[11.25 Current] apache restarting every hour

[trhosting.net]

I think the best solution for this is reverting back to the old system, while working on a fix for this issue.
I don't understand why cPanel didn't make a temporary solution like this.

Lots of people having problem when cPanel is working on this issue and there is no ETA for fixing this issue.

[InterServed]

I don't know what is changed but after upgrading to 42400 (64bit), apache started to restarting periodically and it became unreacheable 3-4 times a day.

Before upgrade we don't have any problems after setting log schedules and disabling log rotation.
After upgrade, no settings are changed but we are having problems.

I attached daily and weekly mrtg graphs, you can see the problem there.
We upgraded on saturday.

As i said i don't know it is related with upgrade but problems started after upgrade.

EDIT: both statistic schedules are set to 24 hours and log rotations (all) are disabled.

Web Traffic Statistics every 24 hours
Bandwidth Statistics every 24 hours

Current Time: Monday, 18-Jan-2010 21:22:07 EET
Restart Time: Monday, 18-Jan-2010 17:41:44 EET
Parent Server Generation: 13
Server uptime: 3 hours 40 minutes 23 seconds
Total accesses: 657141 - Total Traffic: 19.8 GB
CPU Usage: u32.01 s32.02 cu651.23 cs0 - 5.41% CPU load
49.7 requests/sec - 1.5 MB/second - 31.5 kB/request
192 requests currently being processed, 82 idle workers
WKW__RWKCWCWWK_CW_KC_K_WK__WKWKWWKW__.WW__WWK__W_KCKRK_WKK.WWKK_
WW__.W._WC_W_.KW.K.K.WK.._K_.W.C.W..W_RW.K_WKKKWWCCW._WW._._.._.
_.WW.WK....W.C_.._._CW_KW._..WK_WC.W_.._K__._KW.W_.K._K..C.W_KWW
.KKW_C...K_._WWKKW.._K...W.K_W..CKW_.K...CW...._....W._K_.._..._
W.._.CWW.W..WW..K..WK.KWK_.C_._.W_.___.KW__..W...W.._W._W.....K.
.WW___....._..WW..WW.__WW_.K.W.W.C_.R.W..W_.K.KK.WWWW_WWWW.W._.W
.WW__KCCW.WCWK.WKWW_WWW.....G...................................
................................................................

I haven't experienced any of the described issues:
Server Version: Apache/2.2.14 (Unix) mod_ssl/2.2.14 OpenSSL/0.9.8e-fips-rhel5 DAV/2 mod_auth_passthrough/2.1 mod_bwlimited/1.4 FrontPage/5.0.2.2635
Server Built: Dec 28 2009 13:02:21
Current Time: Tuesday, 19-Jan-2010 16:37:01 EET
Restart Time: Saturday, 16-Jan-2010 21:20:50 EET
Parent Server Generation: 26
Server uptime: 2 days 19 hours 16 minutes 10 seconds
Total accesses: 2510241 - Total Traffic: 58.9 GB

cPanel 11.25.0-C42404

[pjman]

I have a high traffic site that throttles a high output from Monday thru Friday. Sat and Sun it's light on traffic. On Monday thru Friday I have the problem. This weekend, no problems at all.

InterServed looks like you have a good bit data being transferred. But, do you have a single account that has tons of traffic?

I'm thinking Cpanel put something in Release build that stops individual sites from being resource hogs?

[cPanelDon]

I'm thinking Cpanel put something in Release build that stops individual sites from being resource hogs?

Any such behavior would be specific to the individual system configurations and applicable software customizations and is not inherently performed by cPanel/WHM.

We too have some bad high load issues since the move to 11.25 release.
Does anyone know if rolling back to 11.24 STABLE resolves the issues?

Inquiries for investigating load and unique server resource usage issues are best suited for the following forums area: cPanel and WHM Optimization - cPanel Forums

I will move your post momentarily and reply in the new thread to assist further; for reference, here is the new thread: High load issues - cPanel Forums

Moderator Note: Please remember to keep the discussion on-topic; this thread is about specific difficulty with Apache graceful restarts triggered by log processing while using cPanel 11.25.0.

[pjman]

Don, I completely understand that this should have been discussed in: cPanel and WHM Optimization - cPanel Forums

I placed a thread in there last week when it started. No one looks in that forum area at all since you guys switched up the forum. As a result, I got no feedback at all. I just got referred to this thread.

Since I posted here, I got a great amount of feedback. I now know many people are having the same issue since 11.25 R went Build 42400. Makes me feel better that this not an issue related to my own unique situation.

[skratchit]

This problem has been bugging me (and my users) for a few days. I think I've managed to fix it on my server.

DISCLAIMER: This is speculation.

It seems to be related to the logs in "/usr/local/apache/logs"

By default I believe these are set to be rotated when they get to 300 MB (if checked in "Service Configuration >> Apache Configuration >> Log Rotation" in WHM).

I think their file sizes are checked every 15 mins.

What seems to be happening is that over 300 MB they get copied / appended and Gzipped to "/usr/local/apache/logs/archive/" and Apache is gracefully restarted. But the original file is not emptied.

So, 15 minutes later the log is rotated _again_ because it is still over 300 MB. And so the cycle continues.

I've just renamed my logs that were over 300 MB and created empty ones and restarted Apache and it has successfully not restarted for 40 minutes (when it was restarting every 15-17 mins).

[racakg]

This is exactly the same problem i had one month ago with su_exec.log. Disabling all rotation solved my problem with restarts.

http://forums.cpanel.net/f5/11-25-cu...tml#post607253

[InterServed]

InterServed looks like you have a good bit data being transferred. But, do you have a single account that has tons of traffic?

This isn't the case , i have a lot of accounts ( including resellers and each having 50+ accounts ).

[cpanelkenneth]

This problem has been bugging me (and my users) for a few days. I think I've managed to fix it on my server.

DISCLAIMER: This is speculation.

It seems to be related to the logs in "/usr/local/apache/logs"

By default I believe these are set to be rotated when they get to 300 MB (if checked in "Service Configuration >> Apache Configuration >> Log Rotation" in WHM).

I think their file sizes are checked every 15 mins.

What seems to be happening is that over 300 MB they get copied / appended and Gzipped to "/usr/local/apache/logs/archive/" and Apache is gracefully restarted. But the original file is not emptied.

So, 15 minutes later the log is rotated _again_ because it is still over 300 MB. And so the cycle continues.

I've just renamed my logs that were over 300 MB and created empty ones and restarted Apache and it has successfully not restarted for 40 minutes (when it was restarting every 15-17 mins).

Aside from the restart, the above has been the behavior since the cPanel 10 era ( maybe earlier ).

We've changed this behavior in our development builds, and the change will be in the next builds of 11.25.0.

You are correct that the rotation threshold is 300 MB. Unfortunately the way the rotation logic works in all production builds is:

1. Does the log exceed the threshold ( even by one byte )?

2. If yes then shrink the log to the threshold

3. If no then move on

We changed the logic to:

1. Does the log exceed the threshold ( even by one byte )?

2. If yes then shrink to 0[*]

3. If no then move on

Proper archiving/rotation is retained when the file is truncated to 0-length.
[*] This only applies to the service logs (/usr/local/apache/logs). For the domain logs ( /usr/local/apache/domlogs) the threshold is 1 GB. When log rotation is requested for these logs the logic change will reduce the file to 500 MB. We will readdress this particular logic later as there are other parts of the product that need changed to work properly with a 0-length domain log.

We are also exposing the service log rotation threshold in the WHM, rather than requiring you to modify the cPanel configuration file manually.

[cpanelkenneth]

This thread has served its purpose and is now closed. Changes to Apache logging are forthcoming and detailed at http://forums.cpanel.net/f5/cpanel-1...ng-145417.html

Please direct all constructive comments to that thread.

Thank you.