RecoversAndUnblocksAfterStepdownWhileStoppingBlocking and RecoversWhenBlockingWithNoOperationsOnDisk trigger failovers and attempt to acquire a handle to the MigrationBlockingOperationCoordinatorV2 instance after step up, however the coordinator at that point is in a state where it is actively trying to clean up (i.e. there are no ongoing operations). If the threads execute such that the cleanup completes before we do the lookup, the tests will spuriously fail.