Process Automation

Previous Next

This final session covers how to manage long-running tasks and schedule scripts to run automatically at predefined times, which is crucial for system administration and automation.

Learning Objectives

  • Understand Unix process states and how to view them (ps, top).
  • Learn techniques for running commands reliably in the background (nohup, &).
  • Master the use of crontab for scheduling recurring tasks.
  • Understand the logging implications of automated jobs.

Topics Covered

1. Process Control and Signals

  • Viewing Processes:
  • ps aux: Full process list.
  • top / htop: Real-time monitoring.
  • pgrep / pkill: Finding or killing processes by name.
  • Signals (Advanced):
  • SIGINT (2): Interrupt (Ctrl+C).
  • SIGTERM (15): Graceful termination.
  • SIGKILL (9): Immediate forced termination.
  • The trap Command (Advanced): Handling signals within scripts for cleanup.

bash trap "echo 'Cleaning up...'; rm -f temp_file; exit" SIGINT SIGTERM

  • Foreground/Background: fg, bg, and the & operator.
  • Reliability: Using nohup or tmux/screen for persistent sessions.

2. Task Scheduling with cron

  • crontab Access: crontab -e (edit), crontab -l (list).
  • Cron Syntax: Minute Hour DayOfMonth Month DayOfWeek command.
  • Advanced Scheduling:
  • */15 * * * *: Every 15 minutes.
  • 0 9-17 * * 1-5: Hourly during work hours (Mon-Fri).
  • anacron (Less Common): Useful for servers or laptops that are not running 24/7; ensures jobs run even if the system was off at the scheduled time.

3. Script Automation Context

  • Environment in cron: Understanding that cron jobs run with a minimal, non-interactive environment. This often means using absolute paths for commands and scripts inside crontabs.
  • Redirecting Output in Cron: Explicitly redirecting standard output and error to log files (>> logfile 2>&1).

Lab/Assessment Focus

Goal: Automate a robust background task.

  1. Script with Cleanup: Create monitor.sh that writes system stats to a log file. Use trap to ensure it logs "Monitoring stopped" when interrupted.
  2. Background Execution: Run monitor.sh using nohup and verify it continues after closing the terminal.
  3. Process Control: Find the PID of monitor.sh using pgrep and terminate it gracefully using kill -15.
  4. Cron Automation: Schedule a job that runs every Sunday at midnight to archive the logs created during the week. Use absolute paths and redirect both stdout and stderr to a central log file.

Advanced Topic References

SQL Intro AI Admin