-
Type: New Feature
-
Resolution: Duplicate
-
Priority: Minor - P4
-
None
-
Affects Version/s: 2.4.4
-
Component/s: None
-
None
When we bring up a secondary via initial sync to compact all the databases and collections and then elect it primary, the performance suffers for weeks or months afterwards.
We think this is because the padding factor gets reset for all collections, and it takes the db a long time to relearn how much space it should allocate for each write that moves things around on disk. So it ends up doing a lot more moves and writes to disk than it really should.
If we were able to set the padding factor manually, we could run a sync script to iterate through all the collections and set the padding factor on the secondary to match that on the primary, before doing the swap. We would LOVE to have this feature.
- duplicates
-
SERVER-3752 padding factor overcompaction and related needed improvements
- Closed
- related to
-
SERVER-1810 ability to set minimum allocation size per collection
- Closed