Currently for every automation test run - we deploy a new AWS setup and then run the k3s/rke2 install and tests on them.
There should be a way to say, provide IPs and the kubeconfig(only if re-using a pre-installed k3s/rke2) from the test options/tfvars and re-use the same for the next testrun.
We might need to think about, if the setup has a pre-installed k3s/rke2 - would we want to uninstall it first and then re-install k3s/rke2 before continuing the new test as well.
Have a deploy flag(default true) - which will be for deploying a new setup. If False, we provide IPs for the pre-existing setup.
If kubeconfig was provided, we continue with the test without uninstall. if not, we uninstall k3s/rke2 and then continue with the test.
Currently for every automation test run - we deploy a new AWS setup and then run the k3s/rke2 install and tests on them.
There should be a way to say, provide IPs and the kubeconfig(only if re-using a pre-installed k3s/rke2) from the test options/tfvars and re-use the same for the next testrun.
We might need to think about, if the setup has a pre-installed k3s/rke2 - would we want to uninstall it first and then re-install k3s/rke2 before continuing the new test as well.
Have a deploy flag(default true) - which will be for deploying a new setup. If False, we provide IPs for the pre-existing setup. If kubeconfig was provided, we continue with the test without uninstall. if not, we uninstall k3s/rke2 and then continue with the test.
This will save the AWS costs significantly.