System Administration Automation

# CHAPTER 15

System Administration Automation #

1. Introduction #

2. Learning Objectives #

• Synthesize loops and conditions into multi-stage administrative tools.

• Write a script to autonomously read a text file and provision multiple user accounts.

• Write a monitoring script utilizing df, awk, and mathematical conditions to detect full hard drives.

• Write a self-healing script that checks the status of services and restarts them if they fail.

• Generate automated, timestamped system health reports.

3. Bulk User Management Script #

The Workflow:

1. 1. The script reads new_hires.txt.

1. 2. It loops through each name.

1. 3. It checks if the user already exists (to prevent errors).

1. 4. If they don't exist, it creates the user.

#!/bin/sh

# Ensure script is run as root
if [ "$(id -u)" -ne 0 ]; then
  echo "Please run as root (sudo)"
  exit 1
fi

FILE="new_hires.txt"

# Loop through the file line by line
while read USERNAME; do
    # Check if user exists by searching /etc/passwd quietly
    if grep -q "^$USERNAME:" /etc/passwd; then
        echo "User $USERNAME already exists. Skipping."
    else
        echo "Creating user $USERNAME..."
        useradd -m -s /bin/bash "$USERNAME"
    fi
done < "$FILE"

*(Note the syntax at the very bottom: < "$FILE". This feeds the physical text file directly into the while loop's read command!)*

4. Disk Space Monitoring Alert #

The Workflow:

1. 1. Run df -h /.

1. 2. Use awk and sed to extract just the raw percentage number.

1. 3. Compare the number against our 90 threshold.

#!/bin/sh

THRESHOLD=90

# 1. df -h / (Get disk info)
# 2. awk '{print $5}' (Grab the 5th column: Use%)
# 3. tail -n 1 (Grab the actual data row, ignore the header)
# 4. sed 's/%//g' (Strip the '%' sign so the shell can do math on the raw number)
USAGE=$(df -h / | awk '{print $5}' | tail -n 1 | sed 's/%//g')

if [ $USAGE -ge $THRESHOLD ]; then
    echo "CRITICAL WARNING: Disk space is at ${USAGE}%!"
    # In a real environment, you would use curl here to send a Slack or Email alert.
else
    echo "Disk space healthy at ${USAGE}%."
fi

5. Self-Healing Service Monitor #

The Workflow:

1. 1. Check the active status of nginx.

1. 2. If the response is not active, restart the service.

1. 3. Log the intervention.

#!/bin/sh

SERVICE="nginx"
LOG="/var/log/service_monitor.log"
TIMESTAMP=$(date +"%Y-%m-%d %H:%M:%S")

# Check if the service is running (Suppress stdout)
systemctl is-active --quiet $SERVICE

# If exit code is not 0, it means it crashed
if [ $? -ne 0 ]; then
    echo "[$TIMESTAMP] ALERT: $SERVICE crashed! Attempting restart..." >> $LOG
    systemctl restart $SERVICE
    
    # Check again to see if the restart worked
    systemctl is-active --quiet $SERVICE
    if [ $? -eq 0 ]; then
        echo "[$TIMESTAMP] SUCCESS: $SERVICE recovered." >> $LOG
    else
        echo "[$TIMESTAMP] FATAL: $SERVICE failed to recover! Manual intervention required." >> $LOG
    fi
fi

6. Diagrams/Visual Suggestions #

• YES -> Arrow loops harmlessly back to the top.

• NO -> Arrow points to a red box: Execute: systemctl restart.

7. Best Practices #

• Idempotency: A professional administrative script should be "Idempotent." This means you can run the script 1 time or 1,000 times, and it will safely result in the exact same state without causing errors or creating duplicates. Notice how our User Management script explicitly checks if grep -q "^$USERNAME:" /etc/passwd first. Without that check, running the script twice would trigger a barrage of messy useradd errors.

8. Common Mistakes #

• Assuming cron has root privileges: A script executing useradd or systemctl restart requires administrative rights. If a standard user schedules this script using their personal crontab -e, it will fail silently. System administration scripts MUST be scheduled in the root user's crontab (sudo crontab -e).

9. Mini Project: Automated System Report #

1. 1. nano system_report.sh

1. 2. Write the code:

#!/bin/sh
REPORT="/tmp/daily_report.txt"

echo "=== DAILY SYSTEM REPORT ===" > $REPORT
echo "Date: $(date)" >> $REPORT
echo "Hostname: $(hostname)" >> $REPORT
echo "" >> $REPORT

echo "--- UPTIME & LOAD ---" >> $REPORT
uptime >> $REPORT
echo "" >> $REPORT

echo "--- MEMORY USAGE ---" >> $REPORT
free -m >> $REPORT

echo "Report generated successfully."
cat $REPORT

10. Practice Exercises #

1. 1. Analyze the Disk Space Monitoring script. Explain the specific text processing function of the sed 's/%//g' command, and why it is mathematically necessary before the if statement can execute.

1. 2. What is "Idempotency" in scripting? Provide an example of how you would make a directory-creation command (mkdir) idempotent.

11. MCQs with Answers #

• Q: Explain the concept of "Idempotent" scripting in Unix administration. Walk me through how you would ensure a shell script tasked with appending a specific configuration line to /etc/ssh/sshdconfig is strictly idempotent.

• Q: A junior engineer is tasked with monitoring a critical background service. They write an infinite while true loop that checks the service status every 1 second. Explain the severe architectural flaw in this design and why scheduling a script via cron is vastly superior for service monitoring.

13. FAQs #

14. Summary #

15. Next Chapter Recommendation #