This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(150.4, 65%, 24%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EFC%3C/text%3E%3C/svg%3E)
When creating PersistentVolumeClaims for storage classes provisioned by disk.csi.azure.com, the generated PersistentVolumes are bound immediately, but they are not available when starting Pods that use them. I get events like:
Warning FailedAttachVolume pod/diag-managed AttachVolume.Attach failed for volume "pvc-e63a92ef-f128-448a-bd0a-825a6aa1f5e5" : rpc error: code = Internal desc = Attach volume /subscriptions/3a9986fa-8e53-4d17-98fc-3ddf4042818b/resourceGroups/mc_cnf-ji2249-rg_cnf-ji2249_westeurope/providers/Microsoft.Compute/disks/pvc-e63a92ef-f128-448a-bd0a-825a6aa1f5e5 to instance aks-default-20624410-vmss000000 failed with could not find disk(/subscriptions/3a9986fa-8e53-4d17-98fc-3ddf4042818b/resourcegroups/mc_cnf-ji2249-rg_cnf-ji2249_westeurope/providers/microsoft.compute/disks/pvc-e63a92ef-f128-448a-bd0a-825a6aa1f5e5) in current disk list(len: 0) nor in diskMap(map[])
What is the reason for these errors?
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(76.8, 74%, 17%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EAN%3C/text%3E%3C/svg%3E)
Hello Fiderer, Christian,
Welcome to microsoft Q&A, thankyou for posting you query.
The Kubernetes system's inability to locate the requested disk is indicative of an internal error. The PVC might be expecting a disk that wasn't properly provisioned due to a configuration problem.
Check to see if your Azure account has the disk that the PVC name references.
Try restarting the "diag-managed" pod if the disk was recently provisioned. This could cause the pod to refresh and find the newly available disk.
Make sure the PVC definition uses a StorageClass that provisioned disks as expected, or that the disk name specified is accurate.
Hope this helps you.
Hello Anveshreddy Nimmala,
thank you for looking into my issue.
My use case is searching for a storage class that provides NFS-compatible volumes.
My cluster has nine StorageClasses defined, so for every storage class in the cluster I create a PVC and a pod that mounts the corresponding PV.
For each PVC, I get a PV, and the PVC are all bound to the PVCs.
For the four PVs provisioned by file.csi.azure.com, the pod gets scheduled, it runs and reports that the storage class uses CIFS (that is not NFS-compatible).
For the five PVs provisioned by disk.csi.azure.com, I first get the error mentioned above. After approx. 15 minutes, three pods are scheduled and complete successfully; the remaining two pods remain unscheduled because the "premium" storage class is not available in the Standard_D2_v2 VM class.
I still wonder why it takes approx. 15 minutes to mount the volumes from disk.csi.azure.com ...
Hello Anveshreddy Nimmala,
thank you for looking into my issue.
My use case is searching for a storage class that provides NFS-compatible volumes. My cluster has nine StorageClasses defined, so for every storage class in the cluster I create a PVC and a pod that mounts the corresponding PV.
For each PVC, I get a PV, and the PVC are all bound to the PVCs.
For the four PVs provisioned by file.csi.azure.com, the pod gets scheduled, it runs and reports that the storage class uses CIFS that is not NFS-compatible.
For the five PVs provisioned by disk.csi.azure.com, the pods remain in ContainerCreating and I get the error mentioned above. After several minutes, three pods are scheduled and complete successfully; the remaining two pods remain unscheduled because the 'premium' storage classes are not available in the Standard_D2_v2 VM class. ok....
But I still wonder why it takes several minutes to mount the volumes from disk.csi.azure.com for the 'diag-default', 'diag-managed' and 'diag-managed-csi' pods.
How can I know whether the check pods will eventually succeed or not?
Hello Fiderer, Christian,
check that both the StorageClass and Azure File Share are configured to use NFS explicitly.
https://github.com/kubernetes-sigs/azurefile-csi-driver/tree/master/deploy/example/nfs
Modify the storage class of your PVs to use a standard storage class instead of premium .since D2_V2 Virtual machine does not support premium.
If premium storage is necessary for your application, consider changing the VM size to one that supports Premium Storage.
https://learn.microsoft.com/en-us/azure/virtual-machines/dv2-dsv2-series#dsv2-series
Hope this helps you.
Hello Nimmala, Anveshreddy,
thank you for the explanation and your guidance on VM selection. This explains the issue and the problems that I encounter.
I am therefore now faced with the question of why two StorageClasses are configured in my Kubernetes cluster, which are known not to be usable.
How can such a misconfiguration be recognised and avoided?
Kind regards, Christian
Hello Fiderer, Christian,
When setting up these storage classes, there can be a problem. it is possible that the provisioner they refer to is not installed correctly, or that the class does not have the necessary permissions to access the underlying storage resource.
this is referred from storage-classes
If an answer has been helpful, please consider accepting the answer to help increase visibility of this question for other members of the Microsoft Q&A community. If not, please let us know what is still needed in the comments so the question can be answered. Thank you for helping to improve Microsoft Q&A!
If an answer has been helpful, please consider accepting the answer to help increase visibility of this question for other members of the Microsoft Q&A community. If not, please let us know what is still needed in the comments so the question can be answered. Thank you for helping to improve Microsoft Q&A!
Hello Anveshreddy Nimmala,
Your answer is helpful when I need to configure a new cluster using the Azure console.However, my problem has a different focus: I need to find a storage class with specific properties on a given Kubernetes cluster. I don't have access to a Hyperscaler management console, and I can't influence the configuration of the cluster in any way.
So what can I do to identify and avoid such a configuration where unusable storage classes are preventing pods from starting?
Kind regards, Christian