When I use infinispan 4.0.0, if one or more nodes failed to response are-you-alive request, the nodes will be dropped out as one or more partitions.
Is there anybody knows what the strategy is to merge them back, that means if I use "replication mode" the new merged group will pick which partition's cache to make all nodes in this group same. And further if I use "distributed mode" what's the strategy?
