Here's something I just thought of: you need backup.pem from SDSC for playbooks/ndslabs-k8s-install.yaml.. That's sort of a deal-breaker if we're attempting to package this thing up for others to use.
Granted, the deployment process works 100% up until that point, but it will certainly confuse our users if they get a big red failure at the end.
In light of this, we may need to rethink how our backup is configured and parameterize the backup configuration (i.e. give it a path to a .pem, user, and target IP for SSH-ing the backups).
This ticket is complete when we have discussed the above and done one or more of the following:
- parameterize backup configuration in deploy-tools / cluster-backup repos
- implemented a suitable workaround that prevents the playbook from failing