Feedback on JRC integration

[diocas]

This is a thread to discuss any changes that might be required to our Spawner to allow the JRC integration.

This discussion fit in the discussion about the binder integration, since the changes required to make binder work should be enough to make this integration also work.

[lucamarletta]

Here some changes we did in kubespawner.py to make it work in our environment:

In all volumes for some reasons not clear if we kept the check on 'type' = 'Directory' the volume failed to be mounted

c.SwanSpawner.volumes = [
{
  'name': 'cvmfs',
  'hostPath': {
  'path': '/cvmfs',
  'type': '', # @jrc it was 'Directory'
}
},

In order to make it work in a subdirectory (we cannot manage a new subdomain) we added this
c.JupyterHub.base_url = '/svc/swan/'

So in our network we reach swan in a similar url
https://jeodpp.jrc.eu/svc/swan/