« Back to Systemd

Systemd units and their purpose

Which systemd unit types are available and what is their goal? In this article we cover them and show some useful commands related to these units.

Showing all available unit types

Let’s start by displaying some of the units that are available on your system.

systemctl list-unit-files

This will reveal types like:

  • Automount
  • Device
  • Mount
  • Path
  • Scope
  • Service
  • Slice
  • Snapshot
  • Socket
  • Target
  • Timer

Retrieve basic information from a unit

Before we dive into all unit types, it is good to know a few commands upfront. These help to inspect or learn more about a unit.

  • Receive help: systemctl help ssh.service
  • Show the unit file: systemctl cat ssh.service
  • Show all unit settings: systemctl show ssh.service

Unit types

.service

This is a very common unit and refers to a service that runs in the background. Typically these are daemons that perform the whole type the system is running. Another option is a script that is executed by a trigger, like a repeating task.

.timer

Another common unit is the timer. It’s a scheduler, similar to cronjobs. Usually a timer and service file are paired together. The timer does the scheduling and then triggers the service.

systemctl list-timers

.target

A state within the service manager. It groups services and performs synchronization during the boot process or when the state changes. A good example of this is services that require the network stack to be up and running, like nginx. The network target can only be started when the basic functionality of the system is active, like some devices. In other words, it is a chain of events.

To see the time-critical tasks during the boot, we can use the systemd-analyze tool.

# systemd-analyze critical-chain
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the "+" character.

graphical.target @19.498s
└─multi-user.target @19.498s
  └─snapd.seeded.service @17.537s +1.874s
    └─basic.target @17.467s
      └─sockets.target @17.466s
        └─snapd.socket @17.444s +15ms
          └─sysinit.target @17.419s
            └─cloud-init.service @16.952s +465ms
              └─systemd-networkd-wait-online.service @3.048s +13.900s
                └─systemd-networkd.service @3.010s +36ms
                  └─network-pre.target @3.009s
                    └─cloud-init-local.service @2.383s +625ms
                      └─systemd-remount-fs.service @445ms +62ms
                        └─systemd-journald.socket @380ms
                          └─-.mount @345ms
                            └─-.slice @345ms

If we want to see how specific targets are linked, we can also use systemctl list-dependencies that shows the relationship between different units.

# systemctl list-dependencies local-fs.target
local-fs.target
● ├─-.mount
● ├─boot.mount
○ ├─systemd-fsck-root.service
● └─systemd-remount-fs.service

To have the local-fs.target up and running, a set of mount and file system checks have to be performed. We can see that the ‘fsck’ service is not running, as that only gets triggered when it is actually needed. The other three units (two mount units, one service) are active.

.socket

A socket description for socket based activation. This is typically something like a network socket, file system FIFO file, or other forms of inter-process communication (IPC).

systemctl list-sockets

.device

Devices that demand systemd management, are stored in this unit type, like the initrd (ramdisk).

.mount

Defines a mount point on the system

.automount

Defines a mount point that will be mounted automatically.

systemctl list-automounts

.path

Defines a file or directory to monitor and take trigger a related service. This is for example great to monitor for changes and then initiate a related action.

systemctl list-paths

.swap

Define the swap space, which is used when there is not enough memory.

.snapshot

Dynamic snapshot of the systemd runtime state. This is not a file type like the others, but an autogenerated definition.

.slice

Defines cgroups, which are the Linux Control Groups. They restrict access to resources, such as memory and CPU.

.scope

These are created automatically and are used to manage sets of systemd processes, such as a user sesion.

Learn more about systemctl

This article uses the systemctl command to achieve its tasks. For this popular tool there is a cheat sheet available!

» Mastering the tool: systemctl

systemctl cheat sheet

Feedback

Small picture of Michael Boelen

This article has been written by our Linux security expert Michael Boelen. With focus on creating high-quality articles and relevant examples, he wants to improve the field of Linux security. No more web full of copy-pasted blog posts.

Discovered outdated information or have a question? Share your thoughts. Thanks for your contribution.

Mastodon icon