We have grid gain setup across 5 nodes and very occasionally we see a network interruption happening across the nodes for 3 - 5 secs.During this period the nodes gets disconnected. The network gets back online after that duration but since the grid nodes got disconnected they cannot communicate. Is there any way to specify the grid nodes to have a ping or heartbeat to be happening > 5 secs so that they are never disconnected due to the network interruptions ?
Right now GridGain will consider a node that is out for 5 seconds failed. I have filed an internal ticket to allow for a node to stay disconnected for a small while before it is considered failed. It should be fixed in a future release.
Having said that, it does seem rather unusual to have such break downs in your network.