This short report brings to light some interesting commits to scylla-cluster-tests.git master from the last week.
Commits in the 2cf61970…baf5542e range are covered.
There were 36 non-merge commits from 7 authors in that period. Some notable commits:
Added support for TEST_ERROR Argus status, indicating a run failed due to a test error, such as problems with cloud resources (e.g., insufficient capacity or spot termination).
Performance tests can now use BYO ScyllaDB.
The RPC transport error that occurs when one node boot is interrupted and Scylla does not start is ignored in BootstrapStreamingError nemesis.
New jobs were created for tablet-enabled runs.
We created a new branch for performance tests from master and refactored the directory structure. All the perf regression tests were moved into their own folder in each release (master/enterprise), based on the branch they originated from.
We fixed BYO ScyllaDB to work with multidc setups, which required copying AMIs to multiple regions.
A new feature in ScyllaDB enables advanced compression of internode RPC messaging, now supported by SCT, and was enabled in two tests: multidc rolling-upgrade and one longevity.
An elasticity test measuring the time to scale out/in a cluster with parallel nodes bootstrap/decommission and doubling the load after growing was finalized. It supports both cql-stress and cassandra-stress tools as loaders. Doubling the load in GrowShrinkNemesis can also be set for longevity tests with the nemesis_double_load_during_grow_shrink_duration parameter.
The default version of Scylla Manager was updated to 3.3.0.
See you in the next issue of last week in scylla-cluster-tests.git master!