Copyright 2010 Lennart Poettering
systemd is free software; you can redistribute it and/or modify it
- under the terms of the GNU General Public License as published by
- the Free Software Foundation; either version 2 of the License, or
+ under the terms of the GNU Lesser General Public License as published by
+ the Free Software Foundation; either version 2.1 of the License, or
(at your option) any later version.
systemd is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- General Public License for more details.
+ Lesser General Public License for more details.
- You should have received a copy of the GNU General Public License
+ You should have received a copy of the GNU Lesser General Public License
along with systemd; If not, see <http://www.gnu.org/licenses/>.
-->
<refnamediv>
<refname>systemd.service</refname>
- <refpurpose>systemd service configuration files</refpurpose>
+ <refpurpose>Service unit configuration</refpurpose>
</refnamediv>
<refsynopsisdiv>
<para>Additional options are listed in
<citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
which define the execution environment the commands
- are executed in.</para>
+ are executed in, and in
+ <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>
+ which define the way the processes of the service are
+ terminated.</para>
<para>Unless <varname>DefaultDependencies=</varname>
is set to <option>false</option>, service units will
<filename>.service</filename> suffix removed) and
dynamically creates a service unit from that
script. This is useful for compatibility with
- SysV.</para>
+ SysV. Note that this compatibility is quite
+ comprehensive but not 100%. For details about the
+ incomptibilities see the <ulink
+ url="http://www.freedesktop.org/wiki/Software/systemd/Incompatibilities">Incompatibilities
+ with SysV</ulink> document.
+ </para>
</refsect1>
<refsect1>
supervises. A number of options that may be used in
this section are shared with other unit types. These
options are documented in
- <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>. The
+ <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>
+ and
+ <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>. The
options specific to the <literal>[Service]</literal>
section of service units are the following:</para>
<option>forking</option>,
<option>oneshot</option>,
<option>dbus</option>,
- <option>notify</option>.</para>
+ <option>notify</option> or
+ <option>idle</option>.</para>
<para>If set to
<option>simple</option> (the default
- value) it is expected that the process
- configured with
+ value if <varname>BusName=</varname>
+ is not specified) it is expected that
+ the process configured with
<varname>ExecStart=</varname> is the
main process of the service. In this
mode, if the process offers
acquired. Service units with this
option configured implicitly gain
dependencies on the
- <filename>dbus.target</filename>
- unit.</para>
+ <filename>dbus.socket</filename>
+ unit. This type is the default if
+ <varname>BusName=</varname> is
+ specified.</para>
<para>Behaviour of
<option>notify</option> is similar to
below) should be set to open access to
the notification socket provided by
systemd. If
- <varname>NotifyAccess=</varname> is not
- set, it will implicitly be set to
+ <varname>NotifyAccess=</varname> is
+ not set, it will be implicitly set to
<option>main</option>.</para>
+
+ <para>Behaviour of
+ <option>idle</option> is very similar
+ to <option>simple</option>, however
+ actual execution of a the service
+ binary is delayed until all jobs are
+ dispatched. This may be used to avoid
+ interleaving of output of shell
+ services with the status output on the
+ console.</para>
</listitem>
</varlistentry>
</listitem>
</varlistentry>
+ <varlistentry>
+ <term><varname>GuessMainPID=</varname></term>
+
+ <listitem><para>Takes a boolean value
+ that specifies whether systemd should
+ try to guess the main PID of a service
+ should if it cannot be determined
+ reliably. This option is ignored
+ unless <option>Type=forking</option>
+ is set and <option>PIDFile=</option>
+ is unset because for the other types
+ or with an explicitly configured PID
+ file the main PID is always known. The
+ guessing algorithm might come to
+ incorrect conclusions if a daemon
+ consists of more than one process. If
+ the main PID cannot be determined
+ failure detection and automatic
+ restarting of a service will not work
+ reliably. Defaults to
+ <option>yes</option>.</para>
+ </listitem>
+ </varlistentry>
+
<varlistentry>
<term><varname>PIDFile=</varname></term>
daemon. Use of this option is
recommended for services where
<varname>Type=</varname> is set to
- <option>forking</option>.</para>
+ <option>forking</option>. systemd will
+ read the PID of the main process of
+ the daemon after start-up of the
+ service. systemd will not write to the
+ file configured here.</para>
</listitem>
</varlistentry>
<literal>-</literal> an exit code of
the command normally considered a
failure (i.e. non-zero exit status or
- abormal exit due to signal) is ignored
+ abnormal exit due to signal) is ignored
and considered success. If both
<literal>-</literal> and
<literal>@</literal> are used for the
- same command the former must preceed
+ same command the former must precede
the latter. Unless
<varname>Type=forking</varname> is
set, the process started via this
main process of the daemon. The
command line accepts % specifiers as
described in
- <citerefentry><refentrytitle>systemd.unit</refentrytitle><manvolnum>5</manvolnum></citerefentry>. On
- top of that basic environment variable
- substitution is supported, where
- <literal>${FOO}</literal> is replaced
- by the string value of the environment
- variable of the same name. Also
- <literal>$FOO</literal> may appear as
- separate word on the command line in
- which case the variable is replaced by
- its value split at whitespaces. Note
- that the first argument (i.e. the
- binary to execute) may not be a
- variable, and must be a literal and
- absolute path name.</para></listitem>
+ <citerefentry><refentrytitle>systemd.unit</refentrytitle><manvolnum>5</manvolnum></citerefentry>.</para>
+
+ <para>On top of that basic environment
+ variable substitution is
+ supported. Use
+ <literal>${FOO}</literal> as part of a
+ word, or as word of its own on the
+ command line, in which case it will be
+ replaced by the value of the
+ environment variable including all
+ whitespace it contains, resulting in a
+ single argument. Use
+ <literal>$FOO</literal> as a separate
+ word on the command line, in which
+ case it will be replaced by the value
+ of the environment variable split up
+ at whitespace, resulting in no or more
+ arguments. Note that the first
+ argument (i.e. the program to execute)
+ may not be a variable, and must be a
+ literal and absolute path
+ name.</para></listitem>
</varlistentry>
<varlistentry>
the command in
<varname>ExecStart=</varname>. Multiple
command lines may be concatenated in a
- single directive, by seperating them
+ single directive, by separating them
by semicolons (these semicolons must
be passed as separate words). In that
case, the commands are executed one
after the other,
serially. Alternatively, these
directives may be specified more than
- once whith the same effect. However,
+ once with the same effect. However,
the latter syntax is not recommended
for compatibility with parsers
suitable for XDG
daemon, and may be used for command
lines like the following:
<command>/bin/kill -HUP
- $(MAINPID)</command>.</para></listitem>
+ $MAINPID</command>.</para></listitem>
</varlistentry>
<varlistentry>
configured in this option are run are
terminated according to the
<varname>KillMode=</varname> setting
- (see below). If this option is not
- specified the process is terminated
- right-away when service stop is
- requested. Specifier and environment
- variable substitution is supported
- (including
- <literal>$(MAINPID)</literal>, see
+ (see
+ <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>). If
+ this option is not specified the
+ process is terminated right-away when
+ service stop is requested. Specifier
+ and environment variable substitution
+ is supported (including
+ <literal>$MAINPID</literal>, see
above).</para></listitem>
</varlistentry>
another delay of this time with
SIGKILL. (See
<varname>KillMode=</varname>
- below.) Takes a unit-less value in seconds, or a
+ in <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>) Takes a unit-less value in seconds, or a
time span value such as "5min
20s". Pass 0 to disable the timeout
logic. Defaults to
- 60s.</para></listitem>
+ 90s, except when <varname>Type=oneshot</varname> is
+ used in which case the timeout
+ is disabled by default.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>WatchdogSec=</varname></term>
+ <listitem><para>Configures the
+ watchdog timeout for a service. This
+ is activated when the start-up is
+ completed. The service must call
+ <citerefentry><refentrytitle>sd_notify</refentrytitle><manvolnum>3</manvolnum></citerefentry>
+ regularly with "WATCHDOG=1" (i.e. the
+ "keep-alive ping"). If the time
+ between two such calls is larger than
+ the configured time then the service
+ is placed in a failure state. By
+ setting <varname>Restart=</varname> to
+ <option>on-failure</option> or
+ <option>always</option> the service
+ will be automatically restarted. The
+ time configured here will be passed to
+ the executed service process in the
+ <varname>WATCHDOG_USEC=</varname>
+ environment variable. This allows
+ daemons to automatically enable the
+ keep-alive pinging logic if watchdog
+ support is enabled for the service. If
+ this option is used
+ <varname>NotifyAccess=</varname> (see
+ below) should be set to open access to
+ the notification socket provided by
+ systemd. If
+ <varname>NotifyAccess=</varname> is
+ not set, it will be implicitly set to
+ <option>main</option>. Defaults to 0,
+ which disables this
+ feature.</para></listitem>
</varlistentry>
<varlistentry>
<term><varname>Restart=</varname></term>
<listitem><para>Configures whether the
- main service process shall be restarted when
- it exists. Takes one of
+ main service process shall be
+ restarted when it exits. Takes one of
<option>no</option>,
- <option>on-success</option> or
- <option>always</option>. If
- set to <option>no</option> (the
- default) the service will not be
- restarted when it exits. If set to
- <option>on-success</option> it
- will be restarted only when it exited
- cleanly, i.e. terminated with an exit
- code of 0. If set to
- <option>always</option> the
- service will be restarted regardless
- whether it exited cleanly or not, or
- got terminated abnormally by a
- signal.</para></listitem>
+ <option>on-success</option>,
+ <option>on-failure</option>,
+ <option>on-abort</option> or
+ <option>always</option>. If set to
+ <option>no</option> (the default) the
+ service will not be restarted when it
+ exits. If set to
+ <option>on-success</option> it will be
+ restarted only when it exited cleanly,
+ i.e. terminated with an exit code of
+ 0. If set to
+ <option>on-failure</option> it will be
+ restarted only when it exited with an
+ exit code not equalling 0, when
+ terminated by a signal (including on
+ core dump), when an operation (such as
+ service reload) times out or when the
+ configured watchdog timeout is
+ triggered. If set to
+ <option>on-abort</option> it will be
+ restarted only if it exits due to
+ reception of an uncaught signal
+ (including on core dump). If set to
+ <option>always</option> the service
+ will be restarted regardless whether
+ it exited cleanly or not, got
+ terminated abnormally by a signal or
+ hit a timeout.</para></listitem>
</varlistentry>
<varlistentry>
false.</para></listitem>
</varlistentry>
- <varlistentry>
- <term><varname>SysVStartPriority=</varname></term>
- <listitem><para>Set the SysV start
- priority to use to order this service
- in relation to SysV services lacking
- LSB headers. This option is only
- necessary to fix ordering in relation
- to legacy SysV services, that have no
- ordering information encoded in the
- script headers. As such it should only
- be used as temporary compatibility
- option, and not be used in new unit
- files. Almost always it is a better
- choice to add explicit ordering
- directives via
- <varname>After=</varname> or
- <varname>Before=</varname>,
- instead. For more details see
- <citerefentry><refentrytitle>systemd.unit</refentrytitle><manvolnum>5</manvolnum></citerefentry>. If
- used, pass an integer value in the
- range 0-99.</para></listitem>
- </varlistentry>
-
- <varlistentry>
- <term><varname>KillMode=</varname></term>
- <listitem><para>Specifies how
- processes of this service shall be
- killed. One of
- <option>control-group</option>,
- <option>process-group</option>,
- <option>process</option>,
- <option>none</option>.</para>
-
- <para>If set to
- <option>control-group</option> all
- remaining processes in the control
- group of this service will be
- terminated on service stop, after the
- stop command (as configured with
- <varname>ExecStop=</varname>) is
- executed. If set to
- <option>process-group</option> only
- the members of the process group of
- the main service process are
- killed. If set to
- <option>process</option> only the main
- process itself is killed. If set to
- <option>none</option> no process is
- killed. In this case only the stop
- command will be executed on service
- stop, but no process be killed
- otherwise. Processes remaining alive
- after stop are left in their control
- group and the control group continues
- to exist after stop unless it is
- empty. Defaults to
- <option>control-croup</option>.</para>
-
- <para>Processes will first be
- terminated via SIGTERM. If then after
- a delay (configured via the
- <varname>TimeoutSec=</varname> option)
- processes still remain, the
- termination request is repeated with
- the SIGKILL signal. See
- <citerefentry><refentrytitle>kill</refentrytitle><manvolnum>2</manvolnum></citerefentry>
- for more
- information.</para></listitem>
- </varlistentry>
-
<varlistentry>
<term><varname>NonBlocking=</varname></term>
<listitem><para>Set O_NONBLOCK flag
<option>main</option> or
<option>all</option>. If
<option>none</option> no daemon status
- updates are accepted by the service
+ updates are accepted from the service
processes, all status update messages
are ignored. If <option>main</option>
only service updates sent from the
accepted. If <option>all</option> all
services updates from all members of
the service's control group are
- accepted. This option must be set to
+ accepted. This option should be set to
open access to the notification socket
when using
- <varname>Type=notify</varname> (see above).</para></listitem>
+ <varname>Type=notify</varname> or
+ <varname>WatchdogUsec=</varname> (see
+ above). If those options are used but
+ <varname>NotifyAccess=</varname> not
+ configured it will be implicitly set
+ to
+ <option>main</option>.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>Sockets=</varname></term>
+ <listitem><para>Specifies the name of
+ the socket units this service shall
+ inherit the sockets from when the
+ service is started. Normally it
+ should not be necessary to use this
+ setting as all sockets whose unit
+ shares the same name as the service
+ (ignoring the different suffix of course)
+ are passed to the spawned
+ process.</para>
+
+ <para>Note that the same socket may be
+ passed to multiple processes at the
+ same time. Also note that a different
+ service may be activated on incoming
+ traffic than inherits the sockets. Or
+ in other words: The
+ <varname>Service=</varname> setting of
+ <filename>.socket</filename> units
+ doesn't have to match the inverse of the
+ <varname>Sockets=</varname> setting of
+ the <filename>.service</filename> it
+ refers to.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>StartLimitInterval=</varname></term>
+ <term><varname>StartLimitBurst=</varname></term>
+
+ <listitem><para>Configure service
+ start rate limiting. By default
+ services which are started more often
+ than 5 times within 10s are not
+ permitted to start any more times
+ until the 10s interval ends. With
+ these two options this rate limiting
+ may be modified. Use
+ <varname>StartLimitInterval=</varname>
+ to configure the checking interval
+ (defaults to 10s, set to 0 to disable
+ any kind of rate limiting). Use
+ <varname>StartLimitBurst=</varname> to
+ configure how many starts per interval
+ are allowed (defaults to 5). These
+ configuration options are particularly
+ useful in conjunction with
+ <varname>Restart=</varname>, however
+ apply to all kinds of starts
+ (including manual), not just those
+ triggered by the
+ <varname>Restart=</varname> logic.
+ Note that units which are configured
+ for <varname>Restart=</varname> and
+ which reach the start limit are not
+ attempted to be restarted anymore,
+ however they may still be restarted
+ manually at a later point from which
+ point on the restart logic is again
+ activated. Note that
+ <command>systemctl
+ reset-failed</command> will cause the
+ restart rate counter for a service to
+ be flushed, which is useful if the
+ administrator wants to manually start
+ a service and the start limit
+ interferes with
+ that.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>StartLimitAction=</varname></term>
+
+ <listitem><para>Configure the action
+ to take if the rate limit configured
+ with
+ <varname>StartLimitInterval=</varname>
+ and
+ <varname>StartLimitBurst=</varname> is
+ hit. Takes one of
+ <option>none</option>,
+ <option>reboot</option>,
+ <option>reboot-force</option> or
+ <option>reboot-immediate</option>. If
+ <option>none</option> is set,
+ hitting the rate limit will trigger no
+ action besides that the start will not
+ be
+ permitted. <option>reboot</option>
+ causes a reboot following the normal
+ shutdown procedure (i.e. equivalent to
+ <command>systemctl reboot</command>),
+ <option>reboot-force</option> causes
+ an forced reboot which will terminate
+ all processes forcibly but should
+ cause no dirty file systems on reboot
+ (i.e. equivalent to <command>systemctl
+ reboot -f</command>) and
+ <option>reboot-immediate</option>
+ causes immediate execution of the
+ <citerefentry><refentrytitle>reboot</refentrytitle><manvolnum>2</manvolnum></citerefentry>
+ system call, which might result in
+ data loss. Defaults to
+ <option>none</option>.</para></listitem>
+ </varlistentry>
+
+ </variablelist>
+
+ <para>Check
+ <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>
+ and
+ <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>
+ for more settings.</para>
+
+ </refsect1>
+
+ <refsect1>
+ <title>Compatibility Options</title>
+
+ <para>The following options are also available in the
+ <literal>[Service]</literal> section, but exist purely
+ for compatibility reasons and should not be used in
+ newly written service files.</para>
+
+ <variablelist>
+ <varlistentry>
+ <term><varname>SysVStartPriority=</varname></term>
+ <listitem><para>Set the SysV start
+ priority to use to order this service
+ in relation to SysV services lacking
+ LSB headers. This option is only
+ necessary to fix ordering in relation
+ to legacy SysV services, that have no
+ ordering information encoded in the
+ script headers. As such it should only
+ be used as temporary compatibility
+ option, and not be used in new unit
+ files. Almost always it is a better
+ choice to add explicit ordering
+ directives via
+ <varname>After=</varname> or
+ <varname>Before=</varname>,
+ instead. For more details see
+ <citerefentry><refentrytitle>systemd.unit</refentrytitle><manvolnum>5</manvolnum></citerefentry>. If
+ used, pass an integer value in the
+ range 0-99.</para></listitem>
+ </varlistentry>
+
+ <varlistentry>
+ <term><varname>FsckPassNo=</varname></term>
+ <listitem><para>Set the fsck passno
+ priority to use to order this service
+ in relation to other file system
+ checking services. This option is only
+ necessary to fix ordering in relation
+ to fsck jobs automatically created for
+ all <filename>/etc/fstab</filename>
+ entries with a value in the fs_passno
+ column > 0. As such it should only be
+ used as option for fsck
+ services. Almost always it is a better
+ choice to add explicit ordering
+ directives via
+ <varname>After=</varname> or
+ <varname>Before=</varname>,
+ instead. For more details see
+ <citerefentry><refentrytitle>systemd.unit</refentrytitle><manvolnum>5</manvolnum></citerefentry>. If
+ used, pass an integer value in the
+ same range as
+ <filename>/etc/fstab</filename>'s
+ fs_passno column. See
+ <citerefentry><refentrytitle>fstab</refentrytitle><manvolnum>5</manvolnum></citerefentry>
+ for details.</para></listitem>
</varlistentry>
</variablelist>
<citerefentry><refentrytitle>systemd</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
<citerefentry><refentrytitle>systemctl</refentrytitle><manvolnum>8</manvolnum></citerefentry>,
<citerefentry><refentrytitle>systemd.unit</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
- <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>
+ <citerefentry><refentrytitle>systemd.exec</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
+ <citerefentry><refentrytitle>systemd.kill</refentrytitle><manvolnum>5</manvolnum></citerefentry>
</para>
</refsect1>