Re: [squid-users] almost 100% CPU (the revenge...)

From: Henrik Nordstrom <hno@dont-contact.us>
Date: Tue, 20 Apr 2004 19:46:14 +0200 (CEST)

On Tue, 20 Apr 2004, Muthukumar wrote:

> > 2004/04/20 14:36:23| peerDigestFetchReply: peer 10.0.0.5, offset: 0 size:
> > 767.
>
> > 2004/04/20 14:36:23| peerDigestFetchReply: peer 10.0.0.5, offset: 0 size:
> > 767.
>
> <snip>
>
> > cache_peer 10.0.0.5 sibling 6588 3130 default
> > cache_peer 10.0.0.1 sibling 6588 3130 default
>
> Requests are only handled by the peer of 10.0.0.5. You have specified cache_peer as default for the peers.
> It is good to use the round-robin scheduling for the requests handling when you are having more than one peers.
>
> Set store_dir_select_algorithm round-robin

Not relevant.

> > hierarchy_stoplist cgi-bin ?
> > acl QUERY urlpath_regex cgi-bin \?
> > no_cache deny QUERY
> > cache_mem 8 MB
>
> Increase the cache_mem size < = 1/2 of the Physical RAM size.

Generally a very bad thing to do. The 8 MB is generally desireable.
Exception is if you are running a HTTP accelerator infront of a small web
server with only a little content..

See Squid FAQ on memory usage.

Regards
Henrik
Received on Tue Apr 20 2004 - 11:46:16 MDT

This archive was generated by hypermail pre-2.1.9 : Fri Apr 30 2004 - 12:00:02 MDT