Controlling services with systemctl

The idea

systemd is the first process (PID 1). It starts and supervises everything else, described as units. A service unit (name.service) is a daemon; there are also timers, mounts, sockets and targets. You control all of them with systemctl.

Two separate questions are always being answered about a service:

Question	Command to change it	Command to ask
Is it running now?	start, stop, restart, reload	systemctl is-active NAME
Will it start at boot?	enable, disable	systemctl is-enabled NAME

They are independent. systemctl enable --now NAME does both in one go, and it is the form to reach for first when a task says “make sure it is running and starts at boot”. restart stops and starts (connections drop), reload asks the daemon to re-read its configuration without stopping (not every service supports it), and try-restart restarts only if it is already running.

mask goes further than disable: it links the unit to /dev/null, so nothing, not even another service that depends on it, can start it. Use it to forbid a service outright; unmask reverses it.

Unit files live in three places, from lowest to highest priority: /usr/lib/systemd/system/ (packages), /run/systemd/system/ (temporary), /etc/systemd/system/ (yours). Never edit the packaged copy; create an override instead. After creating or changing any unit file, run systemctl daemon-reload so systemd re-reads them.

Try it

systemctl status sshd --no-pager

● sshd.service - OpenSSH server daemon
     Loaded: loaded (/usr/lib/systemd/system/sshd.service; enabled; preset: enabled)
     Active: active (running) since Tue 2026-11-03 08:01:12 +03; 3h 40min ago
       Docs: man:sshd(8)
   Main PID: 842 (sshd)
      Tasks: 1 (limit: 22900)
     Memory: 4.2M
     CGroup: /system.slice/sshd.service
             └─842 "sshd: /usr/sbin/sshd -D [listener]"

Nov 03 08:01:12 rhel10 systemd[1]: Started sshd.service - OpenSSH server daemon.

Read it top to bottom: the Loaded line gives the file and whether it is enabled for boot; Active gives the live state and how long; the last lines are the most recent journal entries for the unit. The bullet is green for running, red for failed.

Run now, start at boot

sudo dnf install -y httpd
sudo systemctl enable --now httpd
systemctl is-active httpd
systemctl is-enabled httpd

Created symlink '/etc/systemd/system/multi-user.target.wants/httpd.service' → '/usr/lib/systemd/system/httpd.service'.
active
enabled

enable simply creates the symlink shown: that is all “starts at boot” means. A service that is enabled but not running will come up at the next boot; one that is running but disabled will not.

Check what is listening

sudo ss -tlnp | head -4

State  Recv-Q Send-Q Local Address:Port Peer Address:Port Process
LISTEN 0      128          0.0.0.0:22        0.0.0.0:*     users:(("sshd",pid=842,fd=3))
LISTEN 0      511                *:80              *:*     users:(("httpd",pid=3120,fd=4))

-t TCP, -l listening, -n numbers instead of names, -p show the process. Add -u for UDP. This is the quickest way to confirm “the network service is up and bound to the right port”.

Your own service

[Unit]
Description=Heartbeat logger

[Service]
ExecStart=/usr/local/bin/heartbeat.sh
Restart=on-failure

[Install]
WantedBy=multi-user.target

ExecStart must be a full path to an executable file. Restart=on-failure brings the process back if it crashes. WantedBy=multi-user.target is what enable links the unit into, so that it starts in the normal multi-user boot.

sudo systemctl daemon-reload
sudo systemctl enable --now heartbeat
journalctl -t heartbeat -n 2 --no-pager

Nov 03 11:50:10 rhel10 heartbeat[4410]: alive at 11:50:10
Nov 03 11:50:20 rhel10 heartbeat[4410]: alive at 11:50:20

Customise a packaged service without editing it

sudo systemctl edit httpd

[Service]
Restart=always
RestartSec=5

systemctl edit writes /etc/systemd/system/httpd.service.d/override.conf and reloads for you. systemctl cat httpd shows the original and the override together. The packaged file stays untouched, so updates do not overwrite your change.

When a service fails

systemctl --failed --no-pager
systemctl status heartbeat --no-pager | head -6
journalctl -u heartbeat -n 3 --no-pager

  UNIT              LOAD   ACTIVE SUB    DESCRIPTION
● heartbeat.service loaded failed failed Heartbeat logger
● heartbeat.service - Heartbeat logger
     Loaded: loaded (/etc/systemd/system/heartbeat.service; enabled; preset: disabled)
     Active: failed (Result: exit-code) since Tue 2026-11-03 11:52:03 +03; 4s ago
    Process: 4502 ExecStart=/usr/local/bin/heartbeat-missing.sh (code=exited, status=203/EXEC)
Nov 03 11:52:03 rhel10 systemd[4502]: heartbeat.service: Failed to locate executable /usr/local/bin/heartbeat-missing.sh: No such file or directory

status=203/EXEC means systemd could not execute the file: wrong path, no execute bit, or a bad shebang. Other common codes: status=1/FAILURE (the program itself exited with an error, so read the program’s own message in the journal) and status=217/USER (the User= does not exist). After fixing, daemon-reload, then systemctl reset-failed and restart.

Going further

systemctl list-units --type=service --state=running lists what is running; systemctl list-unit-files --state=enabled lists what starts at boot. systemctl list-dependencies httpd shows what a service needs, and systemctl show -p After httpd shows ordering. Requires= and Wants= declare dependencies; After= only orders. The SELinux and firewall layers can also stop a service from serving traffic even when systemctl says active, so a “running” service is the start of troubleshooting, not the end.