https://wiki.archlinux.org/index.php/systemd

systemd  is a suite of basic building blocks for a Linux system. It provides a system and service manager that runs as PID 1 and starts the rest of the system. systemd provides aggressive parallelization capabilities, uses socket and  D-Bus  activation for starting services, offers on-demand starting of daemons, keeps track of processes using Linux  control groups , maintains mount and automount points, and implements an elaborate transactional dependency-based service control logic.  systemd  supports SysV and LSB init scripts and works as a replacement for sysvinit. Other parts include a logging daemon, utilities to control basic system configuration like the hostname, date, locale, maintain a list of logged-in users and running containers and virtual machines, system accounts, runtime directories and settings, and daemons to manage simple network configuration, network time synchronization, log forwarding, and name resolution.

Basic systemctl usage

The main command used to introspect and control systemd is systemctl. Some of its uses are examining the system state and managing the system and services. See systemctl(1) for more details.

Analyzing the system state

Show system status using:

$ systemctl status

List running units:

$ systemctl

or:

$ systemctl list-units

List failed units:

$ systemctl --failed

The available unit files can be seen in /usr/lib/systemd/system/ and /etc/systemd/system/ (the latter takes precedence). List installed unit files with:

$ systemctl list-unit-files

Using units

Units can be, for example, services (.service), mount points (.mount), devices (.device) or sockets (.socket).

When using systemctl, you generally have to specify the complete name of the unit file, including its suffix, for example sshd.socket. There are however a few short forms when specifying the unit in the following systemctl commands:

• If you do not specify the suffix, systemctl will assume .service. For example, netctl and netctl.service are equivalent.
• Mount points will automatically be translated into the appropriate .mount unit. For example, specifying /home is equivalent to home.mount.
• Similar to mount points, devices are automatically translated into the appropriate .device unit, therefore specifying /dev/sda2 is equivalent to dev-sda2.device.

See systemd.unit(5) for details.

Start a unit immediately:

# systemctl start unit

Stop a unit immediately:

# systemctl stop unit

Restart a unit:

# systemctl restart unit

Ask a unit to reload its configuration:

# systemctl reload unit

Show the status of a unit, including whether it is running or not:

$ systemctl status unit

Check whether a unit is already enabled or not:

$ systemctl is-enabled unit

Enable a unit to be started on bootup:

# systemctl enable unit

Enable a unit to be started on bootup and Start immediately:

# systemctl enable --now unit

Disable a unit to not start during bootup:

# systemctl disable unit

Mask a unit to make it impossible to start it:

# systemctl mask unit

Unmask a unit:

# systemctl unmask unit

Show the manual page associated with a unit (this has to be supported by the unit file):

$ systemctl help unit

Reload systemd, scanning for new or changed units:

# systemctl daemon-reload

Power management

polkit is necessary for power management as an unprivileged user. If you are in a local systemd-logind user session and no other session is active, the following commands will work without root privileges. If not (for example, because another user is logged into a tty), systemd will automatically ask you for the root password.

Shut down and reboot the system:

$ systemctl reboot

Shut down and power-off the system:

$ systemctl poweroff

Suspend the system:

$ systemctl suspend

Put the system into hibernation:

$ systemctl hibernate

Put the system into hybrid-sleep state (or suspend-to-both):

$ systemctl hybrid-sleep

Writing unit files

The syntax of systemd's unit files is inspired by XDG Desktop Entry Specification .desktop files, which are in turn inspired by Microsoft Windows .ini files. Unit files are loaded from two locations. From lowest to highest precedence they are:

• /usr/lib/systemd/system/: units provided by installed packages
• /etc/systemd/system/: units installed by the system administrator

Look at the units installed by your packages for examples, as well as the annotated example section of systemd.service(5).

Handling dependencies

With systemd, dependencies can be resolved by designing the unit files correctly. The most typical case is that the unit A requires the unit B to be running before A is started. In that case add Requires=B and After=B to the [Unit] section of A. If the dependency is optional, add Wants=B and After=B instead. Note that Wants= and Requires= do not imply After=, meaning that if After= is not specified, the two units will be started in parallel.

注意，After和Before字段只涉及启动顺序，不涉及依赖关系。

设置依赖关系，需要使用Wants字段和Requires字段。

Wants字段：表示sshd.service与sshd-keygen.service之间存在"弱依赖"关系，即如果"sshd-keygen.service"启动失败或停止运行，不影响sshd.service继续执行。

Requires字段则表示"强依赖"关系，即如果该服务启动失败或异常退出，那么sshd.service也必须退出。

注意，Wants字段与Requires字段只涉及依赖关系，与启动顺序无关，默认情况下是同时启动的。

Dependencies are typically placed on services and not on #Targets. For example, network.target is pulled in by whatever service configures your network interfaces, therefore ordering your custom unit after it is sufficient since network.target is started anyway.

Service types

There are several different start-up types to consider when writing a custom service file. This is set with the Type= parameter in the [Service] section:

• Type=simple (default): systemd considers the service to be started up immediately. The process must not fork. Do not use this type if other services need to be ordered on this service, unless it is socket activated.
• Type=forking: systemd considers the service started up once the process forks and the parent has exited. For classic daemons use this type unless you know that it is not necessary. You should specify PIDFile= as well so systemd can keep track of the main process.
• Type=oneshot: this is useful for scripts that do a single job and then exit. You may want to set RemainAfterExit=yes as well so that systemd still considers the service as active after the process has exited.
• Type=notify: identical to Type=simple, but with the stipulation that the daemon will send a signal to systemd when it is ready. The reference implementation for this notification is provided by libsystemd-daemon.so.
• Type=dbus: the service is considered ready when the specified BusName appears on DBus's system bus.
• Type=idle: systemd will delay execution of the service binary until all jobs are dispatched. Other than that behavior is very similar to Type=simple.

See the systemd.service(5) man page for a more detailed explanation of the Type values.

Editing provided units

To avoid conflicts with pacman, unit files provided by packages should not be directly edited. There are two safe ways to modify a unit without touching the original file: create a new unit file which overrides the original unit or create drop-in snippets which are applied on top of the original unit. For both methods, you must reload the unit afterwards to apply your changes. This can be done either by editing the unit with systemctl edit (which reloads the unit automatically) or by reloading all units with:

# systemctl daemon-reload

Replacement unit files

To replace the unit file /usr/lib/systemd/system/unit, create the file /etc/systemd/system/unit and reenable the unit to update the symlinks:

# systemctl reenable unit

Alternatively, run:

# systemctl edit --full unit

This opens /etc/systemd/system/unit in your editor (copying the installed version if it does not exist yet) and automatically reloads it when you finish editing.

Drop-in files

To create drop-in files for the unit file /usr/lib/systemd/system/unit, create the directory /etc/systemd/system/unit.d/ and place .conf files there to override or add new options. systemd will parse and apply these files on top of the original unit.

The easiest way to do this is to run:

# systemctl edit unit

This opens the file /etc/systemd/system/unit.d/override.conf in your text editor (creating it if necessary) and automatically reloads the unit when you are done editing.

Revert to vendor version

To revert any changes to a unit made using systemctl edit do:

将指定的单元恢复成初始版本。 这将会删除对指定单元的所有修改。 例如，对于 "foo.service" 单元来说， 将会删除所有 foo.service.d/ 目录。 如果指定的单元在 /usr/lib/ 目录中 还存在单元文件的初始版本，那么还会进一步删除 /etc/ 与 /run/ 目录中 所有用来覆盖初始单元文件的软连接与自定义单元文件。 如果指定的单元已经被屏蔽，那么将会被解除屏蔽。

从效果上看，该命令相当于撤销 edit, set-property, mask 命令所做的操作， 并且将指定单元的配置恢复成软件包提供的初始值。

centos上貌似没有这个指令

# systemctl revert unit

Examples

For example, if you simply want to add an additional dependency to a unit, you may create the following file:

/etc/systemd/system/unit.d/customdependency.conf

[Unit]
Requires=new dependency
After=new dependency

As another example, in order to replace the ExecStart directive for a unit that is not of type oneshot, create the following file:

/etc/systemd/system/unit.d/customexec.conf

[Service]
ExecStart=
ExecStart=new command

Note how ExecStart must be cleared before being re-assigned [1]. The same holds for every item that can be specified multiple times, e.g. OnCalendar for timers.

One more example to automatically restart a service:

/etc/systemd/system/unit.d/restart.conf

[Service]
Restart=always
RestartSec=30

Targets

systemd uses targets which serve a similar purpose as runlevels but act a little different. Each target is named instead of numbered and is intended to serve a specific purpose with the possibility of having multiple ones active at the same time. Some targets are implemented by inheriting all of the services of another target and adding additional services to it. There are systemd targets that mimic the common SystemVinit runlevels so you can still switch targets using the familiar telinit RUNLEVEL command.

Get current targets

The following should be used under systemd instead of running runlevel:

$ systemctl list-units --type=target

Create custom target

The runlevels that held a defined meaning under sysvinit (i.e., 0, 1, 3, 5, and 6); have a 1:1 mapping with a specific systemd target. Unfortunately, there is no good way to do the same for the user-defined runlevels like 2 and 4. If you make use of those it is suggested that you make a new named systemd target as /etc/systemd/system/your target that takes one of the existing runlevels as a base (you can look at /usr/lib/systemd/system/graphical.target as an example), make a directory /etc/systemd/system/your target.wants, and then symlink the additional services from /usr/lib/systemd/system/ that you wish to enable.

Mapping between SysV runlevels and systemd targets

Change current target

In systemd targets are exposed via target units. You can change them like this:

# systemctl isolate graphical.target

This will only change the current target, and has no effect on the next boot. This is equivalent to commands such as telinit 3 or telinit 5 in Sysvinit.

Change default target to boot into

The standard target is default.target, which is aliased by default to graphical.target (which roughly corresponds to the old runlevel 5). To change the default target at boot-time, append one of the following kernel parameters to your bootloader:

• systemd.unit=multi-user.target (which roughly corresponds to the old runlevel 3),
• systemd.unit=rescue.target (which roughly corresponds to the old runlevel 1).

Alternatively, you may leave the bootloader alone and change default.target. This can be done using systemctl:

# systemctl set-default multi-user.target

To be able to override the previously set default.target, use the force option:

# systemctl set-default -f multi-user.target

The effect of this command is output by systemctl; a symlink to the new default target is made at /etc/systemd/system/default.target.

Temporary files

"systemd-tmpfiles creates, deletes and cleans up volatile and temporary files and directories." It reads configuration files in /etc/tmpfiles.d/ and /usr/lib/tmpfiles.d/ to discover which actions to perform. Configuration files in the former directory take precedence over those in the latter directory.

Configuration files are usually provided together with service files, and they are named in the style of /usr/lib/tmpfiles.d/program.conf. For example, the Samba daemon expects the directory /run/samba to exist and to have the correct permissions. Therefore, the samba package ships with this configuration:

/usr/lib/tmpfiles.d/samba.conf

D /run/samba 0755 root root

Configuration files may also be used to write values into certain files on boot. For example, if you used /etc/rc.local to disable wakeup from USB devices with echo USBE > /proc/acpi/wakeup, you may use the following tmpfile instead:

/etc/tmpfiles.d/disable-usb-wake.conf

w /proc/acpi/wakeup - - - - USBE

See the systemd-tmpfiles(8) and tmpfiles.d(5) man pages for details.

Timers

A timer is a unit configuration file whose name ends with .timer and encodes information about a timer controlled and supervised by systemd, for timer-based activation. See systemd/Timers.

Mounting

Since systemd is a replacement for System V init, it is in charge of the mounts specified in /etc/fstab. In fact, it goes beyond the usual fstab capabilities, implementing special mount options prefixed with x-systemd.. See Fstab#Automount with systemd for an example of automounting (mounting on-demand) using these extensions. See [2] for the complete documentation of these extensions.

Journal

systemd has its own logging system called the journal; therefore, running a syslog daemon is no longer required. To read the log, use:

# journalctl

In Arch Linux, the directory /var/log/journal/ is a part of the systemd package, and the journal (when Storage= is set to auto in /etc/systemd/journald.conf) will write to /var/log/journal/. If you or some program delete that directory, systemd will not recreate it automatically and instead will write its logs to /run/systemd/journal in a nonpersistent way. However, the folder will be recreated when you set Storage=persistent and run systemctl restart systemd-journald (or reboot).

Systemd journal classifies messages by Priority level and Facility. Logging classification corresponds to classic Syslog protocol (RFC 5424).

Priority level

A syslog severity code (in systemd called priority) is used to mark the importance of a message RFC 5424 Section 6.2.1.

If you cannot find a message on the expected priority level, also search a couple of levels above and below: these rules are recommendations, and the developer of the affected application may have a different perception of the issue's importance from yours.

Facility

A syslog facility code is used to specify the type of program that is logging the message RFC 5424 Section 6.2.1.

So, useful facilities to watch: 0,1,3,4,9,10,15.

Filtering output

journalctl allows you to filter the output by specific fields. Be aware that if there are many messages to display or filtering of large time span has to be done, the output of this command can be delayed for quite some time.

$ strings /mnt/arch/var/log/journal/af4967d77fba44c6b093d0e9862f6ddd/system.journal | grep -i message

Examples:

• Show all messages from this boot:

  # journalctl -b

  However, often one is interested in messages not from the current, but from the previous boot (e.g. if an unrecoverable system crash happened). This is possible through optional offset parameter of the -b flag: journalctl -b -0 shows messages from the current boot, journalctl -b -1 from the previous boot, journalctl -b -2 from the second previous and so on. See journalctl(1) for full description, the semantics is much more powerful.
• Show all messages from date (and optional time):

  # journalctl --since="2012-10-30 18:17:16"

• Show all messages since 20 minutes ago:

  # journalctl --since "20 min ago"

• Follow new messages:

  # journalctl -f

• Show all messages by a specific executable:

  # journalctl /usr/lib/systemd/systemd

• Show all messages by a specific process:

  # journalctl _PID=1

• Show all messages by a specific unit:

  # journalctl -u netcfg

• Show kernel ring buffer:

  # journalctl -k

• Show only error, critical, and alert priority messages

  # journalctl -p err..alert

  Numbers also can be used, journalctl -p 3..1. If single number/keyword used, journalctl -p 3 - all higher priority levels also included.
• Show auth.log equivalent by filtering on syslog facility:

  # journalctl SYSLOG_FACILITY=10

See journalctl(1), systemd.journal-fields(7), or Lennart's blog post for details.

$ SYSTEMD_LESS=FRXMK journalctl

Journal size limit

If the journal is persistent (non-volatile), its size limit is set to a default value of 10% of the size of the underlying file system but capped to 4 GiB. For example, with /var/log/journal/located on a 20 GiB partition, journal data may take up to 2 GiB. On a 50 GiB partition, it would max at 4 GiB.

The maximum size of the persistent journal can be controlled by uncommenting and changing the following:

/etc/systemd/journald.conf

SystemMaxUse=50M

It is also possible to use the drop-in snippets configuration override mechanism rather than editing the global configuration file. In this case do not forget to place the overrides under the [Journal] header:

/etc/systemd/journald.conf.d/00-journal-size.conf

[Journal]
SystemMaxUse=50M

Restart the systemd-journald.service after changing this setting to immediately apply the new limit.

See journald.conf(5) for more info.

Clean journal files manually

Journal files can be globally removed from /var/log/journal/ using e.g. rm, or can be trimmed according to various criteria using journalctl. Examples:

• Remove archived journal files until the disk space they use falls below 100M:

  # journalctl --vacuum-size=100M

• Make all journal files contain no data older than 2 weeks.

  # journalctl --vacuum-time=2weeks

See journalctl(1) for more info.

Journald in conjunction with syslog

Compatibility with a classic, non-journald aware syslog implementation can be provided by letting systemd forward all messages via the socket /run/systemd/journal/syslog. To make the syslog daemon work with the journal, it has to bind to this socket instead of /dev/log (official announcement).

The default journald.conf for forwarding to the socket is ForwardToSyslog=no to avoid system overhead, because rsyslog or syslog-ng pull the messages from the journal by itself.

See Syslog-ng#Overview and Syslog-ng#syslog-ng and systemd journal, or rsyslog respectively, for details on configuration.

Forward journald to /dev/tty12

Create a drop-in directory /etc/systemd/journald.conf.d and create a fw-tty12.conf file in it:

/etc/systemd/journald.conf.d/fw-tty12.conf

[Journal]
ForwardToConsole=yes
TTYPath=/dev/tty12
MaxLevelConsole=info

Then restart systemd-journald.

Specify a different journal to view

There may be a need to check the logs of another system that is dead in the water, like booting from a live system to recover a production system. In such case, one can mount the disk in e.g. /mnt, and specify the journal path via -D/--directory, like so:

$ journalctl -D /mnt/var/log/journal -xe

Tips and tricks

Enable installed units by default

Arch Linux ships with /usr/lib/systemd/system-preset/99-default.preset containing disable *. This causes systemctl preset to disable all units by default, such that when a new package is installed, the user must manually enable the unit.

If this behavior is not desired, simply create a symlink from /etc/systemd/system-preset/99-default.preset to /dev/null in order to override the configuration file. This will cause systemctl preset to enable all units that get installed—regardless of unit type—unless specified in another file in one systemctl preset's configuration directories. User units are not affected. See systemd.preset(5) for more information.

Sandboxing application environments

A unit file can be created as a sandbox to isolate applications and their processes within a hardened virtual environment. systemd leverages namespaces, white-/blacklisting of Capabilities, and control groups to container processes through an extensive execution environment configuration.

The enhancement of an existing systemd unit file with application sandboxing typically requires trial-and-error tests accompanied by the generous use of strace, stderr and journalctl error logging and output facilities. You may want to first search upstream documentation for already done tests to base trials on.

Some examples on how sandboxing with systemd can be deployed:

• CapabilityBoundingSet defines a whitelisted set of allowed capabilities, but may also be used to blacklist a specific capability for a unit.
  • The CAP_SYS_ADM capability, for example, which should be one of the goals of a secure sandbox: CapabilityBoundingSet=~ CAP_SYS_ADM

Troubleshooting

Investigating systemd errors

As an example, we will investigate an error with systemd-modules-load service:

1. Lets find the systemd services which fail to start at boot time:

$ systemctl --state=failed

systemd-modules-load.service   loaded failed failed  Load Kernel Modules

Another way is to live log systemd messages:

$ journalctl -fp err

2. Ok, we found a problem with systemd-modules-load service. We want to know more:

$ systemctl status systemd-modules-load

systemd-modules-load.service - Load Kernel Modules
   Loaded: loaded (/usr/lib/systemd/system/systemd-modules-load.service; static)
   Active: failed (Result: exit-code) since So 2013-08-25 11:48:13 CEST; 32s ago
     Docs: man:systemd-modules-load.service(8).
           man:modules-load.d(5)
  Process: 15630 ExecStart=/usr/lib/systemd/systemd-modules-load (code=exited, status=1/FAILURE)

If the Process ID is not listed, just restart the failed service with systemctl restart systemd-modules-load

3. Now we have the process id (PID) to investigate this error in depth. Enter the following command with the current Process ID (here: 15630):

$ journalctl _PID=15630

-- Logs begin at Sa 2013-05-25 10:31:12 CEST, end at So 2013-08-25 11:51:17 CEST. --
Aug 25 11:48:13 mypc systemd-modules-load[15630]: Failed to find module 'blacklist usblp'
Aug 25 11:48:13 mypc systemd-modules-load[15630]: Failed to find module 'install usblp /bin/false'

4. We see that some of the kernel module configs have wrong settings. Therefore we have a look at these settings in /etc/modules-load.d/:

$ ls -Al /etc/modules-load.d/

...
-rw-r--r--   1 root root    79  1. Dez 2012  blacklist.conf
-rw-r--r--   1 root root     1  2. Mär 14:30 encrypt.conf
-rw-r--r--   1 root root     3  5. Dez 2012  printing.conf
-rw-r--r--   1 root root     6 14. Jul 11:01 realtek.conf
-rw-r--r--   1 root root    65  2. Jun 23:01 virtualbox.conf
...

5. The Failed to find module 'blacklist usblp' error message might be related to a wrong setting inside of blacklist.conf. Lets deactivate it with inserting a trailing # before each option we found via step 3:

/etc/modules-load.d/blacklist.conf

# blacklist usblp
# install usblp /bin/false

6. Now, try to start systemd-modules-load:

$ systemctl start systemd-modules-load

If it was successful, this should not prompt anything. If you see any error, go back to step 3 and use the new PID for solving the errors left.

If everything is ok, you can verify that the service was started successfully with:

$ systemctl status systemd-modules-load

systemd-modules-load.service - Load Kernel Modules
   Loaded: loaded (/usr/lib/systemd/system/systemd-modules-load.service; static)
   Active: active (exited) since So 2013-08-25 12:22:31 CEST; 34s ago
     Docs: man:systemd-modules-load.service(8)
           man:modules-load.d(5)
 Process: 19005 ExecStart=/usr/lib/systemd/systemd-modules-load (code=exited, status=0/SUCCESS)
Aug 25 12:22:31 mypc systemd[1]: Started Load Kernel Modules.

Diagnosing boot problems

systemd has several options for diagnosing problems with the boot process. See boot debugging for more general instructions and options to capture boot messages before systemdtakes over the boot process. Also see the systemd debugging documentation.

Diagnosing problems with a specific service

If some systemd service misbehaves and you want to get more information about what is going on, set the SYSTEMD_LOG_LEVEL environment variable to debug. For example, to run the systemd-networkd daemon in debug mode:

# systemctl stop systemd-networkd
# SYSTEMD_LOG_LEVEL=debug /lib/systemd/systemd-networkd

Or, equivalently, modify the service file temporarily for gathering enough output. For example:

/usr/lib/systemd/system/systemd-networkd.service

[Service]
...
Environment=SYSTEMD_LOG_LEVEL=debug
....

If debug information is required long-term, add the variable the regular way.

Shutdown/reboot takes terribly long

If the shutdown process takes a very long time (or seems to freeze) most likely a service not exiting is to blame. systemd waits some time for each service to exit before trying to kill it. To find out if you are affected, see this article.

Short lived processes do not seem to log any output

If journalctl -u foounit does not show any output for a short lived service, look at the PID instead. For example, if systemd-modules-load.service fails, and systemctl status systemd-modules-load shows that it ran as PID 123, then you might be able to see output in the journal for that PID, i.e. journalctl -b _PID=123. Metadata fields for the journal such as _SYSTEMD_UNIT and _COMM are collected asynchronously and rely on the /proc directory for the process existing. Fixing this requires fixing the kernel to provide this data via a socket connection, similar to SCM_CREDENTIALS.

Boot time increasing over time

After using systemd-analyze a number of users have noticed that their boot time has increased significantly in comparison with what it used to be. After using systemd-analyze blameNetworkManager is being reported as taking an unusually large amount of time to start.

The problem for some users has been due to /var/log/journal becoming too large. This may have other impacts on performance, such as for systemctl status or journalctl. As such the solution is to remove every file within the folder (ideally making a backup of it somewhere, at least temporarily) and then setting a journal file size limit as described in #Journal size limit.

systemd-tmpfiles-setup.service fails to start at boot

Starting with systemd 219, /usr/lib/tmpfiles.d/systemd.conf specifies ACL attributes for directories under /var/log/journal and, therefore, requires ACL support to be enabled for the filesystem the journal resides on.

See Access Control Lists#Enabling ACL for instructions on how to enable ACL on the filesystem that houses /var/log/journal.

systemd version printed on boot is not the same as installed package version

You need to regenerate your initramfs and the versions should match.

Disable emergency mode on remote machine

You may want to disable emergency mode on a remote machine, for example, a virtual machine hosted at Azure or Google Cloud. It is because if emergency mode is triggered, the machine will be blocked from connecting to network.

# systemctl mask emergency.service
# systemctl mask emergency.target