On Wed, Nov 22, 2006, Manoj Rajkarnikar wrote: > Hi all. > > I have a question regarding squid and WCCP. I've noticed an issue with the > wccp hash bit alotment to individual cache in a farm for quite some time > and have searched alot with no luck. The problem is that in a farm of > squid cache (lets say 3 cache servers) connected to a WCCP router, if a > server gets disconnected for 10-15 minutes or more and then gets > reconnected, the hash bits alotted to that cache is changed. this means > that the urls previously served by this server would now go to another > cache and thus would take some time to rebuild the cache data. Yup, thats one of the things that needs to be fixed in the Squid WCCPv2 support. > I've searched through alot and have come to understand that its the squid > that informs the wccp router which bits are to be assigned to which squid > server. Is there a way to change the default behaviour of assigning the > hash bits to assign them in order of ip address manybe ?? You could modify the code to change the hash algorithm; or look into better 'routing' to bump requests over to the cache that would've handled it in the past. I haven't looked into it that much but its on my list of things that need fixing with the Squid WCCPv2 support. Adrian
