Shell Scripting Complete 13: Scheduling Scripts with cron and systemd Timers

By Ravindra Bagale, trainer LinkedIn Last updated: 25 October 2026

A script is most useful when it runs by itself: a backup at 2 a.m., a report every Monday, a health check every 5 minutes. Linux has two main schedulers. cron is old, simple and everywhere. systemd timers are newer, log to journald, can catch up on missed runs and give resource limits. In this chapter you learn crontab syntax, the cron environment, logging and locking cron jobs, then .service and .timer units, OnCalendar, Persistent, RandomizedDelaySec, systemctl list-timers and journalctl, and how to choose between them. Every topic starts with an everyday example and then shows examples from many industries.

What you'll learn in this chapter

Read and write crontab lines
Fix the cron environment: PATH, shell, mail
Log and lock cron jobs
Write a systemd service and timer
Use OnCalendar, Persistent and delays
Choose cron or timers

1. An everyday example: the milkman and the alarm clock

Why do we need this?

In Pune the milkman comes at 6 every morning without being called. Your alarm rings at 6:30 on weekdays only. If you were travelling, a smart alarm could tell you what you missed. cron is the milkman: fixed time, no questions. A systemd timer is the smart alarm: it keeps a record, can run a missed job when the machine wakes, and writes everything in a diary.

What is it?

Milkman's fixed time = a crontab line. Weekdays only = day-of-week field 1-5. Smart alarm = .timer unit. The task it starts = .service unit. Running a missed job later = Persistent=true. Diary = journalctl.

How to do it (step by step)

Cron and timers run with a tiny environment, not your login shell.

2. crontab syntax

Why do we need this?

You must read and write schedules like 'every 15 minutes' or 'last working day at 7 p.m.' without mistakes.

What is it?

Five time fields then the command: minute (0-59), hour (0-23), day of month (1-31), month (1-12), day of week (0-7, 0 and 7 are Sunday). * means every, */15 means every 15, 1-5 a range, 1,15 a list. Shortcuts: @reboot, @daily, @hourly, @weekly. If both day fields are set, cron runs when either matches. In crontab lines, % means newline, so escape it as \% (for example in date +\%F).

Line	Runs
/5 * * *	every 5 minutes
0 2 * * *	2:00 every day
30 18 * * 1-5	6:30 p.m. weekdays
0 9 1 * *	9 a.m. on the 1st
@reboot	once at boot

How to do it (step by step)

Edit.

crontab -e

What it means: Opens your crontab in an editor.

List.

crontab -l

Output:

0 2 * * * /opt/bin/backup.sh

What it means: Your jobs.

Every 15 min.

*/15 * * * * /opt/bin/health.sh

What it means: Minutes 0, 15, 30, 45.

Monday report.

0 8 * * 1 /opt/bin/weekly-report.sh

What it means: 8 a.m. Monday.

Escape percent.

0 1 * * * tar czf /backup/etc-$(date +\%F).tgz /etc

What it means: Without \ the line breaks.

3. The cron environment

Why do we need this?

'It works when I run it but not in cron' is the most common cron problem.

What is it?

cron runs with a tiny PATH (often /usr/bin:/bin), sh as the shell, your home as working dir, and no .bashrc. So: use full paths or set PATH at the top of the crontab, set SHELL=/bin/bash if needed, cd inside the script, and load env files yourself. MAILTO=you@x sends output by mail if mail is set up; MAILTO="" turns it off. System jobs go in /etc/cron.d files with an extra user field.

How to do it (step by step)

Crontab header.

SHELL=/bin/bash
PATH=/usr/local/bin:/usr/bin:/bin
MAILTO=""

What it means: Top of crontab -e.

See cron env.

* * * * * env > /tmp/cron-env.txt

Output:

PATH=/usr/bin:/bin
SHELL=/bin/sh
HOME=/home/ec2-user

What it means: Compare with yours.

System job.

echo '0 3 * * * root /opt/bin/rotate.sh' | sudo tee /etc/cron.d/rotate

Output:

0 3 * * * root /opt/bin/rotate.sh

What it means: User field after time.

Is cron running?

systemctl status crond   # Ubuntu: cron

Output:

Active: active (running)

What it means: Service must be up.

4. Log and lock cron jobs

Why do we need this?

Cron output goes nowhere by default, and slow jobs can overlap.

What is it?

Append both streams to a log: >> /var/log/job.log 2>&1. Or send to journald with | logger -t job. Wrap with flock -n so a second copy skips. Add timeout so a hung job does not run forever. Check that cron itself fired with journalctl -u crond or grep CRON /var/log/syslog.

How to do it (step by step)

Full safe line.

*/10 * * * * flock -n /tmp/sync.lock timeout 9m /opt/bin/sync.sh >> /var/log/sync.log 2>&1

What it means: Lock, timeout, log.

To journald.

0 2 * * * /opt/bin/backup.sh 2>&1 | logger -t backup

What it means: Read with journalctl -t backup.

Did cron fire?

journalctl -u crond --since '1 hour ago' | grep sync

Output:

Oct 25 10:10:01 host CROND[4410]: (ec2-user) CMD (flock -n /tmp/sync.lock ...)

What it means: Yes at 10:10.

Ubuntu syslog.

grep CRON /var/log/syslog | tail -2

Output:

Oct 25 10:10:01 host CRON[4410]: (ubuntu) CMD (/opt/bin/sync.sh)

What it means: Same check.

5. systemd service + timer

Why do we need this?

Timers give you logs in journald, exit status, retries and limits, all with systemctl commands you already know.

What is it?

You write two files in /etc/systemd/system. name.service says what to run (Type=oneshot, ExecStart=/opt/bin/script.sh, User=). name.timer says when (OnCalendar=) and has WantedBy=timers.target. Then systemctl daemon-reload and systemctl enable --now name.timer. Run the job by hand with systemctl start name.service. User timers go in ~/.config/systemd/user with systemctl --user.

How to do it (step by step)

Service file.

sudo tee /etc/systemd/system/backup.service > /dev/null <<'EOF'
[Unit]
Description=Nightly backup
[Service]
Type=oneshot
User=backup
ExecStart=/opt/bin/backup.sh
EOF

What it means: What to run.

Timer file.

sudo tee /etc/systemd/system/backup.timer > /dev/null <<'EOF'
[Unit]
Description=Run backup daily at 02:00
[Timer]
OnCalendar=*-*-* 02:00:00
Persistent=true
[Install]
WantedBy=timers.target
EOF

What it means: When to run.

Enable.

sudo systemctl daemon-reload && sudo systemctl enable --now backup.timer

Output:

Created symlink /etc/systemd/system/timers.target.wants/backup.timer

What it means: Scheduled.

Run now.

sudo systemctl start backup.service; systemctl status backup.service --no-pager | head -3

Output:

backup.service - Nightly backup
   Active: inactive (dead) since Sun 2026-10-25 11:20:04 IST

What it means: oneshot ended.

6. OnCalendar, Persistent and delays

Why do we need this?

Laptops sleep, servers reboot, and 500 servers hitting one API at 02:00:00 is a storm.

What is it?

OnCalendar uses DayOfWeek Year-Month-Day Hour:Minute:Second, like Mon..Fri *-*-* 18:30 or *-*-01 09:00 or shortcuts daily, weekly, hourly. Check it with systemd-analyze calendar. OnBootSec=5min and OnUnitActiveSec=15min run relative to boot or last run. Persistent=true runs a missed job at next boot. RandomizedDelaySec=10m spreads servers out. AccuracySec sets how exact it must be.

Setting	Meaning
OnCalendar=daily	midnight every day
OnCalendar=Mon..Fri 18:30	weekdays 6:30 p.m.
OnUnitActiveSec=15min	15 min after last run
Persistent=true	catch up missed run
RandomizedDelaySec=10m	random 0-10 min delay

How to do it (step by step)

Check a calendar.

systemd-analyze calendar 'Mon..Fri 18:30'

Output:

  Original form: Mon..Fri 18:30
Normalized form: Mon..Fri *-*-* 18:30:00
    Next elapse: Mon 2026-10-26 18:30:00 IST

What it means: Next run shown.

Every 15 minutes.

OnCalendar=*:0/15

What it means: Like */15 in cron.

First of month.

OnCalendar=*-*-01 09:00:00

What it means: Monthly invoice run.

Spread a fleet.

[Timer]
OnCalendar=daily
RandomizedDelaySec=30m

What it means: No thundering herd.

7. Watch timers: list-timers and journalctl

Why do we need this?

You must know when the job ran last, when it runs next and whether it worked.

What is it?

systemctl list-timers shows NEXT, LEFT, LAST and PASSED for every timer. journalctl -u name.service shows the script output for each run. systemctl status name.service shows the last exit code. Add OnFailure=alert@%n.service in [Unit] to run an alert unit when the job fails.

How to do it (step by step)

Next runs.

systemctl list-timers backup.timer --no-pager

Output:

NEXT                        LEFT     LAST                        PASSED  UNIT
Mon 2026-10-26 02:00:00 IST 14h left Sun 2026-10-25 02:00:03 IST 9h ago  backup.timer

What it means: Last and next.

Output of runs.

journalctl -u backup.service --since yesterday --no-pager | tail -2

Output:

Oct 25 02:00:41 host backup.sh[1201]: backup ok 2.1G
Oct 25 02:00:41 host systemd[1]: backup.service: Deactivated successfully.

What it means: Script output in journal.

Failed?

systemctl is-failed backup.service

Output:

inactive

What it means: Not failed.

Alert on failure.

[Unit]
OnFailure=notify-fail@%n.service

What it means: Runs another unit on failure.

8. cron or systemd timer?

Why do we need this?

Both work; the right choice saves you trouble later.

What is it?

Use cron for simple per-user jobs, on old or minimal systems, or when the team already knows it. Use timers when you want journald logs, catch-up after downtime, random delay across a fleet, dependencies (After=network-online.target), resource limits (MemoryMax=, CPUQuota=) and status in systemctl. Many teams move important server jobs to timers and keep cron for quick personal tasks.

Need	cron / timer
Simple and everywhere	yes / mostly
Logs in journald	manual / built in
Catch up missed run	no (anacron) / Persistent
Random delay	manual sleep / RandomizedDelaySec
Resource limits	no / MemoryMax, CPUQuota

How to do it (step by step)

Limit memory.

[Service]
MemoryMax=512M
CPUQuota=50%

What it means: Job cannot eat the server.

Wait for network.

[Unit]
After=network-online.target
Wants=network-online.target

What it means: No early failures at boot.

One-off later.

sudo systemd-run --on-calendar='2026-10-26 01:00' /opt/bin/migrate.sh

Output:

Running timer as unit: run-u91.timer

What it means: Like at, with logs.

9. Real examples from many industries

Why do we need this?

The same Linux skill shows up in every kind of company. Seeing it in many places helps you remember it.

What is it?

Each row is a made-up team. It shows who uses this and how.

How to do it (step by step)

Team and city	How they use it	Command
Shruti, bank, Pune	Nightly EOD backup at 2 a.m.	`0 2 * * * /opt/bin/eod-backup.sh`
Tom, newspaper, Manchester	Publish e-paper weekdays 5 a.m.	`OnCalendar=Mon..Fri 05:00`
Wei, IoT, Shanghai	Sensor upload every 15 min with lock	`/15 * * * flock -n /tmp/u.lock /opt/bin/up.sh`
Noura, retail, Doha	Fleet-wide cache warm with spread	`RandomizedDelaySec=20m`
Pedro, hospital, Lisbon	Catch up missed audit export	`Persistent=true`
Mangesh, APMC market, Nashik	Daily price list at 6 a.m.	`0 6 * * * /opt/bin/prices.sh >> /var/log/prices.log 2>&1`

10. Common errors and how to fix them

Why do we need this?

Everyone hits these errors. Knowing the cause saves hours.

What is it?

Read the exact message first. Then match it to a row below.

How to do it (step by step)

Error you see	Why it happens	How to fix
`Job never runs`	cron service down or bad line	Check service and syntax. `systemctl status crond`
`command not found in cron only`	Small PATH	Full paths or PATH line. `PATH=/usr/local/bin:/usr/bin:/bin`
`Command cut at % sign`	% means newline in crontab	Escape as \%. `date +\%F`
`No output anywhere`	No redirect, no mail	Redirect to a log. `>> /var/log/job.log 2>&1`
`bad minute / errors in crontab file`	Wrong field count	Five fields then command. `crontab -e`
`Timer not in list-timers`	Not enabled or no daemon-reload	Reload and enable. `sudo systemctl daemon-reload && sudo systemctl enable --now x.timer`
`Failed to parse calendar specification`	Bad OnCalendar	Test the expression. `systemd-analyze calendar 'Mon --* 08:00'`
`Service runs as root unexpectedly`	No User= line	Add User=. `User=backup`
`Overlapping runs`	Slow job, short interval	flock or OnUnitActiveSec. `flock -n /tmp/j.lock ./j.sh`

11. Troubleshooting step by step

Why do we need this?

Find out if the scheduler fired, then if the script failed, then why.

What is it?

Go top to bottom. Stop at the first step that shows the problem.

How to do it (step by step)

Cron fired?

journalctl -u crond --since today | tail

Timer status.

systemctl list-timers --all

Last run output.

journalctl -u backup.service -n 30

Exit status.

systemctl status backup.service

Same env test.

env -i PATH=/usr/bin:/bin /bin/sh -c /opt/bin/backup.sh

Calendar check.

systemd-analyze calendar daily

12. Lab 1: A safe cron job with log and lock

Why do we need this?

Schedule a small disk report every minute with flock, timeout and a log, confirm it runs, then remove it.

What is it?

A full hands-on lab. Run every step on a test Linux machine (an EC2 instance, a VM, or WSL).

How to do it (step by step)

Script.

mkdir -p ~/bin && cat > ~/bin/diskrep.sh <<'X'
#!/usr/bin/env bash
set -euo pipefail
printf '%(%F %T)T ' -1; df -h / | tail -1
X
chmod +x ~/bin/diskrep.sh && ~/bin/diskrep.sh

Output:

2026-10-25 11:40:02 /dev/xvda1 20G 8.1G 12G 41% /

Add to crontab.

( crontab -l 2>/dev/null; echo "* * * * * flock -n /tmp/diskrep.lock timeout 50 $HOME/bin/diskrep.sh >> $HOME/diskrep.log 2>&1" ) | crontab -

Check line.

crontab -l | tail -1

Output:

* * * * * flock -n /tmp/diskrep.lock timeout 50 /home/ec2-user/bin/diskrep.sh >> /home/ec2-user/diskrep.log 2>&1

Wait 2 min and read.

sleep 125; cat ~/diskrep.log

Output:

2026-10-25 11:41:01 /dev/xvda1 20G 8.1G 12G 41% /
2026-10-25 11:42:01 /dev/xvda1 20G 8.1G 12G 41% /

Remove.

crontab -l | grep -v diskrep.sh | crontab -; rm ~/diskrep.log

13. Lab 2: The same job as a systemd timer

Why do we need this?

Turn the disk report into a service and timer that runs every 2 minutes, see it in list-timers and journalctl, then remove it.

What is it?

A full hands-on lab. Run every step on a test Linux machine (an EC2 instance, a VM, or WSL).

How to do it (step by step)

Copy script.

sudo install -m 755 ~/bin/diskrep.sh /usr/local/bin/diskrep.sh

Service.

sudo tee /etc/systemd/system/diskrep.service > /dev/null <<'EOF'
[Unit]
Description=Disk report
[Service]
Type=oneshot
ExecStart=/usr/local/bin/diskrep.sh
EOF

Timer.

sudo tee /etc/systemd/system/diskrep.timer > /dev/null <<'EOF'
[Unit]
Description=Disk report every 2 minutes
[Timer]
OnCalendar=*:0/2
Persistent=true
[Install]
WantedBy=timers.target
EOF

Enable and list.

sudo systemctl daemon-reload && sudo systemctl enable --now diskrep.timer && systemctl list-timers diskrep.timer --no-pager

Output:

NEXT                        LEFT       UNIT
Sun 2026-10-25 11:52:00 IST 1min left  diskrep.timer

Logs after a run.

sleep 130; journalctl -u diskrep.service -n 2 --no-pager

Output:

Oct 25 11:52:00 host diskrep.sh[3101]: 2026-10-25 11:52:00 /dev/xvda1 20G 8.1G 12G 41% /
Oct 25 11:52:00 host systemd[1]: diskrep.service: Deactivated successfully.

Remove.

sudo systemctl disable --now diskrep.timer && sudo rm /etc/systemd/system/diskrep.{service,timer} && sudo systemctl daemon-reload

Output:

Removed /etc/systemd/system/timers.target.wants/diskrep.timer.

Ravindra Bagale's Tip

Before you schedule anything, make the script work with full paths, set -euo pipefail, a log and a flock guard. Then test the schedule: crontab -l for cron, systemd-analyze calendar and list-timers for timers. Always check the first real run in the log instead of assuming it worked.

14. Frequently asked questions

Q. Cron time zone?

The server's zone unless CRON_TZ is set.

Q. Run every 30 seconds?

Timer with OnUnitActiveSec=30s.

Q. Where are user crontabs?

/var/spool/cron.

Q. What is anacron?

Runs daily jobs missed while off.

Q. Timer for a user?

systemctl --user and loginctl enable-linger.

Q. Test a timer now?

systemctl start x.service.

Q. Day 0 or 7?

Both Sunday.

Q. Remove all my cron jobs?

crontab -r, careful.

15. Interview questions and answers

Beginner level

Q. Edit crontab?

crontab -e.

Q. Fields in cron?

Five, then command.

Q. Every 5 minutes?

*/5 * * * *.

Q. List timers?

systemctl list-timers.

Mid level level

Q. Why fails only in cron?

PATH and env differ.

Q. Timer needs which files?

.service and .timer.

Q. Catch up missed run?

Persistent=true.

Q. Percent in crontab?

Escape as \%.

Senior level

Q. Choose timers when?

Need logs, catch-up, limits, dependencies.

Q. Avoid fleet storms?

RandomizedDelaySec or sleep $((RANDOM%600)).

Q. Alert on failure?

OnFailure= unit or exit code monitor.

Q. Stop overlap in timers?

A oneshot service is not started twice while running.

Try at home

Write a crontab line that runs a script at 7:30 p.m. on every weekday, Monday to Friday, and check the line carefully.
Use systemd-analyze calendar to test three schedules: every Sunday 10 a.m., 1st of every month, every 10 minutes.

Practice Tasks

About this scenario

Ahmednagar Sugar Co-op is a sugar co-operative business in Ahilyanagar. Balasaheb Pawar looks after its IT and cloud work. They count money in INR and run their AWS resources in the ap-south-1 Region. In the tasks below you do real jobs for them using what this lesson teaches. All names, account IDs, IP addresses and numbers are made up for practice.

0 / 21