This spawned out of the investigation from WT-14636. The ticket here fixes the drop EBUSY loop that occurs in MongoDB following SERVER-106947 and SERVER-106835.
There are multiple criterias to create the drop/EBUSY loop issue.
- The table must have logging enabled
- The table must be empty
- Background thread checkpoint cleanup dirties the same table
All three criterias must be reached for the drop/EBUSY issue to occur.
- depends on
-
WT-14636 Investigate pending table drop
-
- In Progress
-
- is related to
-
SERVER-106835 Hanging drop-pending collections without holders in MongoDB 8.x
-
- Open
-
-
SERVER-106947 Infinite log messages after dropping database
-
- Blocked
-
-
WT-15228 Create the prepared_id_transaction and prepared_id_transaction_uint api
-
- Closed
-
-
WT-15234 Test deleting local files on startup in disagg
-
- Closed
-
- related to
-
WT-15243 Bulk cursor and drop segmentation fault
-
- Open
-
-
WT-15173 Coverity analysis defect 175782: Dereference null return value
-
- Closed
-
-
WT-15002 Remove the WT_DELTA_CELL_LEAF structure and change leaf page delta format
-
- Open
-
-
WT-14781 Disagg python testing: change "skipped recovery RTS due to disagg" from warning to info message
-
- Closed
-
-
WT-14994 Make non-read-only connection flags atomic
-
- Closed
-
-
WT-15030 Improve TSAN script to work on new code
-
- Closed
-
-
WT-15202 Disagg python testing: Update hook_disgg.fail and skip unsupported tests in disagg mode
-
- Closed
-
-
WT-15095 Refine eviction selection of pages with few updates
-
- Closed
-
-
WT-12027 ex_backup_block issue with cleanup
-
- Closed
-