Cannot access RStudio part / jupyter + command line is fine

@rrrrrok Christine told me to ask questions here …

I have a project:

… and I’m struggling to get both RStudio and Jupyter (command line) working together. It’s built from a Dockerfile that we built last year (together maybe?) and this may be the problem, as renku has been updated, many components have probably changed and so on. I tried to start from one of the templates (bio311, for example), but that lead me to an alternative wall of pain. The one above actually managed to build.

And, I just don’t know enough about the renku innards to correct the course.

I got the following messages …

  1. I get this below on startup, but can click DISMISS and life goes on:

Failed to load the jupyterlab-git server extensionThe versions of the JupyterLab Git server frontend and backend do not match. The @jupyterlab/git frontend extension has version: 0.11.0 while the python package has version 0.22.3. Please install identical version of jupyterlab-git Python package and the @jupyterlab/git extension. Try running: pip install --upgrade jupyterlab-git

  1. (when I change /lab/ to /rstudio/) RStudio ServerThe requested page was not found.
    Please contact your system administrator for assistance, or click here to go back.

  2. A new version of renku is availabe.
    This project is currently using an older version of renku. A migration is necessary to make changes to datasets and is recommended for all projects.
    You do not have sufficient rights to upgrade this project, but a project maintainer can do this. Contact a maintainer.

Do you have any suggestions to get us up to date?

Best, Mark

Hi @markrobinsonuzh thanks for posting the question here! There have been a few changes since that project was made :slight_smile:

I’m trying a build here:

I took out the R packages for now just to make sure the base + conda works fine.

The build is done and from what I can see it all looks fine - all the R packages compiled, but it’s possible some apt packages are missing. Those can always be added to the Dockerfile, but I think it’s much simpler now than what you started with.

You can try it out by going to and spinning up an environment from the latest commit. Hope that helps!

Thanks a bunch @rrrrrok … I also created, which is expanded from the bio311 template (as that template seems to ignore environment.yaml) … so we should be in very good shape. I’ll pass this over to Kathi who will lead Monday’s session. Thanks again!

Hi @markrobinsonuzh, no worries, glad I could help. Btw, are these builds going to be as heavy as usual? In that case we may want to make sure the build is set up in such a way that the students don’t happen to wait super long for their environments to build on Monday. If you let me know which project you’ll use in the end I can set up the runners to avoid that situation.

Btw, we are just about to merge some work that will make this problem go away for good (it will let you specify a single fixed image for a project). I thought we’d have it ready by the next time you teach a class but it looks like we’ll be a week late :slight_smile:

Hi @rrrrrok … thanks. We actually sent the students both links to fork from …