-
Type:
Task
-
Resolution: Fixed
-
Priority:
Major - P3
-
Affects Version/s: None
-
Component/s: Test Csuite
-
Storage Engines - Foundations
-
337.007
-
SE Foundations - 2026-09-01, SE Foundations - 2026-09-15
-
3
A follower node may pick up a checkpoint which is ahead of schema operations currently being applied. When it happens, then shared metadata (from the new checkpoint) maybe different from the current state.
Currently, WT will panic if such checkpoint is picked up.
TODO:
Schema Disagg Abort test needs to accommodate for this scenario and delay checkpoint pickup until workers caught up with the checkpoint.
- blocks
-
WT-18528 schema_disagg_abort: FAILED: Node 0: peer did not adopt the step-down checkpoint (lsn 1684) in 30 seconds
-
- Closed
-
-
WT-18529 Schema disagg abort: Drop for EBUSY 30 seconds timeout
-
- Closed
-
- is duplicated by
-
WT-18510 failed: schema-disagg-abort-multi-switch-crash-test-disagg-1 on amazon2023-disagg-asan-stress [wiredtiger-disagg @ 5cc20887]
-
- Closed
-