Description
The data range logic is overly complicated and very difficult to debug. We also store the same information in 2 places the range.json and the data directory, It also looks like the range.json serves not much purpose at all so it would be good to remove it and simplify all the logic so that is easier to mantain as well as avoiding reading and writing the range json every time we backup a bucket.
Attachments
Issue Links
- blocks
-
MB-38686 [CBM] Add support for resuming restores
- Closed
- causes
-
MB-58194 [CBM] --resume loses data when backing up to S3 with less than 5MB completed in a vBucket
- Closed
-
MB-58094 [RECOVERY] Data is not always transferred concurrently
- Closed
- relates to
-
MB-47639 [CBM] collect-logs should collect both the old/new style range files
- Closed
-
MB-47640 [CBM] DCP code should use "options" structures
- Closed
-
MB-47641 [CBM] Couchbase sink should use "options" structures
- Closed
-
MB-47645 [CBM] The Couchbase sink should maintain/persist data ranges
- Closed
-
MB-47646 [CBM] Tidy up the VBucketBackupWriter 'Set' function
- Closed
-
MB-47638 [CBM] Tidy up the getPrevBackup/getMostRecentBackup code
- Closed
-
MB-47644 [CBM] Restore worker pool should use "options" structures
- Closed
-
MB-47691 [CBM] The Couchbase source should use an "options" structure
- Closed
-
MB-47706 [CBM] Tidy up the handling of mutations/deletions/expirations
- Closed
-
MB-47719 [CBM] [Objstore] Add new style range files to archive metadata compression
- Closed
-
MB-47744 [CBM] Re-add snapshot marker/seqno validation
- Closed
-
MB-47642 [CBM] Remove the snapshot/failover file implementations
- Closed