-
Type: Improvement
-
Resolution: Unresolved
-
Priority: Major - P3
-
None
-
Affects Version/s: 3.0.7
-
Component/s: WiredTiger
-
Storage Engines
-
Storage 2017-01-23
-
(copied to CRM)
This is a spinoff from SERVER-19954. Repro scripts are there.
Once there are more than 10K tables, I now see all the time being spent opening and closing sessions and cursors, because the caching layer bounds how many we keep open since 3.0.6. I can code around that, but then I'm not testing how the system actually behaves.
- related to
-
SERVER-19954 Extended stall during checkpoint with large number of dirty tables under WiredTiger
- Closed