Developer box cluster

[gjcarneiro]

So I created a cluster in my devbox k8s (native, not minikube).

After holidays, I come back to it, and I find that the k8s starts fine by the etcd pods are all in Completed state. The etcd operator logs this over and over:

time="2019-01-07T17:33:06Z" level=warning msg="all etcd pods are dead." cluster-name=example-etcd-cluster pkg=cluster

Even deleting the pods doesn't cause operator to create new pods.

I just want developer box to have a similar deployment method as production. But I also want the etcd cluster to come back to life on its own if I restart the developer desktop PC. In this case, loss of etcd data is of no importance.

Is there way to reconcile this use case? Or is it out of scope?

[hexfusion]

Is there way to reconcile this use case? Or is it out of scope?

@gjcarneiro we are working towards this, but it is currently not possible.

ref: #1323