Correct rolling restart procedure for a couchbase community cluster

Couchbase Server
1

Hi All, we are using the couchbase community edition and have a 3 node cluster.

We do our monthly patching and automatically reboot our servers at different time to maintain availability. Most months this works without an issue, but some months certain nodes have issues rejoin the cluster and we need to add the node back in manually with a delta recovery.

Does anyone know if the rolling restart procedure is meant to work as we have intended here or there is a safe way to rollout restart the cluster with additional steps after rebooting the server or could this be because of another issue we need to look into.

2

Hi @inor , it’s going to be tough to say for sure without looking at logs and statistics, etc.

Your procedure “should” work, but as a best practice, we would recommend using graceful failover to do that a bit more cleanly: Perform Graceful Failover | Couchbase Docs

3

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.