-
Type:
Task
-
Resolution: Fixed
-
Priority:
Major - P3
-
Affects Version/s: None
-
Component/s: None
-
None
-
Query Execution
-
Fully Compatible
-
QE 2026-06-08, QE 2026-06-22
-
None
-
None
-
None
-
None
-
None
-
None
-
None
- Chunk migrations do not cause problems with the sharded scan (e.g. a migration causes an invalid document to move and we don't 'see' it). I'm assuming that this isn't a problem because we should have range preservers which protect us from this, but we should have some test coverage to verify that we are good here.
- Let's say you have a single invalid document, and you initiate a sharded scan. A chunk migration then concurrently moves that invalid document. Then, before the scan completes such that the document is updated to make it valid. I assume that in this situation the scan will report that we have found an invalid doc and the user has to retry (even though it was concurrently update to be valid), is that right?
- related to
-
-
- Backlog
-