Hi Jens, Thanks for the explanation, its really helpful in understanding how its happening. If i start the server 2 prior to server 1 then server 1 becomes inefficient, thus it agrees with normal behaviour as you explained above. Even if i allow clients to use server pool instead of locator pool i.e. client1 to server1 and client2 to server2, same behaviour is observed. This validates your point ' From the client, this also means that connections to the secondary are in use longer...' . It seems process is busy in some kind of waiting like distributing operation etc.. I think setting property 'thread-local-connections' to true will not work for this case because 1. Here thousands of threads use client connections and do a single put operation 2. This will not improve server response time which is main cause behind why connections are in long use. I would definately like to jump to Gemfire 7 but problem is that we are using evaluation version and this Async event queue feature is not licensed in evaluation version. Please suggest if there is a way to evaluate this feature in Gemfire 7 ? Thanks Mahesh
↧
Re: Impact of Gateway Hub on cache server cluster performance
↧