This topic is related to https://www.mongodb.com/docs/v6.0/tutorial/mitigate-psa-performance-issues/ in a PSA ReplicaSet configuration.

When you try to execute a command with writeConcern {w: "majority"} in a three-member Primary-Secondary-Arbiter configuration where one data bearing node is not available, then the command hangs forever - unless the missing member becomes available again or you reconfigure the ReplicaSet to {votes: 0, priority: 0} on the non-available member.

Instead of waiting forever, MongoDB should rollback the change and raise an error. This behavior should apply mainly for operations where {w: "majority"} is set implicitly and cannot be changed by the user, for example at "renameCollection" issued on 'mongos'. You may introduce a new optional parameter which controls whether the command waits forever (the current behavior) or raises an error after a given timeout.