Description of problem: Could not list CSI provisioner on web when create storage class Version-Release number of selected component (if applicable): 4.7.0-0.nightly-2020-12-21-131655 How reproducible: Always Steps to Reproduce: 1.In the OCP web console, click storage--storage classes 2.Click Create Storage Class 3.When click the drop down list of Provisioners, some csi related provisioner listed. Actual results: No csi related provisioner listed when create storage class on web Expected results: There should be some csi provisioners, such as pd.csi.storage.gke.io Master Log: Node Log (of failed PODs): PV Dump: PVC Dump: StorageClass Dump (if StorageClass used by PV/PVC): Additional info:
Moving to Console. "Create Storage Class" page needs some updates: 1. It should allow users (admins) to set a custom provisioner, outside of the hardcoded list - they may want to use a custom provisioner for e.g. flex volumes. 2. In 4.6 we added some CSI drivers to the default installation and these new Provisioners should be added to the list: ebs.csi.aws.com csi.ovirt.org manila.csi.openstack.org 3. In 4.7, we're adding also: pd.csi.storage.gke.io cinder.csi.openstack.org 4. Instead of hardcoding the list of CSI drivers (= Provisioners) in 2. and 3., it's possible to list installed CSI drivers dynamically. List CSIDriver objects, "oc get csidriver", and collect CSIDriver.Name fields to get values allowed in StorageClass.Provisioner. Still, please keep the list of kubernetes.io/* provisioners there, they're hardcoded in Kubernetes and can't be listed (they're not CSI drivers by in-tree volume plugins).
Didn’t have time to work on this. Reassigning this to Jakub
Created attachment 1763916 [details] CSIDrivers are listed dynamically I confirmed that on AWS cluster, CSIDrivers in the Provisioner list are dynamically listed. # oc get csidrivers NAME ATTACHREQUIRED PODINFOONMOUNT MODES AGE block.csi.ibm.com true false Persistent 54s (newly added CSIDriver) ebs.csi.aws.com true false Persistent 8h (default CSIDriver on this AWS cluster) example.ember-csi.io true true Persistent 4m37s (newly added CSIDriver) All above CSIDrivers are listed in the Provisioner list, see screenshot
console should be working as expected. this seems a storage issue: missing CSIDriver for GCP PD
Thanks, Yadan Pei, for confirming. This PR on console https://github.com/openshift/console/pull/8337, the user will add the flexibility to add additional parameters if the provisioner parameters are not known to the console. Probably we can have a cloned bug to fix the storage backed for the GCP PD driver.
> This PR on console https://github.com/openshift/console/pull/8337, the user will add the flexibility to add additional parameters if the provisioner parameters are not known to the console This sounds reasonable, if this is the issue we gonna fix in this bug I think the description should be totally different. > Probably we can have a cloned bug to fix the storage backed for the GCP PD driver Open a new bug tracking the storage backend issue is the best
This is the storage bug for gcp csi driver https://bugzilla.redhat.com/show_bug.cgi?id=1939865
1. check CSI drivers with CLI # oc get csidriver NAME ATTACHREQUIRED PODINFOONMOUNT STORAGECAPACITY TOKENREQUESTS REQUIRESREPUBLISH MODES AGE pd.csi.storage.gke.io true false false <unset> false Persistent 84m 2. pd.csi.storage.gke.io is in the provisioners list on create storage class page 3. Install Operator for IBM block storage CSI driver then create some other csi drivers which are unknown to console # oc get csv -n default NAME DISPLAY VERSION REPLACES PHASE ibm-block-csi-operator.v1.5.0 Operator for IBM block storage CSI driver 1.5.0 ibm-block-csi-operator.v1.4.0 Succeeded # oc get csidriver NAME ATTACHREQUIRED PODINFOONMOUNT STORAGECAPACITY TOKENREQUESTS REQUIRESREPUBLISH MODES AGE block.csi.ibm.com true false false <unset> false Persistent 6s pd.csi.storage.gke.io true false false <unset> false Persistent 94m 4. when `block.csi.ibm.com` is selected as provisioner, we show a section for use to add `Additional parameters` Verified on 4.8.0-0.nightly-2021-04-09-222447
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Moderate: OpenShift Container Platform 4.8.2 bug fix and security update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2021:2438