How to identify process owner / control resource intensive process

[rebouy]

Hi, the server's load has increase this few last days.

When i go to process manager, this entry is always the first one :
16827 (Trace) (Kill) 32006 0
96.2 0.3 /var/cpanel/3rdparty/bin/php -c /usr/local/cpanel/3rdparty/etc/roundcube /usr/local/cpanel/base/3rdparty/roundcube/index.php

This process by owner 32006 is using 96.2% cpu.
How can i really identify whats going on here?

We have other servers besides the one with the problem, and the others doesnt have this issue.

TNX!

[cPanelDon]

Hi, the server's load has increase this few last days.

When i go to process manager, this entry is always the first one :
16827 (Trace) (Kill) 32006 0
96.2 0.3 /var/cpanel/3rdparty/bin/php -c /usr/local/cpanel/3rdparty/etc/roundcube /usr/local/cpanel/base/3rdparty/roundcube/index.php

This process by owner 32006 is using 96.2% cpu.
How can i really identify whats going on here?

We have other servers besides the one with the problem, and the others doesnt have this issue.

TNX!

Roundcube processes should be relatively short-lived and not run for several hours. Do you see the same process ID displayed for several hours, or does the process ID change upon subsequent status checks when viewing running processes at a later time? For reference, the process ID (PID) is displayed on the left at the start or beginning of the line.

If you would like to see which users are accessing Roundcube, you may use the following command via root SSH access to search the cPanel access_log:

# egrep "GET (/cpsess[0-9]+)?/3rdparty/roundcube/\?.* HTTP/1.[01]" /usr/local/cpanel/logs/access_log

Are any "webmaild" processes running? The following commands should list any that may be running along with the IP address of the user connected:

# pgrep -l -f webmaild

Please let us know your cPanel&WHM version and Roundcube version; these can be obtained using the following command via root SSH access:

# grep -H '' /usr/local/cpanel/version /var/cpanel/roundcube/version

[rebouy]

It was running for several hours. i just checked and is the same pid as my first post. 16827.

Yes, there are several webmaild processes running

/usr/local/cpanel/version:11.28.64-RELEASE_51024
/var/cpanel/roundcube/version:0.4.2.cp3

This only happens in 1 of our servers, the rest is fine. Also it started a just few days ago.

According to the process log, it started on 12-23-2010 and since this day this process is responsible for 10% of the daily cpu load.

This is a closed server, we havent add any new accounts to it in quite a long time. Normal avg load was around 1.5 - 2.0 at peek hours, now its 2.0 - 4.0 at any give time.

[rebouy]

Is it ok to just kill the process and restart apache?

How can we prevent same thing to keep happenin?

[cPanelDon]

It was running for several hours. i just checked and is the same pid as my first post. 16827.

Yes, there are several webmaild processes running

/usr/local/cpanel/version:11.28.64-RELEASE_51024
/var/cpanel/roundcube/version:0.4.2.cp3

This only happens in 1 of our servers, the rest is fine. Also it started a just few days ago.

According to the process log, it started on 12-23-2010 and since this day this process is responsible for 10% of the daily cpu load.

This is a closed server, we havent add any new accounts to it in quite a long time. Normal avg load was around 1.5 - 2.0 at peek hours, now its 2.0 - 4.0 at any give time.

Is it ok to just kill the process and restart apache?

How can we prevent same thing to keep happenin?

It is OK to kill the Roundcube process(es); however, please note that restarting Apache is not necessary as Roundcube is not being served from Apache/httpd.

Was it just one (1) Roundcube process that was running for several days, or are there multiple Roundcube processes that run for several days? If, after killing the Roundcube process(es), the problem recurs again then I would consider submitting a support request.