httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jean-Jacques Clar" <JJC...@novell.com>
Subject Re: Seg fault: Possible race conditions in mod_mem_cache.c
Date Wed, 08 Sep 2004 16:13:40 GMT
I am using a set of files I got from VeriTest
http://www.etestinglabs.com/benchmarks/webbench/default.asp
The tar file include about 6000 files of a total size around 60 MB.
Has you can see in the configuration MCacheSize and 
MCacheMaxObjectCount force cached objects to be ejected 
often. This is the key to reproduce that problem. 
The numbers have to be smaller than your sample. I am sure
the problem could be reproduced with a smaller sample.
For the worker mpm, I am just using the default configuration. 
I am clueless as far as debugging that case on a Linux platform,
but I know that protecting decrement_refcount() in mod_mem_cache.c
with a lock prevents segfaults.
 
LoadModule cache_module modules/mod_cache.so
<IfModule mod_cache.c>
   CacheMaxExpire 600
   CacheDefaultExpire 12000000
   CacheIgnoreCacheControl On
   LoadModule mem_cache_module modules/mod_mem_cache.so
   <IfModule mod_mem_cache.c>
      CacheEnable mem /
      MCacheSize 7000
      MCacheMaxObjectCount 3001 
      MCacheMinObjectSize 1
      MCacheMaxObjectSize 69000
   </IfModule>
</IfModule>
 


>>> trawick@gmail.com 09/08/04 9:50 AM >>>

>Sorry but when running the same test on SLES9 with 2.0.51rc1, 
>my error_log is full of threads segfaulting. 

That is terrible.  Do you have a simple testcase for cache-ignorant
folks like myself to play with?


Mime
View raw message