-
Allowing for purging these without a stop-remove-start cycle. With this, validate together with the new API should be enough to purge nodes of bad sstables.
-
We've seen cases in the field where deleting a large number of sstables in parallel (e.g. thousands of sstables holding over a hundred thousands files) by offstrategy compaction following repair that …
-
Can't seem to get it to work
```
[automaton@ip-172-31-9-120 device_monitoring_timestamps]$ pwd
/home/automaton/36134/sstables_10.246.171.127/health/device_monitoring_timestamps
[automaton@ip-172…
-
HEAD: 65e42e4166cef5acafa7ce7f884d6cc2b55adc35
*Description*
Today `scylla sstable validate-checksums` if it succeeds to run will always return `0` regardless if malformed sstables were found or n…
-
I think that the compaction process has a memory leak.
I will try to figure out why.
-
Recently, we have seen numerous examples of a need for a procedure which reliably rids a cluster of all bad sstables. At the heart of such a procedure is scrub, yet it has some deficiencies which mean…
-
These were introduced recently, and permit using the vanilla sstable readers when reading foreign ssts, which is more efficient than wrappers. Documented in https://github.com/cockroachdb/pebble/blob/…
-
Consider a workload where the keyspaces a-f and u-z are write heavy, while the keyspace g-t is read only and heavy. Incoming writes to a-f and u-z keyspaces can easily be included in the same sstable,…
-
Currently, we only check checksum when reading sstables that have compression enabled. We should also check checksum when reading non-compressed sstables using the CRC component. CRC component stores …
-
Currently, we say that a sstable needs cleanup if its token range isn't fully contained in any of the owned token ranges (needs_cleanup() at database.cc). That's not very good because a sstable token …