Robert Bindar

An update on this, we narrowed down the problem in the IO scheduler, it is reproducible when the scheduler runs "unconfigured", no io-properties, which is the typical `iotune` use case....

> > thanks for picking this up! please also try to copy-pasta the test from [#24261 (files)](https://github.com/scylladb/scylladb/pull/24261/files#) it's not much but we have no coverage around this manifest file and...

The `min_tablet_count` tests were failing when restoring with a smaller `min_tablet_count` than the table was originally created with, the nodes were streaming to more nodes than it was expected. I...

> @robertbindar https://jenkins.scylladb.com/job/scylla-master/job/scylla-ci/22538/testReport/junit/cluster.object_store/test_backup/dev___test_py___dev___test_restore_with_streaming_scopes_gs_topology_rf_validity0__dev_1/ > > ``` > > await do_load_sstables(ks, cf, servers, topology, sstables, scope, manager, logger, prefix=prefix, object_storage=object_storage, primary_replica_only=pro) > > test/cluster/object_store/test_backup.py:852: > _ _ _ _ _ _...

@scylladbbot trigger-ci

flaky test tracked originally here https://github.com/scylladb/scylladb/issues/27421, I reopened the issue.

@scylladbbot trigger-ci

@bhalevy safe to re-review, the failing tests were unrelated to this PR.

@yaronkaikov do we have an issue with reporting? I see the bot commented everything was green but `Unit Tests Custom - debug` check is still pending for me.