-
Followup to #82898 which created the basic infrastructure, configuration scheme, and did experimentation with regular and elastic kv0 traffic.
- [x] Eliminate the need to provide the disk-name and ba…
-
Backups consume space - which is costly. It also takes time to upload (and certainly to download)
Regretfully, rclone only support gzip[1] (and not zstd) for compression. This is somewhat unfortunat…
-
drt-chaos is in a really bad state now. Several of the nodes are down, with at least two of them complaining about consistency issues:
F241031 14:03:24.454366 1160675458 kv/kvserver/replica_consiste…
-
In the 23.1 test cluster, we're observing higher than expected exercising of the new flushable ingests code path. This code path is triggered one or more of the sstables in an ingest overlaps with one…
-
There's the need to atomically delete a bunch of sstables. For disk-backed sstables this is implemented inside sstable_directory::delete_atomically() which puts sstables names into deletion logs, then…
-
*Installation details*
Scylla version: 5.4.0~dev-0.20230527.e4d6ed7a7048 with build-id 6dc67925bbb00bf45be5af823788c65
688817efb
Cluster size: 3
OS (RHEL/CentOS/Ubuntu/AWS AMI): ami-0bf7ab0d99f14b…
-
Currently tablet_incremental_selector returns empty sstables vector if a range is not on this replica. It should not need to check the tablet's existence.
The responsibility for generating ranges f…
-
We run read repairs on our cluster and were having trouble with old SSTables not getting dropped until we added `'unchecked_tombstone_compaction': 'true'` to the compaction settings in the schema (whi…
-
In https://github.com/scylladb/scylladb/issues/21611
some interesting problems identified by looking at read concurrency semaphore timeout traces:
```
[shard 0:stmt] reader_concurrency_semaphore - (…
-
Sstable compaction has an optimization which ignores whole sstables which are expired and irrelevant (see commit e9076f3). Sstable reads should have such optimization as well.
Refs https://github.c…