Details
-
Improvement
-
Status: Resolved
-
Low
-
Resolution: Duplicate
-
None
-
None
Description
If tables are TTLed in your Cassandra and you try to backup the corresponding SSTable to a new cluster, your backup will be lost automatically after a certain period of time, because TTL are still expiring data in the backup SStable. Once the TTL elapsed, data become inaccessible and finally will be deleted physically by compaction after GCGraceSecondes. So we cannot really keep the backup data in a new cluster without removing these existing TTL.
So it might be good to add ability to do batch TTL removing.
Attachments
Issue Links
- duplicates
-
CASSANDRA-7471 sstableloader should have the ability to strip ttls
- Open