Re: [squid-users] Help peering httpd accelerators

From: Mohamed Lrhazi <mohamed@dont-contact.us>
Date: Fri, 10 Dec 2004 22:09:21 -0500

--On Saturday, December 11, 2004 12:27 AM +0100 Henrik Nordstrom
<hno@squid-cache.org> wrote:

> On Fri, 10 Dec 2004, Mohamed Lrhazi wrote:
>
>> What I notice happening is that the squids do send ICP queries to each
>> other but they allways end up fetching the objects from their own
>> cache, or the origin server. never from another cache. I am trying to
>> get multicast to work, but I see the eaxt same behavior with unicst icp
>> config.
>
> What does cachemgr say about the peering statistics?
>

Basically all counters are zero.

>> cache_peer 224.9.9.9 multicast 8080 3130 ttl=1
>> cache_peer 10.0.20.27 sibling 8080 3130 proxy-only multicast-responder
>> cache_peer 10.0.20.28 sibling 8080 3130 proxy-only multicast-responder
>> cache_peer 10.0.20.29 sibling 8080 3130 proxy-only multicast-responder
>> mcast_groups 224.9.9.9
>
> Not sure if multicast works very well. Has been broken from time to
> time.. (too few people using this, probably because it only provides
> marginal benefit compared to unicast ICP)

Thanks. I moved to unicast ICP... and it does seem to work now, I changed
too many things of course, can't tell what solved it. I do notice something
weird though, but I will post it separately.

Mohamed~
Received on Fri Dec 10 2004 - 20:09:23 MST

This archive was generated by hypermail pre-2.1.9 : Sat Jan 01 2005 - 12:00:02 MST