[BUG] Downstream v2prov clusters do not accurately reflect configured rewrites and mirrors

[Josh-Diamond]

Rancher Server Setup

• Rancher version: v2.10-f95d1dc120a4b4c602fcb1485b36134fc0f4a2bc-head
• Installation option (Docker install/Helm Chart): Helm

Information about the Cluster

• Kubernetes version: v1.31.2+rke2r1, v1.31.2+k3s1
• Cluster Type (Local/Downstream): Downstream Node driver

User Information

• What is the role of the user logged in? Admin

Describe the bug
When provisioning a downstream v2Prov Node driver cluster w/ mirrors and rewrites configured, the cluster comes up active and successfully provisions, but /etc/rancher/agent/registries.yaml, as well as Rancher UI Edit Config > Registries, does not reflect the expected configurations, set during the initial cluster provisioning

To Reproduce

1. Fresh install of Rancher v2.10-head
2. Provision a downstream RKE2 cluster, v1.31.2+rke2r1, with cluster scoped registry enabled and the following configurations set:
   • Registry hostname: docker.io
   • Mirror endpoints: https://registry.example.com
   • Rewrite Config -> Rewrite pattern: ^rancher/(.*)
   • Rewrite Config -> Rewrite replacement: mirrorproject/rancher-images/$1
3. Verified - cluster successfully provisions
4. Reproduced

Result
/etc/rancher/agent/registries.yaml does NOT accurately reflect configured rewrites and mirrors

{"configs":{},"mirrors":null}

Expected Result
/etc/rancher/agent/registries.yaml accurately reflect configured rewrites and mirrors

Additional context
Only affects v2.10 branch

[Josh-Diamond]

Upon further troubleshooting, UI is not passing these fields to API call when creating the provisioning cluster object

[gaktive]

This worked in 2.9.x

[rak-phillip]

Updating milestone to v2.11, will create a v2.10 backport issue.

[rak-phillip]

/backport v2.10.0