[CONTROLLER-1953] Cluster ends up with a leader=null state when running a daexim export on non-leader

Fri, 2 Jul 2021 12:55:17 +0000

As demonstrated,
1. Brought down the ODL6(non-voting node)
2. Change the default-config leader
3. Ran down-unreachable-nodes
4. Started the ODL6 and monitored the leaderId=null messages and ODL6 managed to come up clean.

On leader node ODL1 for default-config

2020-06-22T21:33:03,046 | INFO | opendaylight-cluster-data-shard-dispatcher-41 | Shard | 290 - org.opendaylight.controller.sal-clustering-commons - 1.9.1 | member-1-shard-inventory-config (Follower): Follower is out-of-sync so sending negative reply: AppendEntriesReply [term=15, success=false, followerId=member-1-shard-inventory-config, logLastIndex=6, logLastTerm=7, forceInstallSnapshot=false, needsLeaderAddress=false, payloadVersion=9, raftVersion=4, recipientRaftVersion=3]

On non-voting ODL6 node,

2020-06-22T21:33:17,862 | INFO | opendaylight-cluster-data-shard-dispatcher-37 | ShardManager | 298 - org.opendaylight.controller.sal-distributed-datastore - 1.9.1 | shard-manager-config: Received LeaderStateChanged message: LeaderStateChanged [memberId=member-6-shard-topology-config, leaderId=null, leaderPayloadVersion=9]
2020-06-22T21:33:17,872 | INFO | opendaylight-cluster-data-shard-dispatcher-37 | ShardManager | 298 - org.opendaylight.controller.sal-distributed-datastore - 1.9.1 | shard-manager-operational: Received LeaderStateChanged message: LeaderStateChanged [memberId=member-6-shard-entity-ownership-operational, leaderId=null, leaderPayloadVersion=9]

OpenDaylight JIRA

[CONTROLLER-1953] Cluster ends up with a leader=null state when running a daexim export on non-leader