Research R, Python, Anaconda Update/Optimization

[bryanpaget]

The following issues StatCan/aaw#1784, #485 and #492 revealed a fundamental problem with the way Anaconda is installed in our images. We need to rethink the Anaconda installation.

Possible course of action:

• find any and all settings for Anaconda and set them to the version we want.
• find and move any and all Anaconda related installation and configuration steps to an Anaconda docker-bit shared by all images (Anaconda is currently being installed in both ∞_CMD.Dockerfile and ∞_CMD_remote-desktop.Dockerfile).
• I don't think we need to install rstudio-server. At least on the remote-desktop image we already have rstudio-server installed from the base image. See: Research R, Python, Anaconda Update/Optimization #496 (comment)
• Some images should be upgraded to "2023-08-07".

[bryanpaget]

Notes for Future Work:

• We have the tag for jupyter/datascience-notebook specific three times in the docker-bits. Next sprint let's refactor some more to have it listed only once. The files are: 0_cpu, 0_cpu_sas and the Makefile.

[bryanpaget]

This has proven to be quite difficult. Earlier in the sprint I tried refactoring the docker-bits files, with the goal of separating and isolating the Anaconda installation to its own Anaconda.dockerfile. This has not worked out.

Not working:

• https://github.com/StatCan/aaw-kubeflow-containers/actions/runs/5868964608

Latest attempt:

• https://github.com/StatCan/aaw-kubeflow-containers/actions/runs/5882103018

[bryanpaget]

Blocked by #502

[bryanpaget]

I don't think we need to install rstudio-server:

I made the following update recently to the rstudio-server version

but discovered the new version was not being used when I opened up rstudio-desktop

as you can see the rstudio-server version is stuck at 2023.06.0-421. That version comes from

which pins rstudio-server at 2023.06.0-421

We may be able to remove the rstudio-server installation step, saving time on builds. The following steps could be omitted from docker-bits/6_rstudio.Dockerfile

# install rstudio-server
ARG RSTUDIO_VERSION=2023.06.1-524
ARG SHA256=208897f00b580b45c37dbb787dc482ff6d774b32c230f49578691b1a11f40c37
RUN apt-get update && \
    apt install -y --no-install-recommends software-properties-common dirmngr gdebi-core && \
    wget -qO- https://cloud.r-project.org/bin/linux/ubuntu/marutter_pubkey.asc | sudo tee -a /etc/apt/trusted.gpg.d/cran_ubuntu_key.asc && \
    add-apt-repository "deb https://cloud.r-project.org/bin/linux/ubuntu jammy-cran40/" && \
    apt install -y --no-install-recommends r-base r-base-core r-recommended r-base-dev && \
    apt-get update && apt-get -y dist-upgrade

[bryanpaget]

We had a discussion during the August 31 2023 tech elab, which resulted in the following EPIC: #1822.