From time to time on our pre-prod environment, one or two of our 3-node cluster go down (they become not reachable - root cause has not been identified yet). Then all new transactions are processed on the “surviving” node.
For instance, let’s suppose cluster node N1 and node N2 fail respectively at time T1 and T2 (while the Node N3 stay alive and receive new data updates)
We wonder if data sync works well when the cluster nodes N1 and N2 (which were “down”) are relaunched : will the nodes N1 and N2 receive again all data changes which happened on node N3 respectively since T1 and T2?
Also, do you know possible reasons why a cluster-node can go down?
Thanks!