search

aiidalab / aiidalab-docker-stack

Docker images with the basic software stack for AiiDAlab
https://aiidalab.net
Other
9 stars 14 forks source link

Downsides of using jupyter stack as the base image #471

Open danielhollas opened 3 months ago

danielhollas commented 3 months ago

Hi @danielhollas, I reconsidered if use jupyter stack as the base image is a good option, instead if we move to using aiida-core images could gaining more flexibility. Major two reasons are:

  1. The jupyter/minimal-notebook we use now as you did in this PR, we using python version tag, therefore has no control on the notebook version. If fact the python version is not we really care but mostly the aiida-core version and jupyter backend version.
  2. We cannot set the system username which is minor but still not so good. Plus we can not using s6-overlay which I think manage well for the services when we need more.

Originally posted by @unkcpz in https://github.com/aiidalab/aiidalab-docker-stack/issues/455#issuecomment-2182685961

danielhollas commented 3 months ago

@unkcpz I agree that those are downsides, but, to be blunt, they are imho not enough to be doing yet another refactor of the stack, which is a huge time sink and I think we have better things to do now.

Also, we'd need to manage the Jupyter stack ourselves, and it's not clear how much extra work it would be. The Jupyter stack is I think the most complicated stack we have in the image (just by the number of dependencies etc) so I think it is quite convenient to rely on others to build it for us.

Plus we can not using s6-overlay

I don't think we have enough experience yet with s6-overlay to know if we should be moving in this direction.