Try if it helps if you use "half_closed_clients off". The support for
half closed clients might be partially broken, causing CPU spikes...
Regards
Henrik
Deac Nkisetlein wrote:
>
> Hello everybody!
>
> Iam running a squid 2.4 stable 3 happily on a solaris 8 box with
> 1 GB of Ram. Cache partition size is about 2 GB. cache_mem is
> set to 70 MB.
>
> Everything worked fine until today. I was changing cache_peer
> from one host, to another one and gave squid a reconfigure.
>
> Before that it consumed around 1-2% CPU. Now it takes
> up to 35% _WITHOUT_ (!) any clients accessing!
>
> What cachemgr is telling me, is that select loop seems to be
> called WAY to often:
>
> ------------snip
> Connection information for squid:
> Number of clients accessing cache: 5
> Number of HTTP requests received: 1928
> Number of ICP messages received: 0
> Number of ICP messages sent: 0
> Number of queued ICP replies: 0
> Request failure ratio: 0.00%
> HTTP requests per minute: 23.5
> ICP messages per minute: 0.0
> Select loop called: 46788784 times, 0.105 ms avg
> ------------snip
>
> Any help would be greatly appreciated!
>
> Thanks in advance!
>
> Regards,
> Tweeg
>
> --
> GMX - Die Kommunikationsplattform im Internet.
> http://www.gmx.net
Received on Wed May 29 2002 - 16:51:23 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:08:15 MST