scylla-operator icon indicating copy to clipboard operation
scylla-operator copied to clipboard

run-e2e-gke-release script has no default storage class

Open rzetelskik opened this issue 8 months ago • 2 comments

What happened?

Running run-e2e-gke-release.sh script results in majority of our tests failing due to there being no default StorageClass.

https://prow.scylla-operator.scylladb.com/view/gs/scylla-operator-prow/logs/ci-scylla-operator-latest-e2e-gke-arm64-parallel/1805118759006900224#1:test-build-log.txt%3A2729

Contrary to ci-deploy script, ci-deploy-release used here creates local-csi-driver from manifests generated from /examples/common/local-volme-provisioner/local-csi-driver. The StorageClass manifest there is missing the storageclass.kubernetes.io/is-default-class: "true" annotation. SO_CSI_DRIVER_PATH is not taken into account, contrary to ci-deploy.sh script.

/kind bug /priority important-soon

What did you expect to happen?

There should be a default StorageClass or we should wire the scripts to explicitly set StorageClasses to all PVCs created.

How can we reproduce it (as minimally and precisely as possible)?

ditto

Scylla Operator version

master

Kubernetes platform name and version

n/a

Please attach the must-gather archive.

n/a

Anything else we need to know?

No response

rzetelskik avatar Jun 24 '24 14:06 rzetelskik