Hello,
I have a windows 2008 R2 cluster set up with two nodes and a virtual IP. We have one product which runs as a service on windows. It has been installed on both these nodes.
I'm trying to ensure that our product is highly available under following circumstances:
1) Node 1 fails (hardware failure): We have configured our product to listen on virtual IP of the cluster. So in this case node2 comes up and gets the virtual IP and as a result all incoming requests to our product are served by our product server running on
node2.
2) Our product service goes down on node1: This is the scenario where i'm bit confused. As per my understanding i have two options as described below:
Option 1: Use the "services and applications" feature of the Failover Cluster Manager and configure our product as a highly available service. In this case i have to ensure that service listens on a different IP. Now, in this case will my service
be highly available under both these cases:
case 1) when our service on node1 goes down. Will it bring up the service on node2 and start routing the requests to node2
case 2) when node1 itself goes down. Will it bring up the service on node2.
Option 2: Make my service cluster aware. Our service when it's running on passive node, should continuously ping and check the health of service on primary node. If for some reason we don't get a response from service on active node, then passive node should become active node.
Can you please suggest, which will be the best approach for us?
Regards,
Avadhut