We have two clusters, each cluster consists of 2 servers, and each server has 8 nodes。
Then we kill one server from each cluster,we find one cluster can lead one node failover,
cluster 1 :msg":"safe mode, loop ratio 0.40, allnodes: 16, failnodes: 7
cluster 2:msg":"safe mode, loop ratio 0.40, allnodes: 16, failnodes: 8
and in cluster 1,when we startup broken server‘s node , we find one shard have two master,and two master have different version。
We have two clusters, each cluster consists of 2 servers, and each server has 8 nodes。 Then we kill one server from each cluster,we find one cluster can lead one node failover, cluster 1 :msg":"safe mode, loop ratio 0.40, allnodes: 16, failnodes: 7 cluster 2:msg":"safe mode, loop ratio 0.40, allnodes: 16, failnodes: 8
and in cluster 1,when we startup broken server‘s node , we find one shard have two master,and two master have different version。