If any voting node in the set does not journal (inMemory or journaling disabled), you must set writeConcernMajorityJournalDefault to false. When set to false, replication waits will no longer wait for writes to be durable. This includes w:majority writes, the user/group admin commands, the replSetStepDown command, and possibly others I'm not thinking of.
With this flag set to false, majority acknowledged writes could possibly roll back, in the face of one node failure.
- is related to
-
SERVER-27138 Cannot run replSetStepDown for a 3.4 replica set with journaling disabled
- Closed