Common causes of server unavailability

Summary

A server can become unavailable for several reasons: out of memory (OOM), full disk, network issues, DDoS attack, or OS freeze. Below are the main causes and initial diagnostic steps for each.

Applies to:
✔ VPS
✔ Dedicated servers
✔ Linux, Windows

Out of memory (OOM)

When RAM runs out, the Linux kernel forcefully terminates processes (OOM Killer). This can affect SSH, the web server, or the database.

Symptoms:

• Processes restart or disappear
• Log entries containing Out of memory

Diagnostics (via VNC/IPMI console):

dmesg | grep -i "out of memory"

free -m

Find the process consuming the most memory:

ps aux --sort=-%mem | head -10

Disk full (100%)

Without free disk space, services cannot write logs, create temporary files, or PID files. SSH may stop working.

Symptoms:

• "No space left on device" errors
• Services fail to start

Diagnostics:

df -h

Quick cleanup:

journalctl --vacuum-time=1d

apt clean

DDoS attack

During a DDoS attack, the server receives an abnormally large volume of traffic. The network becomes unavailable or works with delays.

Symptoms:

• Server is accessible via VNC/IPMI but not over the network
• High load on the network interface

Diagnostics:

iftop -n

netstat -ntu | awk '{print $5}' | cut -d: -f1 | sort | uniq -c | sort -rn | head -20

Network issues

Routing problems, issues on the data center or ISP side.

Symptoms:

• Server works via console but doesn't respond over the network
• Packet loss at intermediate nodes (visible via MTR)

Diagnostics from another device:

mtr -r SERVER_IP

On the server via console:

ip a

ip route

ping 8.8.8.8

Operating system freeze

The OS kernel can freeze due to hardware issues, driver errors, or kernel panic.

Symptoms:

• Console doesn't respond to input
• IPMI/VNC screen shows kernel panic or a static image

Solution: reboot the server through the control panel or IPMI.

SSH/RDP service not running

The server may be working, but the remote access service is stopped.

Diagnostics via VNC/IPMI:

systemctl status sshd

If not running:

systemctl start sshd

Verification

After identifying and resolving the cause:

ping -c 4 SERVER_IP

ssh root@SERVER_IP

If you cannot determine the cause or the server remains unavailable, open a support ticket. Include:
- server IP address
- what you see in the VNC/IPMI console
- MTR results from your computer