Start of change

6027-4211 [I] CCR leader election: lease overdue; must declare quorum loss before taking over.

Explanation

The current cluster leader has failed and the local node has successfully challenged the old leader and been elected as the new leader by using tiebreaker disks. However, too much time has passed since the challenger's last successful lease renewal. Because of that, it is possible that the old leader has expelled the challenger and started log recovery (or will start log recovery) while the old leader still has a valid lease. As a result, data or metadata that the challenger has cached may become stale. To prevent file system corruption, the challenger must declare quorum loss to reset its state before it takes over.

User response

None.
End of change