Restart is required when raft_election_timeout_ms changes

When you change the value of raft_election_timeout_ms in your cluster, you won't see cluster restart is required. "rpk cluster config status" returns false for all NEEDS-RESTART columns.

However, the new value is not applied to existing partitions unless you restart the brokers.

 

Please restart the brokers one by one to apply the new value to all existing partitions. Redpanda official doc has the instruction, https://docs.redpanda.com/current/manage/cluster-maintenance/rolling-restart/

 

Since Redpanda release >= v24.2, the behavior has changed and you no longer need to restart the cluster when you change the property value.

 

If you have any further questions, or need further guidance, please visit support.redpanda.com