly
3417 different sets of services.
3419 * Services which hit the restart limit will now be placed in a
3422 Contributions from: Bertram Poettering, Dave Reisner, Huang
3423 Hang, Kay Sievers, Lennart Poettering, Lukas Nykryn, Martin
3424 Pitt, Simon Peeters, Zbigniew Jędrzejewski-Szmek
3428 * When running in --user mode systemd will now become a
3429 subreaper (PR_SET_CHILD_SUBREAPER). This should make the ps
3430 tree a lot more organized.
3432 * A new PartOf= unit dependency type has been introduced that
3433 may be used to group services in a natural way.
3435 * "systemctl enable" may now be used to enable instances of
3438 * journalctl now prints error log levels in red, and
3439 warning/notice log levels in bright white. It also supports
3440 filtering by log level now.
3442 * cgtop gained a new -n switch (similar to top), to configure
3443 the maximum number of iterations to run for. It also gained
3444 -b, to run in batch mode (accepting no input).
3446 * The suffix ".service" may now be omitted on most systemctl
3447 command lines involving service unit names.
3449 * There's a new bus call in logind to lock all sessions, as
3450 well as a loginctl verb for it "lock-sessions".
3452 * libsystemd-logind.so gained a new call sd_journal_perror()
3453 that works similar to libc perror() but logs to the journal
3454 and encodes structured information about the error number.
3456 * /etc/crypttab entries now understand the new keyfile-size=
3459 * shutdown(8) now can send a (configurable) wall message when
3460 a shutdown is cancelled.
3462 * The mount propagation mode for the root file system will now
3463 default to "shared", which is useful to make containers work
3464 nicely out-of-the-box so that they receive new mounts from
3465 the host. This can be undone locally by running "mount
3466 --make-rprivate /" if needed.
3468 * The prefdm.service file has been removed. Distributions
3469 should maintain this unit downstream if they intend to keep
3470 it around. However, we recommend writing normal unit files
3471 for display managers instead.
3473 * Since systemd is a crucial part of the OS we will now
3474 default to a number of compiler switches that improve
3475 security (hardening) such as read-only relocations, stack
3476 protection, and suchlike.
3478 * The TimeoutSec= setting for services is now split into
3479 TimeoutStartSec= and TimeoutStopSec= to allow configuration
3480 of individual time outs for the start and the stop phase of
3483 Contributions from: Artur Zaprzala, Arvydas Sidorenko, Auke
3484 Kok, Bryan Kadzban, Dave Reisner, David Strauss, Harald Hoyer,
3485 Jim Meyering, Kay Sievers, Lennart Poettering, Mantas
3486 Mikulėnas, Martin Pitt, Michal Schmidt, Michal Sekletar, Peter
3487 Alfredsen, Shawn Landden, Simon Peeters, Terence Honles, Tom
3488 Gundersen, Zbigniew Jędrzejewski-Szmek
3492 * The journal and id128 C APIs are now fully documented as man
3495 * Extra safety checks have been added when transitioning from
3496 the initial RAM disk to the main system to avoid accidental
3499 * /etc/crypttab entries now understand the new keyfile-offset=
3502 * systemctl -t can now be used to filter by unit load state.
3504 * The journal C API gained the new sd_journal_wait() call to
3505 make writing synchronous journal clients easier.
3507 * journalctl gained the new -D switch to show journals from a
3510 * journalctl now displays a special marker between log
3511 messages of two different boots.
3513 * The journal is now explicitly flushed to /var via a service
3514 systemd-journal-flush.service, rather than implicitly simply
3515 by seeing /var/log/journal to be writable.
3517 * journalctl (and the journal C APIs) can now match for much
3518 more complex expressions, with alternatives and
3521 * When transitioning from the initial RAM disk to the main
3522 system we will now kill all processes in a killing spree to
3523 ensure no processes stay around by accident.
3525 * Three new specifiers may be used in unit files: %u, %h, %s
3526 resolve to the user name, user home directory resp. user
3527 shell. This is useful for running systemd user instances.
3529 * We now automatically rotate journal files if their data
3530 object hash table gets a fill level > 75%. We also size the
3531 hash table based on the configured maximum file size. This
3532 together should lower hash collisions drastically and thus
3533 speed things up a bit.
3535 * journalctl gained the new "--header" switch to introspect
3536 header data of journal files.
3538 * A new setting SystemCallFilters= has been added to services
3539 which may be used to apply blacklists or whitelists to
3540 system calls. This is based on SECCOMP Mode 2 of Linux 3.5.
3542 * nspawn gained a new --link-journal= switch (and quicker: -j)
3543 to link the container journal with the host. This makes it
3544 very easy to centralize log viewing on the host for all
3545 guests while still keeping the journal files separated.
3547 * Many bugfixes and optimizations
3549 Contributions from: Auke Kok, Eelco Dolstra, Harald Hoyer, Kay
3550 Sievers, Lennart Poettering, Malte Starostik, Paul Menzel, Rex
3551 Tsai, Shawn Landden, Tom Gundersen, Ville Skyttä, Zbigniew
3556 * Several tools now understand kernel command line arguments,
3557 which are only read when run in an initial RAM disk. They
3558 usually follow closely their normal counterparts, but are
3561 * There's a new tool to analyze the readahead files that are
3562 automatically generated at boot. Use:
3564 /usr/lib/systemd/systemd-readahead analyze /.readahead
3566 * We now provide an early debug shell on tty9 if this enabled. Use:
3568 systemctl enable debug-shell.service
3570 * All plymouth related units have been moved into the Plymouth
3571 package. Please make sure to upgrade your Plymouth version
3574 * systemd-tmpfiles now supports getting passed the basename of
3575 a configuration file only, in which case it will look for it
3576 in all appropriate directories automatically.
3578 * udevadm info now takes a /dev or /sys path as argument, and
3579 does the right thing. Example:
3581 udevadm info /dev/sda
3582 udevadm info /sys/class/block/sda
3584 * systemctl now prints a warning if a unit is stopped but a
3585 unit that might trigger it continues to run. Example: a
3586 service is stopped but the socket that activates it is left
3589 * "systemctl status" will now mention if the log output was
3590 shortened due to rotation since a service has been started.
3592 * The journal API now exposes functions to determine the
3593 "cutoff" times due to rotation.
3595 * journald now understands SIGUSR1 and SIGUSR2 for triggering
3596 immediately flushing of runtime logs to /var if possible,
3597 resp. for triggering immediate rotation of the journal
3600 * It is now considered an error if a service is attempted to
3601 be stopped that is not loaded.
3603 * XDG_RUNTIME_DIR now uses numeric UIDs instead of usernames.
3605 * systemd-analyze now supports Python 3
3607 * tmpfiles now supports cleaning up directories via aging
3608 where the first level dirs are always kept around but
3609 directories beneath it automatically aged. This is enabled
3610 by prefixing the age field with '~'.
3612 * Seat objects now expose CanGraphical, CanTTY properties
3613 which is required to deal with very fast bootups where the
3614 display manager might be running before the graphics drivers
3615 completed initialization.
3617 * Seat objects now expose a State property.
3619 * We now include RPM macros for service enabling/disabling
3620 based on the preset logic. We recommend RPM based
3621 distributions to make use of these macros if possible. This
3622 makes it simpler to reuse RPM spec files across
3625 * We now make sure that the collected systemd unit name is
3626 always valid when services log to the journal via
3629 * There's a new man page kernel-command-line(7) detailing all
3630 command line options we understand.
3632 * The fstab generator may now be disabled at boot by passing
3633 fstab=0 on the kernel command line.
3635 * A new kernel command line option modules-load= is now understood
3636 to load a specific kernel module statically, early at boot.
3638 * Unit names specified on the systemctl command line are now
3639 automatically escaped as needed. Also, if file system or
3640 device paths are specified they are automatically turned
3641 into the appropriate mount or device unit names. Example:
3643 systemctl status /home
3644 systemctl status /dev/sda
3646 * The SysVConsole= configuration option has been removed from
3647 system.conf parsing.
3649 * The SysV search path is no longer exported on the D-Bus
3652 * The Names= option is been removed from unit file parsing.
3654 * There's a new man page bootup(7) detailing the boot process.
3656 * Every unit and every generator we ship with systemd now
3657 comes with full documentation. The self-explanatory boot is
3660 * A couple of services gained "systemd-" prefixes in their
3661 name if they wrap systemd code, rather than only external
3662 code. Among them fsck@.service which is now
3663 systemd-fsck@.service.
3665 * The HaveWatchdog property has been removed from the D-Bus
3668 * systemd.confirm_spawn= on the kernel command line should now
3671 * There's a new man page crypttab(5) which details all options
3672 we actually understand.
3674 * systemd-nspawn gained a new --capability= switch to pass
3675 additional capabilities to the container.
3677 * timedated will now read known NTP implementation unit names
3678 from /usr/lib/systemd/ntp-units.d/*.list,
3679 systemd-timedated-ntp.target has been removed.
3681 * journalctl gained a new switch "-b" that lists log data of
3682 the current boot only.
3684 * The notify socket is in the abstract namespace again, in
3685 order to support daemons which chroot() at start-up.
3687 * There is a new Storage= configuration option for journald
3688 which allows configuration of where log data should go. This
3689 also provides a way to disable journal logging entirely, so
3690 that data collected is only forwarded to the console, the
3691 kernel log buffer or another syslog implementation.
3693 * Many bugfixes and optimizations
3695 Contributions from: Auke Kok, Colin Guthrie, Dave Reisner,
3696 David Strauss, Eelco Dolstra, Kay Sievers, Lennart Poettering,
3697 Lukas Nykryn, Michal Schmidt, Michal Sekletar, Paul Menzel,
3698 Shawn Landden, Tom Gundersen
3702 * "systemctl help <unit>" now shows the man page if one is
3705 * Several new man pages have been added.
3707 * MaxLevelStore=, MaxLevelSyslog=, MaxLevelKMsg=,
3708 MaxLevelConsole= can now be specified in
3709 journald.conf. These options allow reducing the amount of
3710 data stored on disk or forwarded by the log level.
3712 * TimerSlackNSec= can now be specified in system.conf for
3713 PID1. This allows system-wide power savings.
3715 Contributions from: Dave Reisner, Kay Sievers, Lauri Kasanen,
3716 Lennart Poettering, Malte Starostik, Marc-Antoine Perennou,
3721 * logind is now capable of (optionally) handling power and
3722 sleep keys as well as the lid switch.
3724 * journalctl now understands the syntax "journalctl
3725 /usr/bin/avahi-daemon" to get all log output of a specific
3728 * CapabilityBoundingSet= in system.conf now also influences
3729 the capability bound set of usermode helpers of the kernel.
3731 Contributions from: Daniel Drake, Daniel J. Walsh, Gert
3732 Michael Kulyk, Harald Hoyer, Jean Delvare, Kay Sievers,
3733 Lennart Poettering, Matthew Garrett, Matthias Clasen, Paul
3734 Menzel, Shawn Landden, Tero Roponen, Tom Gundersen
3738 * Note that we skipped 139 releases here in order to set the
3739 new version to something that is greater than both udev's
3740 and systemd's most recent version number.
3742 * udev: all udev sources are merged into the systemd source tree now.
3743 All future udev development will happen in the systemd tree. It
3744 is still fully supported to use the udev daemon and tools without
3745 systemd running, like in initramfs or other init systems. Building
3746 udev though, will require the *build* of the systemd tree, but
3747 udev can be properly *run* without systemd.
3749 * udev: /lib/udev/devices/ are not read anymore; systemd-tmpfiles
3750 should be used to create dead device nodes as workarounds for broken
3753 * udev: RUN+="socket:..." and udev_monitor_new_from_socket() is
3754 no longer supported. udev_monitor_new_from_netlink() needs to be
3755 used to subscribe to events.
3757 * udev: when udevd is started by systemd, processes which are left
3758 behind by forking them off of udev rules, are unconditionally cleaned
3759 up and killed now after the event handling has finished. Services or
3760 daemons must be started as systemd services. Services can be
3761 pulled-in by udev to get started, but they can no longer be directly
3762 forked by udev rules.
3764 * udev: the daemon binary is called systemd-udevd now and installed
3765 in /usr/lib/systemd/. Standalone builds or non-systemd systems need
3766 to adapt to that, create symlink, or rename the binary after building
3769 * libudev no longer provides these symbols:
3770 udev_monitor_from_socket()
3771 udev_queue_get_failed_list_entry()
3772 udev_get_{dev,sys,run}_path()
3773 The versions number was bumped and symbol versioning introduced.
3775 * systemd-loginctl and systemd-journalctl have been renamed
3776 to loginctl and journalctl to match systemctl.
3778 * The config files: /etc/systemd/systemd-logind.conf and
3779 /etc/systemd/systemd-journald.conf have been renamed to
3780 logind.conf and journald.conf. Package updates should rename
3781 the files to the new names on upgrade.
3783 * For almost all files the license is now LGPL2.1+, changed
3784 from the previous GPL2.0+. Exceptions are some minor stuff
3785 of udev (which will be changed to LGPL2.1 eventually, too),
3786 and the MIT licensed sd-daemon.[ch] library that is suitable
3787 to be used as drop-in files.
3789 * systemd and logind now handle system sleep states, in
3790 particular suspending and hibernating.
3792 * logind now implements a sleep/shutdown/idle inhibiting logic
3793 suitable for a variety of uses. Soonishly Lennart will blog
3794 about this in more detail.
3796 * var-run.mount and var-lock.mount are no longer provided
3797 (which prevously bind mounted these directories to their new
3798 places). Distributions which have not converted these
3799 directories to symlinks should consider stealing these files
3800 from git history and add them downstream.
3802 * We introduced the Documentation= field for units and added
3803 this to all our shipped units. This is useful to make it
3804 easier to explore the boot and the purpose of the various
3807 * All smaller setup units (such as
3808 systemd-vconsole-setup.service) now detect properly if they
3809 are run in a container and are skipped when
3810 appropriate. This guarantees an entirely noise-free boot in
3811 Linux container environments such as systemd-nspawn.
3813 * A framework for implementing offline system updates is now
3814 integrated, for details see:
3815 http://freedesktop.org/wiki/Software/systemd/SystemUpdates
3817 * A new service type Type=idle is available now which helps us
3818 avoiding ugly interleaving of getty output and boot status
3821 * There's now a system-wide CapabilityBoundingSet= option to
3822 globally reduce the set of capabilities for the
3823 system. This is useful to drop CAP_SYS_MKNOD, CAP_SYS_RAWIO,
3824 CAP_NET_RAW, CAP_SYS_MODULE, CAP_SYS_TIME, CAP_SYS_PTRACE or
3825 even CAP_NET_ADMIN system-wide for secure systems.
3827 * There are now system-wide DefaultLimitXXX= options to
3828 globally change the defaults of the various resource limits
3829 for all units started by PID 1.
3831 * Harald Hoyer's systemd test suite has been integrated into
3832 systemd which allows easy testing of systemd builds in qemu
3833 and nspawn. (This is really awesome! Ask us for details!)
3835 * The fstab parser is now implemented as generator, not inside
3838 * systemctl will now warn you if .mount units generated from
3839 /etc/fstab are out of date due to changes in fstab that
3840 have not been read by systemd yet.
3842 * systemd is now suitable for usage in initrds. Dracut has
3843 already been updated to make use of this. With this in place
3844 initrds get a slight bit faster but primarily are much
3845 easier to introspect and debug since "systemctl status" in
3846 the host system can be used to introspect initrd services,
3847 and the journal from the initrd is kept around too.
3849 * systemd-delta has been added, a tool to explore differences
3850 between user/admin configuration and vendor defaults.
3852 * PrivateTmp= now affects both /tmp and /var/tmp.
3854 * Boot time status messages are now much prettier and feature
3855 proper english language. Booting up systemd has never been
3858 * Read-ahead pack files now include the inode number of all
3859 files to pre-cache. When the inode changes the pre-caching
3860 is not attempted. This should be nicer to deal with updated
3861 packages which might result in changes of read-ahead
3864 * We now temporaritly lower the kernel's read_ahead_kb variable
3865 when collecting read-ahead data to ensure the kernel's
3866 built-in read-ahead does not add noise to our measurements
3867 of necessary blocks to pre-cache.
3869 * There's now RequiresMountsFor= to add automatic dependencies
3870 for all mounts necessary for a specific file system path.
3872 * MountAuto= and SwapAuto= have been removed from
3873 system.conf. Mounting file systems at boot has to take place
3876 * nspawn now learned a new switch --uuid= to set the machine
3877 ID on the command line.
3879 * nspawn now learned the -b switch to automatically search
3882 * vt102 is now the default TERM for serial TTYs, upgraded from
3885 * systemd-logind now works on VT-less systems.
3887 * The build tree has been reorganized. The individual
3888 components now have directories of their own.
3890 * A new condition type ConditionPathIsReadWrite= is now available.
3892 * nspawn learned the new -C switch to create cgroups for the
3893 container in other hierarchies.
3895 * We now have support for hardware watchdogs, configurable in
3898 * The scheduled shutdown logic now has a public API.
3900 * We now mount /tmp as tmpfs by default, but this can be
3901 masked and /etc/fstab can override it.
3903 * Since udisks does not make use of /media anymore we are not
3904 mounting a tmpfs on it anymore.
3906 * journalctl gained a new --local switch to only interleave
3907 locally generated journal files.
3909 * We can now load the IMA policy at boot automatically.
3911 * The GTK tools have been split off into a systemd-ui.
3913 Contributions from: Andreas Schwab, Auke Kok, Ayan George,
3914 Colin Guthrie, Daniel Mack, Dave Reisner, David Ward, Elan
3915 Ruusamäe, Frederic Crozat, Gergely Nagy, Guillermo Vidal,
3916 Hannes Reinecke, Harald Hoyer, Javier Jardón, Kay Sievers,
3917 Lennart Poettering, Lucas De Marchi, Léo Gillot-Lamure,
3918 Marc-Antoine Perennou, Martin Pitt, Matthew Monaco, Maxim
3919 A. Mikityanskiy, Michael Biebl, Michael Olbrich, Michal
3920 Schmidt, Nis Martensen, Patrick McCarty, Roberto Sassu, Shawn
3921 Landden, Sjoerd Simons, Sven Anders, Tollef Fog Heen, Tom
3926 * This is mostly a bugfix release
3928 * Support optional initialization of the machine ID from the
3929 KVM or container configured UUID.
3931 * Support immediate reboots with "systemctl reboot -ff"
3933 * Show /etc/os-release data in systemd-analyze output
3935 * Many bugfixes for the journal, including endianness fixes and
3936 ensuring that disk space enforcement works
3938 * sd-login.h is C++ comptaible again
3940 * Extend the /etc/os-release format on request of the Debian
3943 * We now refuse non-UTF8 strings used in various configuration
3944 and unit files. This is done to ensure we do not pass invalid
3945 data over D-Bus or expose it elsewhere.
3947 * Register Mimo USB Screens as suitable for automatic seat
3950 * Read SELinux client context from journal clients in a race
3953 * Reorder configuration file lookup order. /etc now always
3954 overrides /run in order to allow the administrator to always
3955 and unconditionally override vendor supplied or
3956 automatically generated data.
3958 * The various user visible bits of the journal now have man
3959 pages. We still lack man pages for the journal API calls
3962 * We now ship all man pages in HTML format again in the
3965 Contributions from: Dave Reisner, Dirk Eibach, Frederic
3966 Crozat, Harald Hoyer, Kay Sievers, Lennart Poettering, Marti
3967 Raudsepp, Michal Schmidt, Shawn Landden, Tero Roponen, Thierry
3972 * This is mostly a bugfix release
3974 * systems lacking /etc/os-release are no longer supported.
3976 * Various functionality updates to libsystemd-login.so
3978 * Track class of PAM logins to distuingish greeters from
3981 Contributions from: Kay Sievers, Lennart Poettering, Michael
3986 * This is an important bugfix release for v41.
3988 * Building man pages is now optional which should be useful
3989 for those building systemd from git but unwilling to install
3992 * Watchdog support for supervising services is now usable. In
3993 a future release support for hardware watchdogs
3994 (i.e. /dev/watchdog) will be added building on this.
3996 * Service start rate limiting is now configurable and can be
3997 turned off per service. When a start rate limit is hit a
3998 reboot can automatically be triggered.
4000 * New CanReboot(), CanPowerOff() bus calls in systemd-logind.
4002 Contributions from: Benjamin Franzke, Bill Nottingham,
4003 Frederic Crozat, Lennart Poettering, Michael Olbrich, Michal
4004 Schmidt, Michał Górny, Piotr Drąg
4008 * The systemd binary is installed /usr/lib/systemd/systemd now;
4009 An existing /sbin/init symlink needs to be adapted with the
4012 * The code that loads kernel modules has been ported to invoke
4013 libkmod directly, instead of modprobe. This means we do not
4014 support systems with module-init-tools anymore.
4016 * Watchdog support is now already useful, but still not
4019 * A new kernel command line option systemd.setenv= is
4020 understood to set system wide environment variables
4021 dynamically at boot.
4023 * We now limit the set of capabilities of systemd-journald.
4025 * We now set SIGPIPE to ignore by default, since it only is
4026 useful in shell pipelines, and has little use in general
4027 code. This can be disabled with IgnoreSIPIPE=no in unit
4030 Contributions from: Benjamin Franzke, Kay Sievers, Lennart
4031 Poettering, Michael Olbrich, Michal Schmidt, Tom Gundersen,
4036 * This is mostly a bugfix release
4038 * We now expose the reason why a service failed in the
4039 "Result" D-Bus property.
4041 * Rudimentary service watchdog support (will be completed over
4042 the next few releases.)
4044 * When systemd forks off in order execute some service we will
4045 now immediately changes its argv[0] to reflect which process
4046 it will execute. This is useful to minimize the time window
4047 with a generic argv[0], which makes bootcharts more useful
4049 Contributions from: Alvaro Soliverez, Chris Paulson-Ellis, Kay
4050 Sievers, Lennart Poettering, Michael Olbrich, Michal Schmidt,
4051 Mike Kazantsev, Ray Strode
4055 * This is mostly a test release, but incorporates many
4058 * New systemd-cgtop tool to show control groups by their
4061 * Linking against libacl for ACLs is optional again. If
4062 disabled, support tracking device access for active logins
4063 goes becomes unavailable, and so does access to the user
4064 journals by the respective users.
4066 * If a group "adm" exists, journal files are automatically
4067 owned by them, thus allow members of this group full access
4068 to the system journal as well as all user journals.
4070 * The journal now stores the SELinux context of the logging
4071 client for all entries.
4073 * Add C++ inclusion guards to all public headers
4075 * New output mode "cat" in the journal to print only text
4076 messages, without any meta data like date or time.
4078 * Include tiny X server wrapper as a temporary stop-gap to
4079 teach XOrg udev display enumeration. This is used by display
4080 managers such as gdm, and will go away as soon as XOrg
4081 learned native udev hotplugging for display devices.
4083 * Add new systemd-cat tool for executing arbitrary programs
4084 with STDERR/STDOUT connected to the journal. Can also act as
4085 BSD logger replacement, and does so by default.
4087 * Optionally store all locally generated coredumps in the
4088 journal along with meta data.
4090 * systemd-tmpfiles learnt four new commands: n, L, c, b, for
4091 writing short strings to files (for usage for /sys), and for
4092 creating symlinks, character and block device nodes.
4094 * New unit file option ControlGroupPersistent= to make cgroups
4095 persistent, following the mechanisms outlined in
4096 http://www.freedesktop.org/wiki/Software/systemd/PaxControlGroups
4098 * Support multiple local RTCs in a sane way
4100 * No longer monopolize IO when replaying readahead data on
4101 rotating disks, since we might starve non-file-system IO to
4102 death, since fanotify() will not see accesses done by blkid,
4105 * Do not show kernel threads in systemd-cgls anymore, unless
4106 requested with new -k switch.
4108 Contributions from: Dan Horák, Kay Sievers, Lennart
4109 Poettering, Michal Schmidt
4113 * This is mostly a test release, but incorporates many
4116 * The git repository moved to:
4117 git://anongit.freedesktop.org/systemd/systemd
4118 ssh://git.freedesktop.org/git/systemd/systemd
4120 * First release with the journal
4121 http://0pointer.de/blog/projects/the-journal.html
4123 * The journal replaces both systemd-kmsg-syslogd and
4124 systemd-stdout-bridge.
4126 * New sd_pid_get_unit() API call in libsystemd-logind
4128 * Many systemadm clean-ups
4130 * Introduce remote-fs-pre.target which is ordered before all
4131 remote mounts and may be used to start services before all
4134 * Added Mageia support
4136 * Add bash completion for systemd-loginctl
4138 * Actively monitor PID file creation for daemons which exit in
4139 the parent process before having finished writing the PID
4140 file in the daemon process. Daemons which do this need to be
4141 fixed (i.e. PID file creation must have finished before the
4142 parent exits), but we now react a bit more gracefully to them.
4144 * Add colourful boot output, mimicking the well-known output
4145 of existing distributions.
4147 * New option PassCredentials= for socket units, for
4148 compatibility with a recent kernel ABI breakage.
4150 * /etc/rc.local is now hooked in via a generator binary, and
4151 thus will no longer act as synchronization point during
4154 * systemctl list-unit-files now supports --root=.
4156 * systemd-tmpfiles now understands two new commands: z, Z for
4157 relabelling files according to the SELinux database. This is
4158 useful to apply SELinux labels to specific files in /sys,
4161 * Output of SysV services is now forwarded to both the console
4162 and the journal by default, not only just the console.
4164 * New man pages for all APIs from libsystemd-login.
4166 * The build tree got reorganized and a the build system is a
4167 lot more modular allowing embedded setups to specifically
4168 select the components of systemd they are interested in.
4170 * Support for Linux systems lacking the kernel VT subsystem is
4173 * configure's --with-rootdir= got renamed to
4174 --with-rootprefix= to follow the naming used by udev and
4177 * Unless specified otherwise we will now install to /usr instead
4178 of /usr/local by default.
4180 * Processes with '@' in argv[0][0] are now excluded from the
4181 final shut-down killing spree, following the logic explained
4183 http://www.freedesktop.org/wiki/Software/systemd/RootStorageDaemons
4185 * All processes remaining in a service cgroup when we enter
4186 the START or START_PRE states are now killed with
4187 SIGKILL. That means it is no longer possible to spawn
4188 background processes from ExecStart= lines (which was never
4189 supported anyway, and bad style).
4191 * New PropagateReloadTo=/PropagateReloadFrom= options to bind
4192 reloading of units together.
4194 Contributions from: Bill Nottingham, Daniel J. Walsh, Dave
4195 Reisner, Dexter Morgan, Gregs Gregs, Jonathan Nieder, Kay
4196 Sievers, Lennart Poettering, Michael Biebl, Michal Schmidt,
4197 Michał Górny, Ran Benita, Thomas Jarosch, Tim Waugh, Tollef
4198 Fog Heen, Tom Gundersen, Zbigniew Jędrzejewski-Szmek