RAC High Availability Question When Failing Public Network
I have a 2 node cluster running 9.2.0.6 RAC with HP Service Guard. When I fail the public network on node 1, any active sessions connected to instance 1 freeze. They stay frozen until I re-enable the public network, which makes sense. Now when I test a complete node failure by cutting power to the node, then these active sessions connected to instance 1 will freeze forever. It seems sqlplus will wait for the results to be returned from the database forever. Are there any steps that can be taken to fix this issue? Because session failover will not occur unless sqlplus gets a "no longer connected" or "end of communication failure"? Thanks!