Health monitor - Memory usage increase slowly but continuously

[FranckL]

Hello,

I searched on every thread on this forum but nothing helps me.

I upgraded plesk from 10.4.4 to 11.0.9, I had differents problems that are mentionned in this forum (programmed backup task not working, health monitor displaying nothing etc.).

Everything seems to work now but I have memory problems.

1°) On server reboot, health monitor have strange behaviours like in this thread : http://forum.parallels.com/showthread.php?t=261169
It shows me "13865731,3 %" usage of CPU. (see attached picture) After 10 minutes, it shows a correct value.

2°) From the reboot and continuously, the memory usage increase...
About 8 days after sever working, Plesk, my ssh and my server is down because of memory saturation (16Go), I have to make a hard reboot of the server.

3°) I made everything in this thread : http://forum.parallels.com/showthread.php?t=260744
My Interval in "/etc/sw-collectd/collectd.conf" is 300.

4°) Real Memory Values in health monitor interface not reflecting the ssh "top" command values and system statistics values.
In health monitor : 10.2% used (1.59 Go from 15.6 Go)
In system stats and "top" command : 11.32% used (2.28 Go form 15.58 Go)

5°) I had the same problem (memory increase) on the 10.4.4 plesk with Grsec kernel, I used the kernel without Grsec and now it still working correctly on 10.4.4. Is Grsec included on CentOS 6.2 (Final) ? if yes, is it possible to disable ?

Yesterday I rebooted the server and now my memory still increasing. (see attached picture). It will increase until my server is down and I have to reboot.

Thanks for your help.

Additional info :

CPU : GenuineIntel, Intel(R)Core(TM) i5-2400 CPU @ 3.10GHz
Version : Parallels Plesk Panel v11.0.9_build110120608.16 os_CentOS 6
OS : CentOS 6.2 (Final)

 

[FranckL]

Have you got news about my problem

Sorry, I really need to fix my memory problem, my server was inaccessible this afternoon, I had no choice and I made a hard reboot. The memory still increase continuously...

I have no choice, I have to reinstall my server and still on 10.4.4 version, which is more stable !

But I'll be afraid to upgrade to 11.0.9 !! How can I do ?

I can't reboot my server every 3 days ! Please help me.

 

[FranckL]

Nobody can help me ?

PleskTeam, have you got an answer, any reflection? my message is wrong, did I miss something? I did something wrong ?

I rebooted 5 times since my first message...because of memory saturation.

My problem is serious and I have 5 servers with plesk installed.
I'll never upgrade to 11.0.9 because of memory problems ? And I have to reinstall this server with 10.4.4 version for resolving the problem. Is-it the solution ?

Could you help me please, and if I made something wrong, tell me, and tell me if you need more information.
Thank you

 

[IgorG]

I think that there are can be different reasons of this strange behaviour. We can't reproduce it on our test environment. Therefore I suggest you contact Support Team. They will check and deeply investigate this issue directly on your server.

 

[FranckL]

Thank you for this information, I'll try to contact them.

But I can give you some more information, perhaps somebody has the same problem.

To sum up the tread:

After upgrading from 10.4.4 to 11.0.9, my memory grows up continuously every 5 minutes (the same interval as /etc/sw-collectd/collectd.conf Interval 300)

I tried to compare the "Top and (shift+M)" command to compare values with an other server running on 10.4.4 but nothing is really different...

I tried to stop sw-collectd with #/etc/init.d/sw-collectd stop But no changes...

Have I some tools to know what task is using memory every 5 minutes ? it increase the memory and it doesn't free it after finishing the task.

It seems to be a recurrent task every 5 minutes...