Plesk problem please help

[PersianC]

Hi,

I have a debian VPS with Plesk control panel. problem is that every day at 3:20 UTC all VPS's services not working (include SSH, Apache, FTP, Mail) and I only be able to login to plesk panel. when I log in I only see 'Plesk Diagnostic' message and what I do is clicking on restart container. when I restart all services works fine and Plesk go back to normal.

I can not find where is the problem. I have around 5~6 hours downtime because this problem. please let me know if you can assist me.

Thanks
Mehdi

 

[IgorG]

Do you have any OOM killer related messages in /var/log/messages on Virtuozzo node at this time? Is there something suspicious in the same log inside Virtuozzo container?

 

[PersianC]

Do you have any OOM killer related messages in /var/log/messages on Virtuozzo node at this time? Is there something suspicious in the same log inside Virtuozzo container?

there is no oom message:

root@na:/# tail -50000 /var/log/messages | grep oom
root@na:/#

and nothing especial in vrituozzo log at these times.

 

[IgorG]

No. It is incorrect grep. There is should be something like:

# grep exc /var/log/messages

Oct 16 11:10:40 vps005-lin64 kernel: kill_signal(108136.2): selected 1, signalled 1, queued 1, seq 370, exc 1302 0 red 79207 1408 Oct 16 11:20:56 vps005-lin64 kernel: kill_signal(108136.2): selecting to kill, queued 0, seq 371, exc 1303 0 goal 1303 0...
Oct 16 11:20:56 vps005-lin64 kernel: kill_signal(108136.2): selected 1, signalled 1, queued 1, seq 371, exc 1303 0 red 79205 1408

The digits after 'exc' means how much total and low memory is not enough and in this case we see that total memory is the resource which server lacks of.

Or something like this:

Apr 16 17:07:52 photon kernel: OOM killed process httpd (pid=24631, ve=196) exited, free=99257 gen=699.
Apr 16 17:07:52 photon kernel: OOM killed process httpd (pid=7113, ve=196) exited, free=99970 gen=700.
Apr 16 17:07:52 photon kernel: OOM killed process httpd (pid=799, ve=196) exited, free=100838 gen=701.

 

[PersianC]

No. It is incorrect grep. There is should be something like:


The digits after 'exc' means how much total and low memory is not enough and in this case we see that total memory is the resource which server lacks of.

Or something like this:

thank you for reply.
no there is not such signal on messages log. I viewed log with nano too. evey no kill signal is here.

 

[IgorG]

Did you checked cron jobs for this time? Did you tried to check system with chkrootkit?

 

[PersianC]

Did you checked cron jobs for this time? Did you tried to check system with chkrootkit?

nothing found with chkrootkit. I asked for a reinstallation of OS. maybe this fix the problem.

one question:
what ports should be open in iptables for Plesk? can I change SSH port? is it possible that iptables block something such as a restart signal and this cause the problem? it would be great if provide me an example of iptables rules with all neccessary settings for Plesk.

Thanks

 

[IgorG]

Look at this KB article. It should help.