Cluster provisioning fails when BootVolume Name is specified

[anshuman-agarwala]

/kind bug
/area provider/ibmcloud

What steps did you take and what happened:
[A clear and concise description of what the bug is.]
Added the following to the VPCMachineTemplate spec and tried to provision a cluster with multiple worker nodes:

bootVolume:
        name: boot-vol
        sizeGiB: 20

The control-plane and one worker node came up successfully but the other one is failing with the following error in the controller logs:

"Reconciler error" err="failed to reconcile VSI for IBMVPCMachine default/anshuman-test-cluster-md-0-579cn-42htf: Volume with name \"boot-vol\" already exists" control
ler="ibmvpcmachine" controllerGroup="infrastructure.cluster.x-k8s.io" controllerKind="IBMVPCMachine" IBMVPCMachine="default/anshuman-test-cluster-md-0-579cn-42htf" namespace="default" name="anshuman-test-cluster-md-0
-579cn-42htf" reconcileID="ae4f44c7-12e6-4160-ac54-cec7b11b583d"

What did you expect to happen:
Cluster gets provisioned

Anything else you would like to add:
[Miscellaneous information that will assist in solving the issue.]

Environment:

• Cluster-api version:
• Minikube/KIND version:
• Kubernetes version: (use kubectl version):
• OS (e.g. from /etc/os-release):

[Karthik-K-N]

Hey, Can you help me to understand the volume name constraint. Is it supposed to be unique? Also the VPCMachineTemplate you mentioned is of control-plane or worker machine template.

[Prajyot-Parab]

/assign @anshuman-agarwala

[anshuman-agarwala]

Yes, IBMCloud expects volume names to be unique in a workspace. I saw this issue when I added the name to worker machines but I assume that the same issue will pop up for control-plane machines as well.

[Karthik-K-N]

Yes, IBMCloud expects volume names to be unique in a workspace. I saw this issue when I added the name to worker machines but I assume that the same issue will pop up for control-plane machines as well.

Got it, Thanks.