sync all the pipelines

[codificat]

Make sure all the cnbi pipelines are consistent in how builds work

Focus for mvp is python/jupyter images

[VannTen]

I'm looking at the other pipelines to see what to do with it, and while looking at the validate-jupyterhub-image, I'm wondering if we might forego it.

From my reading of it, it tries to check if the image will run. Can't we just try to run it and get rid of that embedded shell script ?

(Besides, it seems a bit fragile to me, stuff like this:

        else
          if grep "\$@" /tmp/entrypoint.txt >/dev/null 2>/dev/null; then
            ok
          else
            if grep -- "--port=8080" /tmp/entrypoint.txt >/dev/null 2>/dev/null; then
              ok
            else
              config=$(grep -oP "(?<=--config=)\S*" /tmp/entrypoint.txt 2>/dev/null)
              if [ -n "$config" ]; then
                fail "error" "start-singleuser.sh script has to either use runtime arguments via '\$@' or specify '--port=8080' or it needs to specify a custom config file."

Greping for particular options could easily break or have false negatives.

wdyt @codificat @harshad16

[VannTen]

At the very least if we can't I'll try to see if we can have a proper shell script in git and generate a configMap from the file at deploy time.

[harshad16]

The run of notebook can be an option, though for that to be done, we would have to use the replicate what jupyterhub and kubeflow-notebook-controller pattern to spawn the pod. That is what i think were used on previous pipelines, to check via bash to see if those params exists.