server is unresponsive ever 8-12 hours

[agrothe]

Ever since upgrading to Plesk 11.5.30 I have been having regular issues with my Ubuntu 12.04.3 LTS server becoming completely unresponsive. It seems to happen anywhere from once to three times a day.

Error logs show no errors as such, except there is a call to backupmng from cron either right before or a few minutes before the crash. This is the only commonality I can find.

For example, my last crash has this in the syslog:

Oct 12 11:53:46 topaz courier-pop3d: Connection, ip=[::ffff:127.0.0.1]
Oct 12 11:53:46 topaz courier-pop3d: LOGOUT, ip=[::ffff:127.0.0.1]
Oct 12 11:53:46 topaz courier-pop3d: Disconnected, ip=[::ffff:127.0.0.1]
Oct 12 11:53:46 topaz courier-pop3s: Connection, ip=[::ffff:127.0.0.1]
Oct 12 11:53:46 topaz courier-pop3s: LOGOUT, ip=[::ffff:127.0.0.1]
Oct 12 11:53:46 topaz courier-pop3s: Disconnected, ip=[::ffff:127.0.0.1]
Oct 12 11:54:01 topaz CRON[10217]: (root) CMD ([ -x /opt/psa/admin/sbin/backupmng ] && /opt/psa/admin/sbin/backupmng >/dev/null 2>&1)
Oct 12 15:43:35 topaz kernel: imklog 5.8.6, log source = /proc/kmsg started.

That call to CRON is found within a few minutes of each time the system becomes unresponsive.

My hosting provider found a few RAM modules which needed replacing, but a few hours after they took the server offline and did a full hardware check and replaced a faulty RAM module the system went unresponsive again.

This is getting pretty frustrating at this point. Hopefully someone can help shed some light on the situation. I just disabled the above cron job in /etc/cron.d/plesk-backup-manager-task so I will see what happens tomorrow. Others have reported that not working so I'm not holding my breath.

My full syslog is here for the time being in case it helps: http://bordereastcreative.com/syslog

 

[agrothe]

The cron job was not the issue. Not sure what to try next.

 

[agrothe]

I stopped the Plesk service and the issue stopped as well.

How can I debug the Plesk service to find the issue? Is there a verbose logging command?

Not sure if that would help either, as the server going unresponsive would probably stop the logging from happening as well.

Any thoughts on how to proceed?

 

[InderS]

Hello,

I will suggest you monitor your server through SSH using the top command so that you can find out which process are running on the server and using the high server resources

 

[agrothe]

Hello,

I will suggest you monitor your server through SSH using the top command so that you can find out which process are running on the server and using the high server resources

I did that. Nothing was using more than 1.2% of RAM or CPU. I used the newrelic system monitor as well and nothing reported as using high server resources. Nothing in logs to indicate an error, just a random and frequent completely unresponsive server.

I'm at the point where I just want to uninstall and switch to ISPConfig or something.

 

[InderS]

Have you attached the KVM console to your server ?

Try to attach the KVM console to your server and check the KVM consol screen when your server is not responding from your end. You will get some massages on your KVM screen