In addition to the main GitHub workflow which runs the driver tests with a 3-node cluster with the newest Scylla, we also have workflows that run Scylla with some custom configuration (e.g. enabled TLS or password authentication). For reasons that are unclear to me, the tests use custom docker images which include the custom configuration. Those containers were created very long time ago (3 years ago I think) so they are running versions that are not supported anymore.
We should fix this situation and perhaps change the tests not to use those images but rather configure the cluster differently.
If those configs can be shared across all drivers, manager, monitoring (and possibly one day, with Scylla itself!) that'd be great. No need to reinvent the wheel in every repo.
In addition to the main GitHub workflow which runs the driver tests with a 3-node cluster with the newest Scylla, we also have workflows that run Scylla with some custom configuration (e.g. enabled TLS or password authentication). For reasons that are unclear to me, the tests use custom docker images which include the custom configuration. Those containers were created very long time ago (3 years ago I think) so they are running versions that are not supported anymore.
We should fix this situation and perhaps change the tests not to use those images but rather configure the cluster differently.