3 Replies Latest reply on Oct 19, 2011 6:09 AM by dan.berindei

    Intercepting when a join task completes succesfully and reloading balance on my benchmark

    renzos

      Hi guys, I created a benchmark that loads an infinispan node and start stressing the cache with an infiniste number of PUTS and GETS using k threads. The PUT/GET probability is driven by a parameter. Then I created a console manager from which I can write "newNode ip_address" in order to run nodes via SSH on the network (or to shut down those). What I would like to do is the following:

       

      1. I run a node on the ip_address1, with k threads, and once the node is up it starts stressing the cache

      2. I run a second node on the ip_address2

      3. If the join completes succesfully the second node should communicate via socket the completed join task and the console manager has to kill k/2 threads on the node1 and starts k/2 threads on the node2

      4. New nodes can be added but the total number of threads has always to be k, and k is the max nodes number for obvious reasons.

       

      It's a rebalance, useful to study a close system scalability and to analyze performaces. The problem I'm getting is that I can't manage to intercept the completed join task on the new node started in order to communicate that to my console manager via tcp sockets I created.