plainblack.com
Username Password
search
Bookmark and Share

    

Apache2::SizeLimit httpd process too big

User zeros
Date 1/3/2006 4:59 am
Views 5844
Rating 16    Rate [
|
]
Previous · Next
User Message
zeros

Hello Everybody !

anyone any idea about this message in my apache error_log.

[Tue Jan  3 11:20:33 2006] (775) Apache2::SizeLimit httpd process too big, exiting at SIZE=53124 KB  SHARE=3256 KB  REQUESTS=6 LIFETIME=0 seconds

I´m using WebGUI 6.7.8 in a seperate apache instance (with modproxy).

The Site has nearly no traffic.

In my httpd.conf I have following PerlRequire

use Apache2::SizeLimit;
$Apache2::SizeLimit::MAX_PROCESS_SIZE = 75000;
$Apache2::SizeLimit::MAX_UNSHARED_SIZE = 25000;
$Apache2::SizeLimit::CHECK_EVERY_N_REQUESTS = 5;

Thanks !

Luke 

 

--- (Edited on 1/ 3/2006 4:59 am [GMT-0600] by zeros) ---



Back to Top
Rate [
|
]
 
 
lunap2

Hi,

I recently installed wre-0.7.2 with webgui-7.4.5-beta, on Suse Linux 10.  I am constantly getting the errors:

  [Fri Sep 21 10:15:56 2007] (22992) Apache2::SizeLimit httpd process too big, exiting at SIZE=175500 KB  SHARE=3764 KB  REQUESTS=6 LIFETIME=0 seconds
[Fri Sep 21 10:18:01 2007] (22996) Apache2::SizeLimit httpd process too big, exiting at SIZE=175880 KB  SHARE=3656 KB  REQUESTS=6 LIFETIME=0 seconds
[Fri Sep 21 10:21:01 2007] (22999) Apache2::SizeLimit httpd process too big, exiting at SIZE=176196 KB  SHARE=3776 KB  REQUESTS=6 LIFETIME=0 seconds

The site has no traffic except for me.  The configuration of the modules are as defined by the wre.  From the dialog, I understand this is not bad in itself.  I looked at the Apache2::SizeLimit docs as described in the response but am not clear whether I need to make any changes.  Any comments will be appreciated.  Thanks.

 

--- (Edited on 9/21/2007 9:55 am [GMT-0500] by lunap2) ---



Back to Top
Rate [
|
]
 
 
jameslyne

I had this problem... it is worth noting that SizeLimit triggering it itself is no bad thing - but if it is happening on a very low traffic site... it's worrying.  Are you by any chance experiencing high CPU usage by depracated httpd processes?  I found that when using RSS or syndicated content feeds in my site I was able to trigger this precise behaviour (it gets worse over time... as the RSS attempts to reload).  If I remove all the syndicated content - it works beautifully....

 Similar?

--- (Edited on 1/ 3/2006 7:09 am [GMT-0600] by jameslyne) ---



Back to Top
Rate [
|
]
 
 
knowmad

 

Are you by any chance experiencing high CPU usage by depracated httpd processes?  I found that when using RSS or syndicated content feeds in my site I was able to trigger this precise behaviour (it gets worse over time... as the RSS attempts to reload).  If I remove all the syndicated content - it works beautifully....

 

I've recently started using RSS feeds from collab. system assets and am experiencing this type of behavior. My modperl error logs show SizeLimit disabling children that have only been running a couple of minutes and are less than 100Mb. In addition, I'm seeing errors from the rss feed asset and have even experienced the server crashing.

My load average is miniscule (currently at 0.04) and I have 1.5Mb free according to top. I've posted a message to the PB support boards and will report back if I get an answer.

 

William 

----
Knowmad Technologies
http://www.knowmad.com

--- (Edited on 10/25/2007 12:55 pm [GMT-0400] by knowmad) ---



Back to Top
Rate [
|
]
 
 
lunap2
I've recently started using RSS feeds from collab. system assets and am experiencing this type of behavior. My modperl error logs show SizeLimit disabling children that have only been running a couple of minutes and are less than 100Mb. In addition, I'm seeing errors from the rss feed asset and have even experienced the server crashing.

My load average is miniscule (currently at 0.04) and I have 1.5Mb free according to top. I've posted a message to the PB support boards and will report back if I get an answer.

 

Thank you.  I am not using RSS feeds.  I'm considering installing the wre 0.8 and see how that behaves.  I'll post back once I get to it.

--- (Edited on 10/25/2007 2:56 pm [GMT-0500] by lunap2) ---



Back to Top
Rate [
|
]
 
 
knowmad

To modify the size limit configurations:

  • WRE 0.7.2 - /data/wre/prereqs/apache/conf/startup.pl
  • WRE 0.8 - /data/wre/etc/modperl.pl

 

--- (Edited on 10/25/2007 4:26 pm [GMT-0400] by knowmad) ---



Back to Top
Rate [
|
]
 
 
zeros

 

thank you, but I´m not using Syndicated Content/RSS. Just Articles.

And there are no depracated cpu processes...

Strange... 

--- (Edited on 1/ 5/2006 11:48 am [GMT-0600] by zeros) ---

 

--- (Edited on 1/ 5/2006 11:48 am [GMT-0600] by zeros) ---



Back to Top
Rate [
|
]
 
 
len

This message in your error_log is not an error, but a notification. 

Apache2::SizeLimit is stopping that httpd child, because it's using too few shared memory (55124 - 3256 = 49868 UNSHARED). 

It's actually a good thing that Apache2::SizeLimit kills off that child, because it's using too much physical RAM.

The child is stopped gracefully after it has served the request, so user won't notice it. A new child is spawned if needed.

Read the Apache2::SizeLimit documentation for more info.

Len Kranendonk, www.primaat.com

---- Are you a chess player ? Challenge me for a game ---

--- (Edited on 1/ 6/2006 7:13 am [GMT-0600] by len) ---



Back to Top
Rate [
|
]
 
 
timothy

As stated in the documentation pointed to by Len:

Since linux 2.6 /proc/self/statm does not report the amount of memory shared by the copy-on-write mechanism as shared memory. Hence decisions made on the basis of MAX_UNSHARED_SIZE or MIN_SHARE_SIZE are inherently wrong.


if you are running a recent 2.6 kernel you can use smaps

check /proc/self/ if there is a map called smaps

if it is there you just have to install the

Linux::Smaps perl module


restart webgui and the problem is solved.Tongue out

--- (Edited on 4/14/2009 8:10 am [GMT-0500] by timothy) ---



Back to Top
Rate [
|
]
 
 
© 2014 Plain Black Corporation | All Rights Reserved