This brief report highlights some intriguing commits to the scylla-cluster-tests.git master over the past week, specifically within the 086c12d2…6a910e04 range.
During this period, we had 27 non-merge commits from 5 authors. Here are some of the notable changes:
For those interested in the nemesis list that SisyphusMonkey is about to create, you can obtain it by running the following command:
hydra nemesis-list --backend aws --config test-cases/PR-provision-test-docker.yaml
This command will output the list of nemesis based on the configuration files used. Check out the changes here.
Tests running on K8S-EKS backend have transitioned to using the real S3 service for Scylla-manager backups, moving away from Minio. This change enables us to start running ‘mgmt restore’ nemesis, and a job for this has been added. You can see these updates here and here.
The EndOfQuota Nemesis has been updated to read all mount options and search for the expected value. This nemesis can now be run on all backends.
Previously, only EventsFilter had the option to set extra time before eviction. Now, DbEventsFilter has been granted the same capability. See the change here.
Lastly, we’ve begun to collect runner metrics using node_exporter and have integrated it into the monitoring stack. This will assist us in understanding the load on the sct-runner, identifying issues with the test code, and determining if we need to choose larger instances/disks for certain tests. A new image version for sct-runner (1.7) was released with this change.
Looking forward to updating you on the next batch of changes in the scylla-cluster-tests.git master!