Description
What's the issue?
We currently ignore DCP stream events (since we don't care about them); this is fine, except we don't handle those sequence numbers (which is not fine). This can lead to a case where the range file suggests data exists, when it doesn't.
What's the fix?
We should handle stream event sequence numbers (we could probably internally translate them into seqno advaced).
Attachments
Issue Links
- is duplicated by
-
MB-49527 CBBM restore fails with "file does not exist" when restoring ranges
- Closed
For Gerrit Dashboard: MB-49530 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
165655,3 | MB-49530 Correctly handle ignored stream events | master | backup | Status: MERGED | +2 | +1 |