[rt-users] httpd processes consuming a lot of memory

[Fran Fabrizio]

May be a false alarm, I think I traced this to a 3.2GB log file that I 
was using that was logging everything at 'warning' and above.  Of 
course, that begs the question "why did I have 3.2GB worth of warnings", 
but once I disabled this logging and went back to syslog-style logging, 
the system became MUCH snappier and mem usage is down.

Fran Fabrizio wrote:
> Within the last week, the httpd processes on the RT server have started 
> consuming much more memory than usual.  On a fresh restart, after I hit 
> the Home page once, my httpd processes are up to 110-130m of memory per 
> process.  This is a lot more than usual, and is causing our server 
> (which has been running happily as a 512MB VM for years) to really 
> thrash its swap.  Could this be the result of a bad ticket (poorly 
> behaving attachment or something?)  MySQL database issue?  How best to 
> hunt the cause?  I've been tuning down the various httpd.conf parameters 
> for number of children, lifespan of children, etc... but as I said, the 
> problem manifests on the very first hit after a restart, so I don't 
> think that's it.  Thoughts?
>
>   


-- 
Fran Fabrizio
Senior Systems Analyst
Department of Computer and Information Sciences
University of Alabama at Birmingham
http://www.cis.uab.edu/
205.934.0653