Our current resumable index builds testing only uses index builds that are small enough to not spill to disk until shutdown occurs. We should add test coverage for the case where the index build was large enough to have already spilled to disk during the collection scan.
- is duplicated by
-
SERVER-51469 Unregister index build if we fail to resume the build
- Closed
- is related to
-
SERVER-41761 MultiIndexBlock::init() fails to clean up temporary record stores on IndexBuildBlock::init() failure
- Closed
-
SERVER-41918 CollectionBulkLoader does not anticipate exceptions from MultiIndexBlock
- Closed
-
SERVER-50293 TemporaryKVRecordStore should reset the _recordStoreHasBeenFinalized flags state on a WriteConflictException
- Closed