For load balanced cluster that I am using, in case of router node fails, setup takes nearly 2 minutes until the VIP comes online on another running node (which earlier acted as real-server node). All requests coming in this period are not handled.
To avoid this scenario, I was thinking weather it is possible to direct future requests on the VIP be served by any other cluster node until some specific period? If yes, how can I do it?
I am using direct routing technique, that means VIP is plumbed on all the real-server nodes.
