Kimmo Lehto
Kimmo Lehto
Still sounds like a good idea, stalebot.
Now it should only display `[default-master-0] SSL_connect returned=1 errno=0 state=error: certificate verify failed (certificate is not yet valid)`
I wonder what's wrong now, why does the `kubectl get nodes` after `up` fail? Is the proxy stuff actually working now and the command doesn't work because of that?
So, `Configurer` should have some way to `register_config` with some kind of version matcher.
`Pharos::Config.new.to_yaml` currently generates: ```yaml --- network: provider: weave service_cidr: 10.96.0.0/12 pod_network_cidr: 10.32.0.0/12 node_local_dns_cache: true firewalld: enabled: false open_ports: - port: '22' protocol: tcp roles: - "*" - port: '80' protocol:...
How about generating a short lived [bootstrap token](https://kubernetes.io/docs/reference/access-authn-authz/bootstrap-tokens/) for each run?
(or locate the long outputs and shorten/split them so they're nice on a "standard" 80 column display.)
I don't know how this is supposed to work anymore.
If we want to update the `~/.kube/config` on master, then I guess the service account should be made during `pharos up` instead of `pharos kubeconfig`? Should it be in the...
Moved to `ConfigureServiceAccount` phase. Still needs some specs, though most of it would be `expect(transport).to receive(:exec!).with('xyz')`.