Basic systemd concepts

This document applies to the following products and product versions:

systemd is the modern replacement for the legacy Linux initialization systems, System V (SysV) and Linux Standard Base init (LSB init). systemd owns PID 1, and is started directly by the kernel. All other processes are either started directly by systemd, or by one of its child processes. systemd is designed to manage the complex demands of modern Linux systems, by dynamically managing all processes from system boot to shutdown.

systemd uses unit files to control related processes. A unit configuration file contains all the elements necessary to operate a service: a listening socket, device, mount point, automount point, swap file or partition, startup target, watched file system path, timers, temporary system state snapshot, resource management slice, or a group of externally created processes. This simplifies system administration, as systemd tracks and manages all the elements necessary for a particular service, and the administrator has only a single service unit file to manage for any service.

Unit files are customizable, and simpler and easier to understand than SysV and LSB init scripts.

systemd provides a number of benefits. It is a single daemon that performs complete process management, from startup to shutdown. It speeds up boot times by parallelizing service starts. It conserves system resources by starting services on demand, and puts them to sleep when there is no demand. systemd reduces complexity and increases efficiency by opening listening sockets, which are then available when services start, and the administrator does not have to worry about starting them in the correct order.

In contrast, SysV and LSB init only start services at boot, then exit. The SysV and LSB init administrator must configure services carefully to start in the correct order, for example networking must start before network services, and the administrator must remember to configure all necessary services to start at boot.

systemd is written in C, in contrast to the large assortment of shell scripts that comprise SysV and LSB init. The compiled single systemd binary is faster and more efficient than masses of scripts.

systemd provides unified logging for all processes with the systemd journal, which is viewed with the journalctl command.

systemd organizes processes with Kernel Control Groups (cgroups). cgroups are inbuilt in the Linux kernel. This provides an extremely efficient way to organize all related processes for a service, and to quickly find which process belongs to which service. cgroups also provide resource management, such as CPU, memory, I/O, and device groups.

The systemctl command is the systemd and service manager.

journalctl and systemctl replace a host of legacy commands and scripts.

systemd is backwards-compatible with SysV and LSB init.

Because systemd is a complete process manager, from startup to shutdown, it controls all operations on a Linux system. Consult the following resources to learn more.