[squid-users] tracking down leaks

From: josh <josh@dont-contact.us>
Date: Mon, 8 Dec 2003 22:41:30 -0500

How do I track down a memory leak? How do I fix one if I find it?

I am running squid compile with the options listed below on a Sun
Cobalt 550. The system seems to swallow up memory as it runs. I don't
know for sure that squid is leaking, but it is a likely suspect. If
squid is leaking it has to be doing it in a subtle way where the
memory is taken, does not really show up as part of the squid process,
and is not returned when squid stops. It is a 2.4 linux redhat
kernel. I have to use RAID formatted disk for the cache.

I had tried to have much larger cache_mem which seemed to be
worse. Though that it seemed to be worse was all that makes me thing
squid might be the problem. I have 2G of ram.

cache_access_log /var/log/squid/access.log
cache_dir ufs /home/squid/cache 7000 64 256
cache_effective_user nobody
cache_log /var/log/squid/cache.log
cache_mem 32 MB
cache_mgr admin
 

Squid Cache: Version 2.5.STABLE4
configure options: --prefix=/usr/local --enable-gnuregex
--enable-carp --enable-delay-pools --enable-ssl --enable-useragent-log
--enable-referer-log -enable-forw-via-db
--enable-default-err-language=English --enable-err-languages=English
--enable-underscores '--enable-auth=basic digest'
'--enable-basic-auth-helpers=getpwnam NCSA SASL LDAP PAM'

-- 
Josh Kuperman                       
josh@saratoga.lib.ny.us
Received on Mon Dec 08 2003 - 20:42:20 MST

This archive was generated by hypermail pre-2.1.9 : Thu Jan 01 2004 - 12:00:07 MST