Use hardware namespace and name as provider ID

[abhinavmpandey08]

Signed-off-by: Abhinav Pandey abhinavmpandey08@gmail.com

Description

This PR changes the machine.spec.providerID to be of the format tinkerbell://{hardware.namespace}/{hardware.name} instead of using UID because UID can change if a hardware is recreated/moved and machine.spec.providerID is immutable so CAPT runs into errors like this:

E0518 23:43:49.671348       1 controller.go:317] controller/tinkerbellmachine "msg"="Reconciler error" "error"="ensuring machine dependencies: ensuring hardware: patching machine object: admission webhook \"validation.tinkerbellmachine.infrastructure.cluster.x-k8s.io\" denied the request: TinkerbellMachine.infrastructure.cluster.x-k8s.io \"abhinav-control-plane-template-1652916169789-67n5d\" is invalid: spec.providerID: Forbidden: is immutable once set" "name"="abhinav-control-plane-template-1652916169789-67n5d" "namespace"="eksa-system" "reconciler group"="infrastructure.cluster.x-k8s.io" "reconciler kind"="TinkerbellMachine"

This new providerID was suggested here: https://github.com/tinkerbell/cluster-api-provider-tinkerbell/pull/160/files#r869807167

How Has This Been Tested?

Created a cluster with the modified CAPT, performed a cluster move operation and verified that the cluster and machines were Ready and there were no errors in CAPT.

How are existing users impacted? What migration steps/scripts do we need?

No user impact

Checklist:

I have:

• updated the documentation and/or roadmap (if required)
• added unit or e2e tests
• provided instructions on how to upgrade

[micahhausler]

LGTM