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
When RAM runs out, the Linux kernel forcefully terminates processes (OOM Killer). This can affect SSH, the web server, or the database.
Symptoms:
Out of memoryDiagnostics (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
Without free disk space, services cannot write logs, create temporary files, or PID files. SSH may stop working.
Symptoms:
Diagnostics:
df -h
Quick cleanup:
journalctl --vacuum-time=1d
apt clean
During a DDoS attack, the server receives an abnormally large volume of traffic. The network becomes unavailable or works with delays.
Symptoms:
Diagnostics:
iftop -n
netstat -ntu | awk '{print $5}' | cut -d: -f1 | sort | uniq -c | sort -rn | head -20
Routing problems, issues on the data center or ISP side.
Symptoms:
Diagnostics from another device:
mtr -r SERVER_IP
On the server via console:
ip a
ip route
ping 8.8.8.8
The OS kernel can freeze due to hardware issues, driver errors, or kernel panic.
Symptoms:
Solution: reboot the server through the control panel or IPMI.
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
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