SME Server Quick Health Check

A safe, mostly read-only checklist for a quick snapshot of server health.

Identity and uptime

uname -a
uptime

Look for:

• Very high load averages (relative to CPU cores)
• Recent unexpected reboots

Disk space

df -h

Rule of thumb:

• Anything above 85% deserves attention
• Anything above 95% is an emergency

Memory and swap

free -m

Notes:

• Linux uses RAM for cache, so free is often low
• Pay attention to available and swap used
• Sustained swap growth can indicate memory pressure

Failed services

systemctl --failed

This should ideally say 0 loaded units listed.

SMART disk health

smartctl -H /dev/sda

If SMART does not say PASSED, investigate immediately.

Tip: list disks first if you are unsure of device names:

lsblk

Recent errors in logs

journalctl -p 3 -xb

Or (older style):

grep -i error /var/log/messages | tail -50

Common noise:

• ACPI BIOS warnings at boot are often harmless on older hardware

Network listening services

ss -tulnp

This gives a quick “what is exposed” view.

What to record (baseline)

When the server is healthy, save these outputs somewhere:

• uname -a
• df -h
• free -m
• systemctl --failed
• smartctl -H /dev/sda

Baselines make later troubleshooting much faster.