-
Type: Build Failure
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: None
-
(copied to CRM)
-
1
-
5
-
2023-02-23 "Stoney Baloney"
-
v6.3, v6.2, v6.0, v5.0, v4.4
If we have a big page with a lot of updates, reconciliation can take a long time. It is a problem for eviction as it will block the read write operations and eviction.
The majority time of reconciliation is spent in inserting updates to the history store but building the disk image can also take more than 1 minute.
We need to figure out how to avoid long reconciliation in eviction.
- causes
-
WT-10723 Avoid checking all updates to decide the need for save update
- Closed
- is depended on by
-
SERVER-70695 Use config fuzzer on sharded tests
- Closed
- is duplicated by
-
WT-9357 Generation drain timeout after calling compact_checkpoint
- Closed
-
WT-8442 test_checkpoint_row_stress_sweep_timestamps : resource took too long to drain
- Closed
-
WT-10436 Fix format eviction generation drain failure (5.0)
- Closed
-
WT-10442 Fix eviction generation drain timeout in test_checkpoint (5.0)
- Closed
-
WT-10772 failed: make-check-test on macos-1014-cmake [wiredtiger-mongo-v5.0 @ c7cea9c7]
- Closed