Closed supersaw-it closed 2 months ago
The Scylla Operator project currently lacks enough contributors to adequately respond to all issues.
This bot triages un-triaged issues according to the following rules:
lifecycle/stale
is appliedlifecycle/stale
was applied, lifecycle/rotten
is appliedlifecycle/rotten
was applied, the issue is closedYou can:
/remove-lifecycle stale
/close
/lifecycle stale
I think you'd have to at very least copy the data to empty volume temporarily mounted somewhere else, so it's identical. I can see how copying the files into a running pod can be racy.
The only path we support is backup+restore https://operator.docs.scylladb.com/stable/nodeoperations/restore.html
Due to the recent operator issue https://github.com/scylladb/scylla-operator/issues/1690 , I decided to completely reinstall my local test cluster.
Before doing that, I moved the data from
/var/lib/scylla
to my machine, hoping I could restore it once I have a running scylla cluster up & running.However upon moving the following directories _commitlog data hints viewhints that exist in the aforementioned scylla directory back to the pod with
kubectl cp
, and queyring the DB, it seems like the database does not recognize the data.First it complained about the fact that the keyspace did not exist, I created it. Then, after a successful query, I got 0 rows. I created a table with the same name too, and moved the corresponding <...>big-*, e.g. me-7173-big-Data.db, files to the directory of newly created table. Same result with 0 rows.
Nodetool output:
Is there a way to make scylla accept the data I have backed up?
Thanks in advance!