Nolan Brubaker

Sorry, but set PYTHONHOME where? In buildout.py, recipes, or somewhere else?

Just an update, if 1.5+ is pinned in versions.cfg, then the buildout will work as long as it's boostrapped with version 1.4.4. So, there's a usable workaround for the moment.

@kris-nova > building kubicorn into containers to run as controllers in a pod. I know there's #482 that mentions some of this, but could you expand on this a little...

@jnardiello This should be working as of https://github.com/kris-nova/kubicorn/commit/79d42702e1b58684bb465f91579999473cfb2def (and some subsequent typo fixes). My primary use case was supporting EBS-backed PVs, so I didn't try ELBs in depth.

fwiw, the kubeadm config files are now being written with https://github.com/kris-nova/kubicorn/commit/79d42702e1b58684bb465f91579999473cfb2def. I haven't confirmed that the kubeconfig cluster name matches, though.

Also, deleting the cluster that I manually spun up appears to work just fine. It reports things are fine on the CLI, and the resources are gone on my console....

After finding the StackDriver logs on my GCP console, I noticed I was getting these errors when Kubicorn was running: ``` error: [ 0: { code: "SERVICE_ACCOUNT_ACCESS_DENIED" detail_message: "" location:...

Something like AWS's newer [resource groups](https://docs.aws.amazon.com/ARG/latest/userguide/welcome.html) could take advantage of this.

Yeah, there's some of that; I think having 1 that identifies all resources (in addition to KubicornMaster and KubicornNode) would be helpful. In AWS some are applied ([example](https://github.com/kris-nova/kubicorn/blob/2f67b7aa28ac21d1542c0ee8bbda3103ac3290d5/cloud/amazon/public/resources/routetable.go#L78)), so it...

So I think I was mistaken - I think I can use the `KubernetesClusterName` tag to do this. I did notice, however, that some resources don't have a `tag` function....