-
Type: Improvement
-
Resolution: Duplicate
-
Priority: Major - P3
-
None
-
Affects Version/s: None
-
Component/s: Replication
-
None
-
Replication
When doing an initial resync for a secondary where large indexes are created it easily ends up to a "too stale" state as the primary's oplog is too small. When the index creation takes several days the oplog size would have to be enormous to handle it or you'd have to take another secondary down to copy the files over.
Why not just keep fetching the oplog during the index creation instead of trying to fetch them all after the indexes are created?
- duplicates
-
SERVER-7200 use oplog as op buffer on secondaries
- Closed
- related to
-
SERVER-14805 Use multithreaded oplog replay during initial sync
- Closed