Hello -
When we were on Couchbase 3.1.1 (past tense), we had the following config:
3 node cluster with replicas set to 1
2 buckets in the cluster
Writes would fail under the following circumstances:
- A single node fails. This node eventually hard-fails based on the auto-failover policy
- While in this interim state, WRITES MAPPING TO THE FAILED NODE would fail
- After kicking off the re-balance (re-balancing the data onto the surviving two nodes)
- During the re-balance operation WRITES STILL FAIL
- Once the re-balance is complete, writes work as expected again
Is this still the case of CB 4.5? Will writes still fail in the period between when the node fails and when the re-balance is complete?
Thanks,
- Jeff