backup and restore tests. This made the testing times explode
with both cluster/object_store/test_backup.py and
cluster/test_refresh.py taking more than an hour each to complete
under test.py and around 14min under pytest directly.
This was painful especially in CI because it runs tests under test.py which
suffers from the issue of not being able to run test cases from within
the same file in parallel (a fix is attempted in 27618).
This patch reduces the dataset of these tests to the minimum and
gets rid of one of the tested topology as it was redundant.
The test times are reduced to 2min under pytest and 14 mins under
test.py.
Signed-off-by: Robert Bindar <robert.bindar@scylladb.com>
Closesscylladb/scylladb#28280
We switched to the size-based load balancing, which now has more
strict requirements for load stats. We no longer need only per-node
stats, but also per-tablet stats.
Bootstrapping a node triggers stats refresh, but allocating tablets on
table creation didn't. So after creating a table, load balancer
couldn't make progress for up to 60s (stats refresh period).
This makes tests take longer, and can even cause failures if tests are
using a low-enough timeout.
Fixes https://github.com/scylladb/scylladb/issues/27921
No backport becuse only master is vulnerable (size-based load balancing).
Closesscylladb/scylladb#27926
* https://github.com/scylladb/scylladb:
test: cluster: Add reproducer for missed notification in topology coordinator
topology_coordinator: Wake up the state machine after stats refresh
topology_coordinator: Move tablet_load_stats_refresh_before_rebalancing injection earlier
topology_coordinator: Fix potential missed notification
topology_coordinator: Refresh load stats after table is created or altered
tablets: Do a group0 read barrier on tablet load stats refresh
topology_coordinator: Ensure stats are refreshed in the gossip scheduling group
test: Use ManagerClient.{disable,enable}_tablet_balancing()
test: Add missing calls to disable_tablet_balancing() in tests which use move_tablet() API
test: pylib: Introduce ManagerClient.{disable,enable}_tablet_balancing()
The refresh api is expected to automatically delete
the sstable files from the uploads/ dir. Verify that.
The code that does that is currently called by
sstables_loader::load_new_sstables:
```c++
if (load_and_stream) {
...
co_await loader.load_and_stream(ks_name, cf_name, table_id, std::move(sstables_on_shards[this_shard_id()]), primary_replica_only(primary), true /* unlink */, scope, {});
```
Signed-off-by: Benny Halevy <bhalevy@scylladb.com>
Closesscylladb/scylladb#27586
This change adds the --scope option to nodetool refresh.
Like in the case of nodetool restore, you can pass either of:
* node - On the local node.
* rack - On the local rack.
* dc - In the datacenter (DC) where the local node lives.
* all (default) - Everywhere across the cluster.
as scope.
The feature is based on the existing load_and_stream paths, so it
requires passing --load-and-stream to the refresh command.
Also, it is not compatible with the --primary-replica-only option.
Signed-off-by: Robert Bindar <robert.bindar@scylladb.com>
Closesscylladb/scylladb#23861
