Rancher 2.12.x Upgrade Fails: RKE1 NodeTemplate Resources Detected

Rancher v2.11.5, RKE2v1.32

After upgrading to Rancher v2.12+, the rancher-pre-upgrade job fails with the error "Rancher v2.12+ does not support RKE1." The logs from this job indicate that RKE1-related resources, such asNodeTemplate, were detected. This prevents the upgrade from completing successfully.

The detected RKE1-related resources(like Node Template) must be either migrated to RKE2/K3s or deleted to resolve the issue. 

Step 1: Verify the RKE1-Related Resources

First, confirm that the rancher-pre-upgrade job is indeed detecting NodeTemplate resources by checking its logs.

kubectl logs rancher-pre-upgrade-x4p64 -n cattle-system

The log output will confirm the presence of the NodeTemplate, as seen in the example below:

...
Found 1 NodeTemplate resource(s):
NAMESPACE          NAME      DISPLAY NAME
----------------  --------  --------------
cattle-global-nt  nt-6t9wh  nutanix-ubuntu

===== SUMMARY =====
NodeTemplate: 1
Total resources detected: 1
Error: Rancher v2.12+ does not support RKE1.
Detected RKE1-related resources (listed above).
Please migrate these clusters to RKE2 or K3s, or delete the related resources.
More info: https://www.suse.com/c/rke-end-of-life-by-july-2025-replatform-to-rke2-or-k3s

Step 2: Delete the NodeTemplate Resource

Use kubectl to delete the detected NodeTemplate resource. You will need to use the name of the template from the logs.

kubectl delete nodetemplate nt-6t9wh -n cattle-global-nt

Note: The name and namespace of your NodeTemplate may differ from the example. Ensure you use the exact NAME (e.g., nt-6t9wh) and NAMESPACE (cattle-global-nt) from your logs.

Step 3: Re-run the Pre-Upgrade Job by upgrading the Rancher.

After deleting the NodeTemplate, re-run the rancher-pre-upgrade job to ensure all RKE1-related resources are gone, and the check now pass

# Delete the old failed job first
kubectl delete job rancher-pre-upgrade -n cattle-system


# Scaledown Rancher deployment pods to zero
 kubectl -n cattle-system scale deployment rancher --replicas=0


# Re-run the upgrade command
# (e.g., your helm upgrade command for Rancher)
helm upgrade --install rancher rancher-stable/rancher \
    --namespace cattle-system \
    --set hostname=<your-hostname> \
    # ... other settings


# once it's successful, scale up Rancher pods 
 kubectl -n cattle-system scale deployment rancher --replicas=3

The new rancher-pre-upgrade job will now complete successfully, and the Rancher v2.12+ upgrade can proceed

Before the upgrade, Nutanix drivers are enabled, and a Node Template is created using Nutanix drivers. Rancher v2.12 and later versions have officially ended support for RKE1 clusters. During the upgrade process, a check is performed for any remaining RKE1-related resources, such as NodeTemplate or  ClusterTemplate, to ensure no legacy RKE1 clusters are being managed. The presence of these resources causes the upgrade to fail.

when upgrading to v2.12.x there must be no RKE1-related objects left - customers are responsible for cleaning them up, and we're providing instructions on how to locate these objects (script is referenced there).