Fix #329

[qsdrqs]

This PR tries to fix #329. After a basic inspection of the codebase, I find the current Acto does have some retry mechanisms. However, because of some subprocess.run for command execution have an argument of check=True, according to the document:

If check is true, and the process exits with a non-zero exit code, a CalledProcessError exception will be raised. Attributes of that exception hold the arguments, the exit code, and stdout and stderr if they were captured.

If check is set to True and the command failed, a CalledProcessError will be raised and the following returncode logic becomes meaningless. So currently, I only correct some check arguments to make the current retry mechanisms meaningful.

Besides, I also mentioned that although create_cluster logic for all 3 tools(kind, minikube, k3s) will retry for 3 times before if fails, the delete_cluster logic will forever try to delete if the command fails:

def delete_cluster(self, name: str, kubeconfig: str):
    ...

    while subprocess.run(cmd, check=False).returncode != 0:
        continue

Is this an expected design? Do we need to change this into the same as create_cluster logic (retry for 3 times and fail)?