Replace a cron job with a systemd timer

Cron is fine for simple jobs, but its output disappears unless mail is set up, a job missed while the server was off simply does not run, and testing “what will happen at 02:30” is awkward. A systemd timer is two small unit files: a .service that says what to run and a .timer that says when.

This example replaces the crontab line 30 2 * * * /usr/local/bin/backup.sh.

1. The service: what to run

sudo nano /etc/systemd/system/backup.service

[Unit]
Description=Nightly backup

[Service]
Type=oneshot
ExecStart=/usr/local/bin/backup.sh
User=root

Type=oneshot means the job runs to completion and exits, which is what a scheduled script does. There is no [Install] section: the timer starts this service, not the boot process.

2. The timer: when to run

Give the timer the same base name. If you do not set Unit=, a timer starts the service with the same name.

sudo nano /etc/systemd/system/backup.timer

[Unit]
Description=Run the nightly backup at 02:30

[Timer]
OnCalendar=*-*-* 02:30:00
Persistent=true
RandomizedDelaySec=10min

[Install]
WantedBy=timers.target

• OnCalendar=: a calendar expression. *-*-* 02:30:00 is every day at 02:30.
• Persistent=true: remember the last run on disk; if the server was off at 02:30, run the job once as soon as the timer starts again.
• RandomizedDelaySec=: add a random delay of up to 10 minutes, so many servers do not all hit the backup target at the same second.

3. Check the schedule before enabling

systemd-analyze calendar parses an expression and shows when it will next fire:

systemd-analyze calendar '*-*-* 02:30:00'
systemd-analyze calendar --iterations=5 'Mon..Fri *-*-* 09:00'

Useful expressions:

• daily is *-*-* 00:00:00; weekly is Monday at 00:00.
• Mon..Fri *-*-* 09:00: weekdays at 09:00.
• *:0/15: every 15 minutes (minute 0, 15, 30, 45).
• *-*-01 04:00:00: the first day of each month at 04:00.

For “every N minutes after the last run” instead of wall-clock times, use monotonic timers such as OnBootSec=5min together with OnUnitActiveSec=1h.

4. Enable the timer and test the job

sudo systemctl daemon-reload
sudo systemctl enable --now backup.timer
systemctl list-timers backup.timer      # NEXT and LAST run times

sudo systemctl start backup.service      # run the job right now, as a test
journalctl -u backup.service -n 30       # its output and exit status

Note that you enable the timer, not the service. Starting the service by hand is the easiest way to test the exact environment the timer will use.

5. Remove the old cron entry

sudo crontab -e      # delete the backup line

Cron to timer cheat sheet

• @reboot becomes OnBootSec=1min.
• 0 * * * * becomes OnCalendar=hourly.
• */10 * * * * becomes OnCalendar=*:0/10.
• 0 3 * * 0 becomes OnCalendar=Sun *-*-* 03:00:00.
• Cron’s MAILTO output becomes journalctl -u name.service.

If the job fails, systemctl status backup.service shows failed and it appears in systemctl --failed, which makes it easy to pick up with monitoring.

Official documentation

• systemd.timer(5)
• systemd.time(7): calendar expressions
• systemd-analyze(1)