Hi Adi, Why do you want failover? It's just a cache, so your application should be able to run ok if part of the cache cluster is unavailable, you would experience a slightly higher cache miss ratio.
If your application requires that all memcached instances are up and running all the time, it's pretty likely that you are doing something wrong, that you are using memcached in a way it was not intended for. Please tell us a bit about your application and we can probably help you. That said, the BeITMemcached client supports consistent hashing, but there is no automatic failover, and no automatic recovery from failover, because those features would in the absolute majority of cases only cause subtle errors in the application, without offering any benefits whatsoever. /Henrik On Fri, Oct 16, 2009 at 13:55, Adi <adeelna...@gmail.com> wrote: > > Hi, > I am using memcached on windows server 2003 in a web cluster > environment setup through NLB (Network load balancer) and using two > different memcached server for caching, using BeIT Memcached client. > > I want to know if memcached doesn't provide failover explicitly than > is there any way to handle it? > > I had also check out the faqs, let me know please which windows client > provide consistent hashing if one memcached node is dead? Can i handle > it manually buy identifying the dead node and remove it from the > server list. > > Regards > Adeel Ahmed
