[backport v2.13] DownstreamResources don't work for the local cluster

[rancherbot]

This is a backport issue for #4559, automatically created via GitHub Actions workflow initiated by @0xavi0

Original issue body:

There is a problem when copying Downstream Resources into the local cluster, because DeepCopy() is being used, which also copies the metadata (including the UID).

When calling CreateOrUpdate, the function does not return any error, but internally Kubernetes is not performing the copy because a resource with the same ID already exists.

[sbulage]

System Information

Rancher Version	Fleet Version
v2.13-4795da16e8c26817e1d10d38a7538094ee92e382-head	108.0.3+up0.14.3-beta.1

Verification steps:

• Enabled EXPERIMENTAL_COPY_RESOURCES_DOWNSTREAM
  • Enable Experimental feature

    extraEnv:
      - name: EXPERIMENTAL_COPY_RESOURCES_DOWNSTREAM
        value: "true"
    

---

• Create ConfigMap using below command:
  • ConfigMap data file values.yaml

    name: testing-downstream-cluster-configmap

    kubectl create configmap -n fleet-local config-values --from-file=values.yaml=values.yaml

---

• Below HelmOps YAML used
  • HelmOps YAML

    apiVersion: fleet.cattle.io/v1alpha1
    kind: HelmOp
    metadata:
      name: test-values-from
      namespace: fleet-local
    spec:
      correctDrift: {}
      downstreamResources:
        - kind: configmap
          name: config-values
      helm:
        chart: >-
          https://github.com/rancher/fleet/raw/refs/heads/main/integrationtests/cli/assets/helmrepository/config-chart-0.1.0.tgz
        valuesFrom:
          - configMapKeyRef:
              key: values.yaml
              name: config-values
      pollingInterval: 5s
      targets:
        - clusterSelector:
            matchExpressions:
              - key: provider.cattle.io
                operator: NotIn
                values:
                  - harvester

---

Verification Screenshots

Rancher 2.13.2	Rancher 2.13-head (Fixed)

Copied resources to Local cluster