-
Type: Improvement
-
Resolution: Done
-
Priority: Blocker - P1
-
Affects Version/s: None
-
Component/s: Replication
-
None
With the new behavior of using minvalid to determine if a initial sync has been done, and regular replication should start, leads to some significant problems.
Cases:
- Upgrade a set without minvalid collectoin: all nodes going into startup2(forever) if no primary/secondary is up, otherwise all nodes drop+resync all data.
- Variations of this will cause the initial sync to fail/elect a primary because of the oplog stale rules
- Removing the minvalid collection causes a full resync (ind. of data/indexes), or host going offline in startup2 until one can be done
The upgrade case is bad since we had no need for the minvalid collection and it was not maintained nor guaranteed (esp. on the primary) or if replicas were seeded with a copy of the files without it.
- is related to
-
SERVER-7652 Initial sync criteria needs to check minValid, not the oplog
- Closed
-
SERVER-8556 Add failpoint tests to initial sync to test various MinValid phases
- Backlog