-
Type:
Task
-
Resolution: Fixed
-
Priority:
Critical - P2
-
Affects Version/s: None
-
Component/s: Checkpoints, Test Model
-
None
-
Storage Engines, Storage Engines - Transactions
-
StorEng - 2025-04-25, SE Transactions - 2025-05-09
-
5
While addressing WT-14376, we uncovered a bug where a crash occurring after the log flush during the checkpoint process leads to a recovery failure with a block_merge error.
Acceptance Criteria:
The acceptance criteria of this ticket is to implement two new crash points (after the log flush and after metadata sync) controlled by the debug.checkpoint_crash_point configuration to simulate and investigate crashes happening after the checkpoint log flush.