Re: 1.2.beta-20 memory leak

From: Alex Rousskov <rousskov@dont-contact.us>
Date: Mon, 4 May 1998 17:44:23 -0600 (MDT)

--MimeMultipartBoundary
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Tue, 5 May 1998, Chris Wedgwood wrote:

> > Note that run-time reconfigure may lead to memory leaks. They are all minor
> > (if any) to be concerned about in practice, but if you reconfigure very
> > often, the accumulated effect might be noticeable.
>
> # ps aux -p `cat /var/run/squid.pid`
> USER PID %CPU %MEM VSZ RSS TT STAT START TIME COMMAND
> squid 14150 0.8 2.7 932 658 11 S 11:16 0:00 (squid)
>
> while [ 1 ] ; do kill -1 `cat /var/run/squid.pid` ; sleep 1 ; done
>
> [left for a little while]
>
> # ps aux -p `cat /var/run/squid.pid`
> USER PID %CPU %MEM VSZ RSS TT STAT START TIME COMMAND
> squid 14150 1.3 2.8 954 680 11 S 11:16 0:08 (squid)
>
>
> Now there were no changes to the configuration here between reloads nor were
> there may accesses (maybe 1/sec or so) so this really only shows that if
> there is a leak, its not a really obvious one.

Squid may leak because not all the "configuration dependent" objects are
destroyed before they get re-created during reconfiguration. Thus, one does
not need to change squid.conf or submit requests to monitor this kind of
behavior. In practice, again, reconfiguration is not likely to happen often..

Alex.

--MimeMultipartBoundary--
Received on Tue Jul 29 2003 - 13:15:49 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:11:46 MST