-
Type: Bug
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: None
-
8
-
Storage Engines - 2022-10-31, Storage Engines - 2022-11-14
I have a CONFIG that is failing on the current mirror branch (which includes the fix from WT-9522). The first mirrored table is FLCS. The failure has the same signature as WT-9522. I have been able to reproduce it twice (it can take a while - with 5 parallel runs it has hit "overnight").
The error is:
mirror: 2236015/2236015 mismatch: table:T00002: {0002236015.00/opqrstuvwxyza}/{0002236015/HMNOPQRSTUVWXYZABCDEFGHIJ} [0x48], table:T 00001: {#}/{0} mirror error: base cursor (table 2): dumping to WT_TEST.1/FAIL.pagedump.1 mirror error: table cursor (table 1): dumping to WT_TEST.1/FAIL.pagedump.2 mirror error: base key number 2236015 in table 3: dumping to WT_TEST.1/FAIL.pagedump.3 mirror error: base key number 2236015 in table 4: dumping to WT_TEST.1/FAIL.pagedump.4 t: FAILED: void table_verify_mirror(WT_CONNECTION *, TABLE *, TABLE *, const char *)/322: failures == 0
- is depended on by
-
WT-10108 Add a data structure encapsulating user level truncate context
- Closed
- is related to
-
WT-9944 Truncate write skew vs. format mirroring
- Closed
-
WT-4158 Fix concurrent behaviour of insert with truncate.
- Backlog
- related to
-
WT-9830 Do not allow the fast truncate on-disk state to go backwards because of visibility
- Closed