Schedule system jobs and clean temporary files

$ run as a normal user# run as root (or with sudo)

Skills in this chapter

Schedule system jobs with cron, systemd timers and tmpfiles

Use the system cron files

System cron jobs live in files, not in crontab -e. Their lines have one extra field: the user the job runs as.

/etc/cron.d/backup

# Run the backup as root every night at 01:30
SHELL=/bin/bash
PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin
30 1 * * *  root  /usr/local/bin/backup.sh >> /var/log/backup.log 2>&1
*/5 * * * * apache /usr/local/bin/clean-sessions.sh

Place	Use
`/etc/crontab`	The main system table (do not change it: package updates may)
`/etc/cron.d/NAME`	One file per job or per application — use this
`/etc/cron.hourly/`, `cron.daily/`, `cron.weekly/`, `cron.monthly/`	Scripts run once per hour / day / week / month
`/var/spool/cron/USER`	The crontabs of the users (made by `crontab -e`)

#ls /etc/cron.d/ /etc/cron.daily/
#install -m 755 cleanup.sh /etc/cron.daily/cleanup
#run-parts --test /etc/cron.daily

Watch out

Scripts in cron.daily/ must be executable and their names must not have a dot — cleanup, not cleanup.sh. Otherwise run-parts skips them.

4.2

Catch up missed jobs with anacron

anacron runs the daily, weekly and monthly scripts even when the machine was off at the planned time. It remembers the last run of each.

/etc/anacrontab

# period  delay  job-id        command
1         5      cron.daily    nice run-parts /etc/cron.daily
7         25     cron.weekly   nice run-parts /etc/cron.weekly
@monthly  45     cron.monthly  nice run-parts /etc/cron.monthly

#cat /etc/anacrontab
#ls -l /var/spool/anacron/
#cat /var/spool/anacron/cron.daily
20261027

Field	Meaning
period	Run every N days (`@monthly` = once a month)
delay	Wait N minutes after anacron starts (spreads the load)
job-id	Name; the date of the last run is saved in `/var/spool/anacron/JOB-ID`
command	What to run

START_HOURS_RANGE=3-22 in the same file limits the hours when anacron may start jobs.

4.3

List and read systemd timers

Many RHEL jobs are timers today: dnf-makecache, logrotate, fstrim, systemd-tmpfiles-clean. list-timers shows when each one runs next.

$systemctl list-timers
NEXT                         LEFT        LAST                         PASSED   UNIT                         ACTIVATES
Mon 2026-10-27 13:02:11 +03  21min       Mon 2026-10-27 12:02:11 +03  38min    dnf-makecache.timer          dnf-makecache.service
Tue 2026-10-28 00:00:00 +03  11h         Mon 2026-10-27 00:00:00 +03  12h      logrotate.timer              logrotate.service
Tue 2026-10-28 08:15:00 +03  19h         Mon 2026-10-27 08:15:00 +03  4h 25min systemd-tmpfiles-clean.timer systemd-tmpfiles-clean.service
$systemctl list-timers --all
$systemctl cat logrotate.timer

4.4

Write your own system timer

A timer needs two units with the same name: NAME.service (what to run) and NAME.timer (when). Put them in /etc/systemd/system/.

/etc/systemd/system/backup.service

[Unit]
Description=Nightly backup of /etc and /srv
 
[Service]
Type=oneshot
ExecStart=/usr/local/bin/backup.sh
Nice=10

/etc/systemd/system/backup.timer

[Unit]
Description=Run the backup every night
 
[Timer]
OnCalendar=*-*-* 01:30:00
RandomizedDelaySec=15min
Persistent=true
 
[Install]
WantedBy=timers.target

#systemctl daemon-reload
#systemctl enable --now backup.timer
#systemctl list-timers backup.timer
#systemctl start backup.service
#journalctl -u backup.service -n 20

Exam tip

You enable the timer, not the service. Start the service by hand once (systemctl start backup.service) to test it.

4.5

Write the right time for OnCalendar

OnCalendar uses the form DayOfWeek Year-Month-Day Hour:Minute:Second. * means "every". systemd-analyze calendar checks it for you.

Value	Runs
`daily` / `weekly` / `monthly`	At 00:00 every day / Monday / first day
`--* 02:30:00`	Every day at 02:30
`Mon..Fri 08:00`	Work days at 08:00
`--01 04:00`	The first day of every month at 04:00
`*:0/15`	Every 15 minutes
`Sat,Sun 10:00`	Weekends at 10:00

$systemd-analyze calendar 'Mon..Fri 08:00'
  Original form: Mon..Fri 08:00
Normalized form: Mon..Fri *-*-* 08:00:00
    Next elapse: Tue 2026-10-28 08:00:00 +03
$systemd-analyze calendar --iterations=3 '*:0/15'

Other timer options: OnBootSec=10min (10 minutes after boot), OnUnitActiveSec=1h (one hour after the last run).

4.6

Manage temporary files with systemd-tmpfiles

systemd-tmpfiles creates folders and files at boot and deletes old files on a schedule. Rules are lines in tmpfiles.d files.

/etc/tmpfiles.d/myapp.conf

# Type  Path               Mode  User   Group  Age
d       /run/myapp         0755  myapp  myapp  -
d       /var/tmp/myapp     0750  myapp  myapp  7d
f       /run/myapp/ready   0644  myapp  myapp  -

Type	Meaning
`d`	Create the folder; with an age, delete files in it that are older
`D`	Like `d`, but empty it at every boot
`f` / `F`	Create a file / create it and empty it
`L`	Create a symbolic link
`x`	Do not clean this path (an exception)

#systemd-tmpfiles --create /etc/tmpfiles.d/myapp.conf
#systemd-tmpfiles --clean /etc/tmpfiles.d/myapp.conf
$cat /usr/lib/tmpfiles.d/tmp.conf
q /tmp 1777 root root 10d
q /var/tmp 1777 root root 30d
$systemctl status systemd-tmpfiles-clean.timer

Remember

Package rules are in /usr/lib/tmpfiles.d/. Your own rules (and changes) go to /etc/tmpfiles.d/ — a file with the same name there replaces the package file.

Exam tip

"Files in /var/tmp/myapp older than 7 days must be removed" → one d line with 7d in /etc/tmpfiles.d/, then test with systemd-tmpfiles --clean.

Practice

Lab 12.1Schedule jobs with cron, at and a timerMedium · 30 minNot started

Practice

All commands in this chapter