Unsolved
2 Posts
0
826
July 14th, 2020 00:00
[CSI][VxFlexOS] Unable to use private mount
Hi Team,
We have successfully deployed the CSI (v1.1.4) on Kubernetes (v1.14.9) for VxFlexOs, and its working fine, however, off late, we found an issue. When ever, the pod that uses persistent volume dies (or is killed) and it moves back to one of the nodes, that it used to be there earlier, we get the message
"Unable to use private mount"
I have looked through the source code, and it seems that it is hitting here https://github.com/dell/csi-vxflexos/blob/v1.1.4/service/mount.go#L227
From my testing, I found that when it moves to a server that has never hosted the pod, all is well, the volume gets remapped between the different nodes, and the mount point is created and Kubernetes mounts the pod easily. But when it comes to a server that had previously run this pod, then the volume mounting to the node happens, but the mount to the pod does not happen with the error
kubelet: E0714 02:26:49.338206 14176 nestedpendingoperations.go:267] Operation for "\"kubernetes.io/csi/csi-vxflexos.dellemc.com^d1ffd2bc0000001e\"" failed. No retries permitted until 2020-07-14 02:28:51.338182008 -0400 EDT m=+1177.001841159 (durationBeforeRetry 2m2s). Error: "MountVolume.SetUp failed for volume \"k8s-4570dfd0b3\" (UniqueName: \"kubernetes.io/csi/csi-vxflexos.dellemc.com^d1ffd2bc0000001e\") pod \"master-2\" (UID: \"a1406d2c-c597-11ea-a314-005056a827f4\") : rpc error: code = Internal desc = Unable to use private mount point"
I even tried to see the mounts and unmounted using the umount command, and its still gave the same error.
So here are my questions
1. What is a good way to clean up the private mount points (Where is it located)
2. Is this fixed in 1.1.5 ?
Thanks,
Alok