Destroying a Cluster

Destroying (tearing down) a Cluster, its network and Bastion is a simple process, handled by the orchestrator.

Tearing down a cluster relies on terraform state files that were generated when the cluster was created. These are located in the orchestrator’s terraform directory. If you’ve lost these files or are running the orchestrator form another host you will not be able to destroy the cluster.

If you’re not in the orchestrator container then, from the working copy of the okd-orchestrator clone that you used to create your cluster, start and enter the orchestrator:

$ ./okdo-start.sh

To destroy the cluster you just need to run destroy.py. If there’s more than one deployment configuration then you will need to name the deployment that defined the cluster you wish to delete. For the example AWS cluster this would require the following command:

$ ./destroy.py compact-aws-frankfurt-3-11

+-----------------------------------------------+
| Compact AWS Deployment (OKD 3.11) [Frankfurt] |
+-----------------------------------------------+

CAUTION You are about to destroy the cluster.
------- Are you sure you want to do this?

Enter "yes" to DESTROY this deployment:

Acknowledge the warning to destroy the cluster.

Destruction is generally very fast. destroy.py, after seeking confirmation, simply relies on the underlying Terraform utility, which rapidly deletes all of the objects that were originally created.