Hi,
Last week, when I was on vacation, the server started to freeze due to 100% cpu usage. There was no updates, config changes or new cron jobs since previous package updates (about 1.5months). The syslogs show nothing suspicious or any kind of errors. It is also not triggered by the cron job, because there are no traces of any job running before the crash. It seems like some kind of memory leak, but why it just started now and not after updates or server reboot? Is there any way to know which process ran away?
The VPS is hosted on Hetzner, which is pretty stable (other VPSes running just fine).
Thanks
This matter of servers acting up when we go on holidays has been an industry wide problem for decades now. I wish someone would do something about it. It could be a billion dollar opportunity.
The top command displays realtime info about resources consumed and which process is using what. The atop command can be invoked to show historical data from the logs. By using the latter command (with appropriate command like parameters) you will know exactly which process was consuming what at the time when the server acted up.
Thanks Caport. Unfortunately, I was unable to login without server hard reboot to see live data. It seems atop is not installed on Ubuntu by default, so I guess it won’t be able to read historical data. Anyways, few days later things calmed down and no problems since. I was looking for traces of DoS attacks, but fail2ban/firewall logs showed nothing. I guess I will install atop just in case…
Cheers,