# Last week in scylla-cluster-tests.git master (issue #153; 2026-09-26)

**URL:** <https://forum.scylladb.com/t/last-week-in-scylla-cluster-tests-git-master-issue-153-2026-09-26/5512>\
**Category:** ScyllaDB\
**Tags:** git-testing-news\
**Created:** [September 26, 2026, 8:16am UTC](https://forum.scylladb.com/t/last-week-in-scylla-cluster-tests-git-master-issue-153-2026-09-26/5512 "2026-09-26T08:16:07Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![soyacz](https://sea2.discourse-cdn.com/flex016/user_avatar/forum.scylladb.com/soyacz/32/557_2.png) [@soyacz](https://forum.scylladb.com/u/soyacz)\
**Post date:** [September 26, 2026, 8:16am UTC](https://forum.scylladb.com/t/last-week-in-scylla-cluster-tests-git-master-issue-153-2026-09-26/5512/1 "2026-09-26T08:16:07Z")

</div>

This short report brings to light some interesting commits to [scylla-cluster-tests.git master](https://github.com/scylladb/scylla-cluster-tests) from the last week. Commits in the bc0f720b…9daf0294 range are covered.

There were 35 non-merge commits from 7 authors in that period. Some notable commits:

The multi-keyspaces longevity scenario was [switched from cassandra-stress to latte](https://github.com/scylladb/scylla-cluster-tests/commit/9daf02940d1beea6605c7c357302a6131e2cb098), driving all 1000 keyspaces from a single command on 2 loaders instead of 10. Reads are now validated, and the load covers all 1000 tables.

The rolling-upgrade test now [runs its complex-schema workload on latte](https://github.com/scylladb/scylla-cluster-tests/commit/8f0a762130e6d723e36dde075968da83bef8cceb) through the new `data_dir/latte/complex_schema.rn` script, with latte equivalents for `verify_more` and `verify_delete`.

A new logstor performance test [drives 60% of the load against a logstor table and 40% against an LSM table](https://github.com/scylladb/scylla-cluster-tests/commit/4ea07ede4b89864393c1304588e2ac49a4b8836f) on the same cluster, with a separate Argus latency row per engine.

The baremetal backend gained [support for `simulated_racks`](https://github.com/scylladb/scylla-cluster-tests/commit/21a4096d7addda43b0dcb9e3b8018050a5036f51), so tests with RF \> 1 keyspaces, such as the baremetal throughput perf-regression case, can run on physical hosts.

Latte was [bumped to 0.52.0](https://github.com/scylladb/scylla-cluster-tests/commit/2e96d9f10f769945fb17d29d439bc98a6d0e88ca). The release moves to version 1.9.0 of the Rust `scylla` driver crate, adds single-partition batches to the row distribution presets, and sends ad-hoc statements unprepared in the `execute` context function.

An implementation plan for [preserving coredumps during the out-of-space nemeses](https://github.com/scylladb/scylla-cluster-tests/commit/b41a8cd7720d7372a6dcd944eedb93341580439a) was added.

A mini-plan proposes [a `*` wildcard for `availability_zone`](https://github.com/scylladb/scylla-cluster-tests/commit/dfae2c8c9af211e3dc4294ea250d88451e6ab157), so a test can ask for “any AZ” or for several distinct zones without naming them.

See you in the next issue of last week in scylla-cluster-tests.git master!