Description
What's the issue?
Now that we've fixed MB-47653, we have the issue that we maintain tombstones during a merge; this will result in tombstones being created on the target cluster during a restore (this may not be desirable).
What's the fix?
We should investigate whether we should skip restoring tombstones which are behind the purge seqno (in a similar fashion to the previous behavior with merge).
Attachments
Issue Links
- relates to
-
MB-47653 [CBM] Merge may discard/disregard tombstones resulting in mutations being "revived"
- Closed