This short report brings to light some interesting commits to scylla-cluster-tests.git master from the last week.
Commits in the 9c63d6d6…0be52b00 range are covered.
There were 33 non-merge commits from 13 authors in that period. Some notable commits:
Refactored HDR histogram support to work with multiple benchmarking tools, not just cassandra-stress. Stress threads now automatically detect HDR support, enabling future integration with tools like latte, scylla-bench, and cql-stress-cassandra-stress.
Enabled a tombstone GC verification thread for large-partitions tests, requiring an adjustment to tombstone lookup to match the new output format of scylla sstable dump-data.
Due to a change in the semantics of the enable_tablets Scylla config parameter, is_tablets_feature_enabled now returns the default tablets setting for new keyspaces.
Nemesis can now be flagged with supports_high_disk_utilization to indicate compatibility with 90% disk utilization tests, and a 90% utilization test has been added.
Fixed Gemini command arguments when no oracle cluster is provided, resolving an issue in upgrade tests.
Enabled parallel node operations by default, improving cluster provisioning speed. This enhances test scenarios that involve adding multiple nodes (nemesis_add_node_cnt > 1), aligning with recent Scylla rapid scaling improvements.
See you in the next issue of last week in scylla-cluster-tests.git master!