Cannot restore "service" with velero build from the main branch

[ywk253100]

Run restore with velero built from the main branch, get the following errors:

error restoring services/kibishii-workload/etcd-client: Service "etcd-client" is invalid: spec.clusterIPs: Invalid value: []string{"100.70.53.8"}: must be empty when clusterIP is empty
error restoring services/kibishii-workload/etcd0: Service "etcd0" is invalid: spec.clusterIPs: Invalid value: []string{"100.66.73.184"}: must be empty when clusterIP is empty
error restoring services/kibishii-workload/etcd1: Service "etcd1" is invalid: spec.clusterIPs: Invalid value: []string{"100.69.23.44"}: must be empty when clusterIP is empty
error restoring services/kibishii-workload/etcd2: Service "etcd2" is invalid: spec.clusterIPs: Invalid value: []string{"100.64.23.70"}: must be empty when clusterIP is empty

This is caused by that we bumped up the version of library k8s.io/api/core/v1 from v0.19.12 to v0.20.9 in PR #4022.
There is a new property ClusterIPs introduced in v0.20.9 https://github.com/kubernetes/api/blob/v0.20.9/core/v1/types.go#L4098 and the restore action empties the ClusterIP but not for ClusterIPs https://github.com/vmware-tanzu/velero/blob/main/pkg/restore/service_action.go#L56

Vote on this issue!

This is an invitation to the Velero community to vote on issues, you can see the project's top voted issues listed here.
Use the "reaction smiley face" up to the right of this comment to vote.

• 👍 for "I would like to see this bug fixed as soon as possible"
• 👎 for "There are more important bugs to focus on right now"

[kycfeel]

I also noticed this issue too. The services cannot be restored due to:

error restoring services/<MY_NAMESPACE>/<MY_SERVICE>: Service "<MY_SERVICE>" is invalid: spec.clusterIPs: Invalid value: []string{"<MY_CNI_IP>"}: must be empty when clusterIP is empty

this error.