fix azure disk attachment error on Linux

[andyzhangx]

What type of PR is this?

Uncomment only one, leave it on its own line:

/kind api-change
/kind bug
/kind cleanup
/kind design
/kind documentation
/kind failing-test
/kind feature
/kind flake

What this PR does / why we need it:
There PR is going to fix drawback after PR: #62612
following getDiskController code will fetch vmCache may lead to below error and the vmCache refresh time period would be 1 min, and MountVolume.WaitForAttach may cost 1 min which is too long time

kubernetes/pkg/volume/azure_dd/attacher.go

Line 162 in ba66014

diskController, err := getDiskController(a.plugin.host)

MountVolume.WaitForAttach failed for volume "pvc-12b458f4-c23f-11e8-8d27-46799c22b7c6" : Cannot find Lun for disk kubernetes-dynamic-pvc-12b458f4-c23f-11e8-8d27-46799c22b7c6

Which issue(s) this PR fixes (optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close the issue(s) when PR gets merged):
Fixes #69262

Special notes for your reviewer:
This PR add a new func getDiskLUN to get disk LUN num from a deviceInfo, deviceInfo could be a LUN number or a device path, e.g. /dev/disk/azure/scsi1/lun2.
While on Windows, since the deviceInfo could be a LUN num or a disk num(on windows), cannot tell whether it's a LUN num since both conditions are a number, we just keep the original code logic.

Does this PR introduce a user-facing change?:

NONE

Release note:

fix azure disk attachment error on Linux

/sig azure
/assign @feiskyer @khenidak
cc @brendandburns

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: andyzhangx

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• pkg/volume/azure_dd/OWNERS [andyzhangx]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[feiskyer]

Could we add a V(2) log here with device and lun number? It could help for troubleshoooting issues.

[andyzhangx]

already addressed in #70012

[feiskyer]

Check whether it is a number or not?

[feiskyer]

Is there any possibility of supporting lun number >= 100 in the future?

[andyzhangx]

it's checking under below:

lun, err := strconv.Atoi(diskLUN)

Azure platform only support up to 64 disk number

[andyzhangx]

ping @feiskyer

[feiskyer]

/lgtm

[andyzhangx]

/test pull-kubernetes-verify

[antoineco]

In spite of some warnings during Pod relocation, I can confirm all volumes eventually get re-attached successfully after 2-3 minutes (Kubernetes 1.10, custom build including this patch).

Related Pod events:

  Warning  FailedAttachVolume      4m               attachdetach-controller                        Multi-Attach error for volume "pvc-1234-5678" Volume is already exclusively attached to one node and can't be attached to another
  Warning  FailedAttachVolume      4m (x2 over 4m)  attachdetach-controller                        AttachVolume.Attach failed for volume "pvc-1234-5678" : Attach volume "k8s-mstr-myclustersub2-d1-kub--pvc-1234-5678" to instance "/subscriptions/abcd-efgh/resourceGroups/myclustersub2-d1/providers/Microsoft.Compute/virtualMachineScaleSets/k8s-mycluster-33570458-vmss/virtualMachines/7" failed with compute.VirtualMachineScaleSetVMsClient#Update: Failure sending request: StatusCode=409 -- Original Error: failed request: autorest/azure: Service returned an error. Status=<nil> Code="ConflictingUserInput" Message="Disk '/subscriptions/abcd-efgh/resourceGroups/myclustersub2-d1/providers/Microsoft.Compute/disks/k8s-mstr-myclustersub2-d1-kub--pvc-1234-5678' cannot be attached as the disk is already owned by VM '/subscriptions/abcd-efgh/resourceGroups/myclustersub2-d1/providers/Microsoft.Compute/virtualMachineScaleSets/k8s-mycluster-33570458-vmss/virtualMachines/k8s-mycluster-33570458-vmss_9'."
  Warning  FailedMount             3m (x7 over 4m)  kubelet, k8s-mycluster-33570458-vmss000007  MountVolume.WaitForAttach failed for volume "pvc-1234-5678" : azureDisk - WaitForAttach failed within timeout node (k8s-mycluster-33570458-vmss000007) diskId:(k8s-mstr-myclustersub2-d1-kub--pvc-1234-5678) lun:(1)
  Normal   SuccessfulAttachVolume  13m (x3 over 15m)  attachdetach-controller                        AttachVolume.Attach succeeded for volume "pvc-1d577a03-d28c-11e8-9585-000d3ad0b3ea"