VMSS + IPv6 backend address pool deletion failure

[lzhecheng]

What happened:

When ensureBackendPoolDeletedFromVMSS, the IPv6 backendAddressPool is not deleted, which leads to "child resources are in use by virtual machine scale set" error. An example:
https://storage.googleapis.com/kubernetes-jenkins/pr-logs/pull/kubernetes-sigs_cluster-api-provider-azure/3398/pull-cluster-api-provider-azure-e2e/1643851112475791360/artifacts/clusters/capz-e2e-17k4mh-ipv6/kube-system/cloud-controller-manager-5fcf8dbddb-qf57j/cloud-controller-manager.log

What you expected to happen:

IPv6 backendAddressPool is deleted from VMSS.

How to reproduce it (as minimally and precisely as possible):

Test on a IPv6 + VMSS cluster.

Anything else we need to know?:

Environment:

• Kubernetes version (use kubectl version):
• Cloud provider or hardware configuration:
• OS (e.g: cat /etc/os-release):
• Kernel (e.g. uname -a):
• Install tools:
• Network plugin and version (if this is a network-related bug):
• Others:

[lzhecheng]

Current investigation shows that in an IPv6 cluster, the primary ipConfigurations is still IPv4, so the deletion fails.

cloud-provider-azure/pkg/provider/azure_vmss.go

Line 1592 in 4554e85

primaryIPConfig, err := getPrimaryIPConfigFromVMSSNetworkConfig(primaryNIC)

[lzhecheng]

Related: kubernetes-sigs/cluster-api-provider-azure#3361
/cc @CecileRobertMichon For a IPv6 cluster, VMSS primary NIC's primary IP config is always IPv4, right?

[CecileRobertMichon]

For a IPv6 cluster, VMSS primary NIC's primary IP config is always IPv4, right?

correct

[lzhecheng]

/close

[k8s-ci-robot]

@lzhecheng: Closing this issue.

In response to this:

/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.