This is similar to CONTROLLER-1679 but the exception happens [0] very quickly.
Karaf.log [1] shows it at 06:34:03,711 which is basically as soon as isolation happens.

As the suites are running after each other without hard reboot in between, this might be just an artifact of previous isolation cycle. The previous suite has confirmed [2] all shards now have leaders, but there might be some residual state affecting the shard behavior (module-based shards, tell-based protocol).

We will probably change the job to do more reboots, but reporting anyway as this might be a useful symptom to track.

[0] https://logs.opendaylight.org/releng/jenkins092/controller-csit-3node-clustering-only-carbon/684/archives/log.html.gz#s1-s37-t3-k2-k14-k1-k1-k1-k1-k1-k1-k1-k1-k1-k1
[1] https://logs.opendaylight.org/releng/jenkins092/controller-csit-3node-clustering-only-carbon/684/archives/odl1_karaf.log.gz
[2] https://logs.opendaylight.org/releng/jenkins092/controller-csit-3node-clustering-only-carbon/684/archives/log.html.gz#s1-s37-t2-k2-k17-k2