Label worker nodes as worker-cnf when 0 workers available (or if failed)

[e-minguez]

I have a 3 masters + 3 worker nodes environment but after applying the setup-test-cluster.sh script, there were no worker-cnf workers... the issue was the script doesn't validate if there are >0 workers to be labeled. In my case, there are but the oc get request failed:

hack/setup-test-cluster.sh

E0427 16:05:22.121357     299 request.go:924] Unexpected error when reading response body: read tcp 10.131.3.197:34740->10.19.138.14:6443: read: connection reset by peer
error: unexpected error when reading response body. Please retry. Original error: read tcp 10.131.3.197:34740->10.19.138.14:6443: read: connection reset by peer
[INFO]: Labeling 0 worker nodes with worker-cnf

I thought the -e flag would prevent that, but it seems it doesn't

[slintes]

connection reset by peer sounds like a (temporary?) network problem.
Does a simple oc get nodes work?
Is this reproducible?

[e-minguez]

connection reset by peer sounds like a (temporary?) network problem.

It was indeed, but the thing is, the script didn't failed and it ended up labeling 0 nodes. I guess the script should check if >0 workers are going to be labeled, otherwise, fail gracefully.

Does a simple oc get nodes work?

Yes.

Is this reproducible?

I guess not, but the point is if the script should allow this scenario or controlling the output and fail if needed.

Thanks!