Swap memory up to 100%

[ZuinqStudio]

I’ve been working with plesk panel from long time ago. Since I upgraded plesk panel from 10.x to 11.5.30 version I’m experiencing several swap memory issues.

There’re no changes in my server webpages and applications, but passing 6 hours from a server restart, swap memory goes 100% and I need to restart the server again in order to release it.

When swap memory reach 100%, memory ram used is about 33%, and the server responses dramatically slow down.

Can’t figure out what’s happening here. Is there any known issue in v11.5.30? Is there any way to downgrade back to 10.x version? Any help will be much appreciated.

Best regards.

 

[IgorG]

Plesk downgrade is impossible.
This issue require detailed investigation directly on your server. With help of top command try to determine which process is consuming the most memory at least.
I suggest you contact Support Team for deep investigation this issue.

 

[ZuinqStudio]

Hi Igorg,

Thanks for your response.

The output of top command is:

top - 09:31:18 up 18:11, 1 user, load average: 0.33, 0.28, 0.25
Tasks: 68 total, 3 running, 65 sleeping, 0 stopped, 0 zombie
Cpu(s): 11.9%us, 6.9%sy, 0.0%ni, 78.1%id, 0.0%wa, 0.0%hi, 0.0%si, 3.1%st
Mem: 1048576k total, 873344k used, 175232k free, 0k buffers
Swap: 131072k total, 131072k used, 0k free, 418508k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
972 drweb 20 0 222m 124m 92 S 0.0 12.1 0:00.52 drwebd.real
25530 drweb 20 0 222m 124m 72 S 0.0 12.1 0:00.66 drwebd.real
937 apache 20 0 370m 30m 7752 S 0.0 2.9 0:49.57 httpd
935 apache 20 0 369m 29m 7916 S 0.0 2.9 0:49.12 httpd
11231 apache 20 0 369m 29m 7100 S 0.0 2.8 0:26.79 httpd
810 mysql 20 0 676m 28m 5088 S 0.0 2.8 14:01.59 mysqld
934 apache 20 0 364m 24m 7848 S 0.0 2.4 0:51.48 httpd
11233 apache 20 0 363m 24m 7896 S 0.0 2.4 0:27.72 httpd
931 apache 20 0 365m 23m 7896 S 0.0 2.3 0:49.64 httpd
936 apache 20 0 364m 23m 7012 S 0.0 2.3 0:47.27 httpd
938 apache 20 0 364m 23m 7232 S 0.0 2.3 0:47.96 httpd
7366 apache 20 0 364m 23m 7000 S 0.0 2.3 0:34.69 httpd
933 apache 20 0 362m 22m 7776 S 0.0 2.2 0:46.58 httpd
932 apache 20 0 364m 22m 6436 S 0.0 2.2 0:50.26 httpd
9945 apache 20 0 363m 22m 7156 S 0.0 2.2 0:29.78 httpd
385 named 20 0 242m 16m 1868 S 0.0 1.6 0:03.14 named
2680 apache 20 0 189m 15m 8056 R 29.3 1.5 321:24.90 php
2085 apache 20 0 188m 14m 7948 R 3.7 1.4 0:00.17 php
928 root 20 0 345m 11m 7116 S 0.0 1.1 0:02.77 httpd
2046 postfix 20 0 77228 5704 3532 S 0.0 0.5 0:00.00 smtpd

This is the screenshot of memmory summary:



I've read that having so many httpd processes is not usual. Also, drwebd.real didn't appear in the top output other times I've experienced the issue, so I'm pretty sure it isn't the culprit. Is there anything else you find suspicious?

Regarding support team, is there any contact email? Or the only communication channel is phone?

Best regards.

 

[IgorG]

Perhaps that you are using FastCGI mode for your websites? This is the most demanding to the size of the memory PHP handler mode. Or any specific script/software?
Parallels support is here http://www.parallels.com/support

 

[Azurel]

I had a similar problem with high memory usage. But after days, not hours: http://forum.parallels.com/showthread.php?292631-Is-this-a-memory-leak
My solution was to deactivate drWeb and SpamAssassin in plesk and the problem (for me) was solved.

 

[ZuinqStudio]

Hi Azurel!

Thanks for the clue. I will try it right now!

Thanks again!