Details
-
Bug
-
Resolution: Not a Bug
-
Critical
-
Cheshire-Cat
-
Centos 7 64 bit; CB EE 7.0.0-4721
-
1
Description
Summary:
Nodes are not completely removed from the cluster after quorum failover
Steps to Reproduce:
1. 2 node init cluster - .215, .217
2. Disable autofailover
3. Load travel-sample bucket
4. Induce failure "stop-server" on .217 node
5. (Quorum) Failover .217 node
So after confirming failover, node .217 is failovered and removed from the cluster from the point of view of .215 node. This can be confirmed by UI logs on .215 node. All is good until this point.
Now, I start the server again on .217 node and try to add it back to the single node cluster, it fails with "Node already part of the cluster". Checking the UI from .217 node, it seems like .217 node is indeed still a part of cluster and was not truly removed from the cluster when it was (quorum) failovered.
UI logs screenshots attached