Details
-
Bug
-
Resolution: Won't Fix
-
Blocker
-
3.0-Beta, 3.0
-
Security Level: Public
-
None
-
10.6.2.144, 10.6.2.146, 10.6.2.147, 10.6.2.150
Description
1105
1. Create 4 Node cluster with beer sample bucket
2. Failover a node (hard/graceful) 10.6.2.146
3. Add-back node using recovery type (delta)
4. Hit rebalance and during rebalance bring down a node 10.6.2.147
5. After Rebalance failover in Step 4, hard failover 10.6.2.147
Observations::
Before we execute Step 5, UI shows graceful failover for Node 10.6.2.146 for which we did delta recovery. Since one node is already down, we do not expect graceful option to be shown. When we select graceful failover and try to run it, it doesn't work.
After Step 5, we see delta recovery option for node 10.6.2.146, which is not expected. After choosing this option, rebalance fails to execute. And the rebalance will fail, the message "Could not Rebalance because requested delta recovery is not possible. You probably add more nodes then cluster had before failover or possible remove some nodes.” We should not show delta recovery as per 3.0.0 implementation expectations. Note that full recovery in this case succeeds.
Attaching logs asap