3 Replies Latest reply: Mar 18, 2014 7:01 AM by Simo Kemppinen RSS

    Oracle RAC/TAF FAilback once 2nd node is available again




      Ok I have a 2 node 11gR2 RAC setup.


      i have created a service to manage to failover options and its all working fine.


      I can create 2 connections, query gv$session and see they have load balanced across both nodes fine.  I then kick off a query on node 2 and then crash the node 2 instance, short pause and the select carries on.  query of gv$session shows sessions now on node 1 only.   All works as expected.


      So in the real world my app servers create a bunch of connections that sit there until required, pooled connections.  Half of which will be on each node hopefully, so when node 2 crashes all of the connections move to node 1 as they should, but now node 2 has become available again.  The connections will not move back automatically, so how do I now rebalance the load across both nodes or will i have to restart the app servers so new connections get load balanced?