diff options
author | Dianna Hohensee <dianna.hohensee@mongodb.com> | 2020-04-30 12:52:44 -0400 |
---|---|---|
committer | Evergreen Agent <no-reply@evergreen.mongodb.com> | 2020-05-06 13:24:38 +0000 |
commit | da4a8e0e85627d6febd8bce3fd87f221e0ff97c6 (patch) | |
tree | 1acabbff065b4995ef8168ed464fefa089c2a07c | |
parent | 74ea6aa4dfd7b3b4acf090084d36891834a49385 (diff) | |
download | mongo-da4a8e0e85627d6febd8bce3fd87f221e0ff97c6.tar.gz |
SERVER-47582 clearing the oplogTruncateAfterPoint during stepdown must not take the PBWM
-rw-r--r-- | src/mongo/db/repl/replication_coordinator_external_state_impl.cpp | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/src/mongo/db/repl/replication_coordinator_external_state_impl.cpp b/src/mongo/db/repl/replication_coordinator_external_state_impl.cpp index c8b740275ad..603d111e5a3 100644 --- a/src/mongo/db/repl/replication_coordinator_external_state_impl.cpp +++ b/src/mongo/db/repl/replication_coordinator_external_state_impl.cpp @@ -805,6 +805,10 @@ void ReplicationCoordinatorExternalStateImpl::_stopAsyncUpdatesOfAndClearOplogTr // above, if writing is imminent, so we must make sure that the code completes fully. StorageControl::waitForJournalFlush(opCtx); + // Writes to non-replicated collections do not need concurrency control with the OplogApplier + // that never accesses them. Skip taking the PBWM. + ShouldNotConflictWithSecondaryBatchApplicationBlock shouldNotConflictBlock(opCtx->lockState()); + // We can clear the oplogTruncateAfterPoint because we know there are no user writes during // stepdown and therefore presently no oplog holes. _replicationProcess->getConsistencyMarkers()->setOplogTruncateAfterPoint(opCtx, Timestamp()); |