Search before asking
Motivation
In #1142. We will introduce a new request of controlled shutdown, which will inform the coordinatorServer to migrate the bucket leader on this server through retries before the tabletServer goes offline. However, we need a mechanism to ensure that if an offline tabletServer sends a controlled shutdown request (abnormal situation), the coordinatorServer should ignore this request; otherwise, it would incorrectly take the leader on the new tabletServer offline.
Solution
No response
Anything else?
No response
Willingness to contribute