Are you using any sotreurl rewriter ? if you do , then rewritten URLs will not be found by ICP On Mon, May 16, 2011 at 8:09 PM, George N. Zavalis <gzabalis@xxxxxxxxx> wrote: > > Hi list, > > I have squid Version 2.7.STABLE9 and I tried to setup a multi instance squid > system as proposed at http://wiki.squid-cache.org/MultipleInstances > > I used the "http_port" setting to create two (2) independent squid instances > running at the same ip address but on different ports : Squid no 1 > listening at port 81 and Squid no 2 listening at port 82 and icp ports at > 3131 & 3132 respectively > > Also I configure the "cache_peer" setting so as to have sibling between the > 2 squid instances : > cache_peer ip sibling 82 3132 no-netdb-exchange no-delay no-digest > proxy-only , at the no 1 squid and > cache_peer ip sibling 81 3131 no-netdb-exchange no-delay no-digest > proxy-only , at the no 2 squid. > > This is where my problems are occurring. > When the first request arrives at squid no 1 - of course > TCP_MISS:FIRST_UP_PARENT - I can see from the access log : GET > http://FQDN:81/resourcepath > I was expecting when I request the same resource and squid no 2 reply to me, > I sould get a cache hit but instead I got a TCP_MISS: FIRST_UP_PARENT and of > course from the access log : GET http://FQDN:82/resourcepath > > Subsequent requests to either http://FQDN:81/resourcepath or > http://FQDN:82/resourcepath are HITS > > I am running two (2) different proxies - one with the prefix ":81" and the > other with the prefix ":82" - without any sibling hit. Looks like I have 2 > isolated squids which is not what I want. > > Why is this happening? Is this happening because of the existence of the > port at the URL? Can I configure squid to ignore it > > Thanks > GnZ >
