e2e: add nvmeof driver testing support with k8s-e2e-external-storage#6214
e2e: add nvmeof driver testing support with k8s-e2e-external-storage#6214nixpanic wants to merge 5 commits intoceph:develfrom
Conversation
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
Job is still running after addressing the MarkDown linter failures. |
44d2d02 to
7f22c0f
Compare
|
@Mergifyio rebase |
7f22c0f to
0a51e55
Compare
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
|
Deprecation notice: This pull request comes from a fork and was rebased using |
✅ Branch has been successfully rebased |
0a51e55 to
9d2cfcf
Compare
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
1 similar comment
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
691d9e4 to
ca6dc6e
Compare
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
ca6dc6e to
9e6d963
Compare
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
9e6d963 to
5a2e14b
Compare
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
|
job is running need to inspect the result of the nvmeof driver once done. /cc @gadididi |
5a2e14b to
783a7fd
Compare
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
1 similar comment
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
8cc3df5 to
47c3972
Compare
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
47c3972 to
51d1351
Compare
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
51d1351 to
f182e4c
Compare
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
|
Disable |
1 hour is too little with the added nvme-of testing. See-also: ceph#6214 Signed-off-by: Niels de Vos <ndevos@ibm.com>
|
/test ci/centos/k8s-e2e-external-storage/1.34 |
|
/test ci/centos/k8s-e2e-external-storage/1.33 |
|
/test ci/centos/mini-e2e/k8s-1.35/nvmeof |
This one is important to check, the nvme-of gw gets deployed by the e2e suite, and now also by |
Looks good 🥳 |
|
|
||
| ```bash | ||
| export GATEWAY_ADDRESS=10.242.64.32 | ||
| export LISTENERS='[{"address": "10.242.64.32"}]' |
There was a problem hiding this comment.
Need to add mandatory hostname?
There was a problem hiding this comment.
indeed, the readme is not in-sync anymore
| driver: nvmeof.csi.ceph.com | ||
| parameters: | ||
| clusterID: @@CLUSTER_ID@@ | ||
| csi.storage.k8s.io/snapshotter-secret-name: rook-csi-nvmeof-provisioner |
There was a problem hiding this comment.
need to use rook-csi-rbd-provisioner ?
There was a problem hiding this comment.
right, probably. Maybe this caused the volume-clone failure 🤔
|
|
||
| cat <<EOF >"${TEMP_DIR}/operatorconfig.yaml" | ||
| apiVersion: csi.ceph.io/v1alpha1 | ||
| apiVersion: csi.ceph.io/v1 |
There was a problem hiding this comment.
This commit can be dropped now?
There was a problem hiding this comment.
that will happen automatically when this PR is rebased
| clusterID: @@CLUSTER_ID@@ | ||
| pool: replicapool | ||
| imageFeatures: layering | ||
| # FIXME: use nvmeof specific secrets, but they don't exist yet? |
There was a problem hiding this comment.
I dont think there is right now nvmeof secrets. As far as I know it uses rbd secrets
(maybe we should reconsider in the future to use dedicated secrets)
There was a problem hiding this comment.
Yes, it would be good to use nvmeof secrets in the future. The e2e/nvmeof-deploy.go does it for our e2e suite. Probably we rely on Rook to create the secrets for the k8s-e2e-external-storage tests.
Add configuration files and scripts to enable NVMeoF driver testing in the k8s-e2e-external-storage CI job. This includes the driver manifest, StorageClass template, and VolumeSnapshotClass template for NVMeoF. The create-storageclasses.sh script now auto-detects the NVMeoF gateway address from the rook-ceph-nvmeof service, or it can be configured via the GATEWAY_ADDRESS and LISTENERS environment variables. Assisted-by: Claude Code <noreply@anthropic.com> Signed-off-by: Niels de Vos <ndevos@ibm.com>
Signed-off-by: Niels de Vos <ndevos@ibm.com>
Deploy the NVMe-oF CSI driver alongside RBD, CephFS, and NFS drivers when using the ceph-csi-operator deployment script. Assisted-by: Claude Code <noreply@anthropic.com> Signed-off-by: Niels de Vos <ndevos@ibm.com>
POD_ADDRESS is needed as the nvmeof-gw is deployed in rook-ceph, and ceph-csi in a dedicated testing namespace. Resolving the short-hostname of the gateway is not possible from outside the rook-ceph namespace. The gateway is configured with a host-id as the short-hostname, which needs to match what is passed in the LISTENERS. Assisted-by: Claude Code <noreply@anthropic.com> Signed-off-by: Niels de Vos <ndevos@ibm.com>
Signed-off-by: Niels de Vos <ndevos@ibm.com>
8b5087e to
095dc98
Compare
|
/test ci/centos/k8s-e2e-external-storage/1.35 |
| # Support ReadWriteMany access modes (block mode only) | ||
| RWX: false |
There was a problem hiding this comment.
this test uses filesystem volumes (works on CephFS), we can not restrict the tests to block-mode only
|
/test ci/centos/mini-e2e-helm/k8s-1.34/rbd |
Running RBD (with VaultKMS) test for #6236 (comment) |
Add configuration files and scripts to enable NVMeoF driver testing in the
k8s-e2e-external-storage CI job. This includes the driver manifest,
StorageClass template, and VolumeSnapshotClass template for NVMeoF.
Also deploy the NVMe-oF gateway with Rook.
Depends-on: #5672 for NVMe-oF gateway support