-
Type:
Bug
-
Resolution: Unresolved
-
Priority:
Major - P3
-
None
-
Affects Version/s: None
-
Component/s: Cursors, Disagg CI-blocker
-
Storage Engines, Storage Engines - Foundations
-
SE Foundations - 2025-10-10
-
5
This comment at the top of the layered cursor modify functions:
* WT_CURSOR->modify method for the layered cursor type. This function assumes the modify will * be done on the btree that we originally calculate the diff from. Currently, we only allow * writes to the stable table so the assumption holds. TODO: revisit this when we enable writing * to the ingest table.
The test/format test was only in leader mode, so it doesn't apply to the situation in item one, but the server is now attempting modifies on followers as part of oplog application.
The "initial support" terminology means this should be functionally complete, even if performance isn't optimal. The current plan is to search the ingest table for the key being modified, then insert a full value if we can't find it (and modify if we do find it). The former case is hopefully relatively uncommon, so maybe performance is acceptable.
(This ticket used to include another problem related to an assert in the layered modify while running test/format. This is now has its own ticket WT-15530).
- is depended on by
-
SERVER-111602 Remove special handling for using 'wiredTigerCursorModify' for DSC
-
- Blocked
-
- is related to
-
WT-15530 Fix WT_MODIFY memory buffer error in test/format
-
- Open
-
-
WT-15386 test/format (disagg.mode=switch) WT_CURSOR.remove assertion failure WT_CURSTD_KEY_INT not set
-
- Closed
-
- related to
-
WT-15475 (disagg.mode=leader) test/format failure: Truncate Invalid argument
-
- Open
-
-
SERVER-111601 Temporarily skip using 'wiredTigerCursorModify' in DSC
-
- Closed
-