If you’ve ever worked with Rancher projects, you know they’re a handy way to group namespaces, manage RBAC, and keep your Kubernetes world a little less chaotic. But what happens if a project or its namespaces vanish?
That’s where CloudCasa comes in. It makes restoring Rancher projects and their workloads surprisingly simple. Let’s break it down into the three main situations you might run into.
Preparation: Back Up Everything
First things first – backup both your local upstream (SUSE management cluster) and downstream cluster(s).
To perform a recovery, you need a backup. Hopefully, you’ve been using CloudCasa to backup and migrate your downstream clusters (you know, the ones running mission-critical applications and data). But users seldom backup the SUSE Rancher local management cluster as well. It’s here where SUSE Rancher keeps all the data it uses to manage those clusters, including projects, user permissions, and memory/CPU thresholds that define a SUSE Rancher managed cluster.
So be sure to back up all clusters including the SUSE local management cluster as part of your disaster recovery and migration plans.
💡Did you know? – SUSE Rancher stores its projects metadata in the SUSE local management cluster but gives the namespaces some cryptic names. Downstream cluster nodes added to the local management cluster with a unique cluster id and project ids will show with a unique project id that can be referenced in the yaml. As an example, look at the screenshot below showing an inquiry of projects, where you can see 4 separate clusters being managed by SUSE Rancher, including the local management cluster (itself). The cluster ID is highlighted in green and will also appear as a separate namespace associated with the project.
You’ll also see project ids as highlighted in blue. Here you’ll see these 3 projects as part of a single cluster. The concatenation of both the cluster-id and the project id create a backing namespace which stores the metadata for all the details of the project with that cluster. So down the road, if you do end up changing the cluster ID or the project ID, to make sure that the definitions for these clusters will match the naming convention. Understanding this structure will simplify your backup and migration strategies. |
Restoration Scenarios
- Oops, I Lost Just the Project (But the Apps Are Still There)
This is a metadata only issue—your project definition is gone, but the application namespaces are still alive and kicking. SUSE Rancher can’t manage them because it doesn’t know they exist. You’ll need to restore the project definition back to the SUSE local management cluster.
Steps to restore:
- Spin up a restore job in CloudCasa and point it at the project namespace.
Browse for projects to restore directly in the CloudCasa Resource Browser
- In the restore settings, make sure you tick Overwrite existing resources.
- Run the job and then refresh Rancher. Your user roles tied to that project should pop back up.
- Reattach the project to its namespace with a quick kubectl command:
kubectl annotate namespace <NAMESPACE> field.cattle.io/projectId=’CLUSTER_ID:PROJECT_ID’ –overwrite
Refresh Rancher again, and you’re back in business.
- The Namespaces Are Gone, but the Project Is Still There
In this case, you’re restoring workloads, not the project itself.
- Create a restore job in CloudCasa from your downstream cluster backup.
- Select all the namespaces that belonged to the project.
Standard restore job, something CloudCasa has done since its creation.
- Run the job.
Once it’s done, Rancher automatically links those namespaces back to the existing project. Easy win.
- Disaster Strikes — Both the Project and Namespaces Are Gone
This is the worst-case scenario, but still manageable. You’ll just combine the first two playbooks:
- Restore the project definition.
- Patch the namespaces if needed.
- Restore the namespaces from your downstream cluster backup.
Boom—you’ve got your Rancher project and workloads back.
Why This Matters
Rancher projects aren’t just a nice-to-have. They’re the glue that holds RBAC, resource quotas, and team organization together. Losing them (or their namespaces) can turn your cluster into the Wild West real quick.
CloudCasa takes the pain out of this by making recovery straightforward—whether it’s a small hiccup (project definition missing) or a full-blown outage (project and namespaces gone).
Conclusion
Whether you’re looking to migrate between versions of Rancher or your Rancher projects go missing, don’t panic. CloudCasa has your back with clear steps to restore everything quickly—without the late-night stress sessions.