Troubleshooting a Unix System

1. Identify the Problem

• Ask: What’s wrong? (slow, unresponsive, network down, process not working, boot failure, etc.)
• Get error messages (logs, user reports, console output).
• Note when it started and any recent changes (updates, config edits, hardware swaps).

2. Check System Health

Run basic commands:

uptime          # load average, uptime, logged-in users
top or htop     # CPU, memory, running processes
free -m         # memory usage
df -h           # disk usage
du -sh /path/*  # find large directories
iostat, vmstat  # I/O bottlenecks

Look for:

• High CPU or memory usage
• Disk full or high I/O wait
• Swapping (memory exhaustion)

3. Check Processes & Services

Find misbehaving processes:

ps aux --sort=-%cpu | head
ps aux --sort=-%mem | head

Restart or kill problematic processes:

kill -9 <pid>
systemctl restart <service>

Check open ports and network services:

netstat -tulnp
ss -ltnp

4. Check Logs

System logs:

less /var/log/syslog    # Debian/Ubuntu
less /var/log/messages  # RHEL/CentOS
journalctl -xe

Application/service logs (e.g., Apache: /var/log/httpd/).

Look for errors, warnings, crashes.

5. Check Network

Verify connectivity:

ping 8.8.8.8
curl -I http://example.com

Check interface status:

ip addr
ip route

DNS issues:

dig example.com
nslookup example.com

6. Hardware Checks

Disk health:

smartctl -a /dev/sda
dmesg | grep -i error

Memory test (requires reboot):

memtest86+

CPU temperature / sensors:

sensors

7. Boot & Filesystem Problems

• If system won’t boot, use single-user mode or rescue mode.
• Check filesystem:

fsck /dev/sda1

Rebuild initramfs or grub if necessary.

8. Security Checks

Look for suspicious logins:

last
w

Check for unusual processes:

ps aux | grep -v root
lsof -i

Verify permissions and firewall rules:

iptables -L -n

9. If Still Stuck

• Isolate: Is it hardware, OS, or application?
• Roll back recent changes (packages, configs).
• Search logs & errors online.
• Check vendor/OS documentation.

Rule of Thumb

✅ Start broad (system load, health, logs), then drill down (processes, services, hardware).