ml-workspace icon indicating copy to clipboard operation
ml-workspace copied to clipboard

Fail to install R and Rstudio

Open jfarland opened this issue 3 years ago • 1 comments

Describe the bug:

Hi there - I'm running the newest container image on an Ubuntu server for a project and would like to also point to some of my legacy code base in R via Rstudio. I followed the steps to get this going as outlined here

Unfortunately, it looks like the Install Tool functionality doesn't fully work for Rstudio Server or Desktop.

The expected behaviour is that a fresh Rstudio install would be available from the Install Tool functionality.

Steps to reproduce the issue:

From the main Jupyter application, I selected Install Tool and found the appropriate selection from the dropdown:

image

Following the instructions, I then copy/pasted into a terminal with the following error.

image

The tool does register in the dropdown.

image

But doesn't function:

image

It also looks like standard R fails with a similar error. Here's where R lives.

image

Here's the error from trying to kick up an R session.

image

Technical details:

  • Workspace version: Not sure where to find this, echo $WORKSPACE_VERSION returns unknown.
  • Docker version: image
  • Host Machine OS (Windows/Linux/Mac): Ubuntu 20.04 (LTS) x64
  • Command used to start the workspace:
docker run -d \
     -v "/${PWD}:/workspace" \
     -p 8080:8080 \
     --name "xxx" \
     --env AUTHENTICATE_VIA_JUPYTER="xxx" \
     --shm-size 2g \
     --restart always \
     dagshub/ml-workspace:latest
  • Browser (Chrome/Firefox/Safari): Microsoft Edge Version 92.0.902.73 (Official build) (x86_64)

Possible Fix:

It looks like this might be related to conflicts in R versioning via conda (some info found here and here

jfarland avatar Aug 18 '21 21:08 jfarland

Hey, @jfarland thanks for reporting this and apologies for the slow response. Would you be open to contributing a fix for this issue?

deanp70 avatar Nov 09 '21 10:11 deanp70