Master Linux Local Mail, at/batch, and cron: A Practical Guide

1. Introduction

Linux provides three task‑scheduling commands: at, batch, and cron. at and batch execute a job once, while cron runs jobs periodically; after a job finishes, the result is emailed to the submitter.

2. Local Mail Service

CentOS installs a local mail service that listens on port 25. You can send mail locally using the mail command. Two methods are shown:

1) Interactive entry: mail username (or username@hostname), then follow prompts for subject and body, finish with CTRL+D.

2) Using a file or pipe: mail -s "subject" user< file or echo "content" | mail -s "subject" user List received messages with mail and read a specific message by its number.

3. at / batch

Schedule a one‑time job with at now + count time‑units (units can be minutes, hours, days) or with an exact time such as hh:mm MMDDYY. Edit the job, finish with CTRL+D. Example screenshots show a task scheduled for a specific time.

View pending jobs with atq or at -l. Cancel a pending job with atrm jobID. The batch command works like at but runs automatically when the system is idle.

4. cron

The cron daemon runs continuously in the background. Verify it is active with service crond status; if it shows “stopped”, start it with service crond start. Enable it at boot with chkconfig crond on.

System‑wide crontab is /etc/crontab; user crontabs reside in /var/spool/cron and have the same format.

Crontab syntax: minute hour day month weekday command. Examples:

7 6 5 4 *   /path/to/script   # runs at 06:07 on April 5 each year

*/3 * * * *   /bin/echo "hello"   # every 3 minutes

20-40/3 * * * *   /path/to/script   # every 3 minutes between minute 20 and 40 each hour

Common crontab options: -l list jobs, -e edit (syntax checking is provided), -r remove all jobs. Administrators can edit a user's crontab with crontab -u USERNAME -e.

5. Summary

The article described how to create, modify, and delete scheduled tasks using at, batch, and cron, and how to use the local mail service to receive job results. It also reminded that in Linux everything is a file and that second‑level scheduling requires supplemental use of sleep.