Closed wilsongis closed 3 years ago
@wilsongis Thanks for your interest! Unfortunately, I have removed the R-flavor in the latest version (0.13). The reason is that the newer versions of RStudio Server have some unresolved problems, making it complicated to update the R-Tooling. However, if there is interest, I might bring it back and try to figure out a way to deal with the RStudio Problem. It would help for deciding the next steps if you could answer the following questions:
Btw. you can find the r-flavor resources of the 0.12.1 version here: https://github.com/ml-tooling/ml-workspace/tree/v0.12.1/r-flavor
I appreciate the reply. In looking at my issue, I realized that I can just use the vanilla version. I just needed to install R so that I could use it in a Jupyter Notebook. Again, I appreciate your work.
Thanks, Mike
From: Lukas Masuch @.> Date: Tuesday, July 6, 2021 at 8:41 AM To: ml-tooling/ml-workspace @.> Cc: Mike Wilson @.>, Mention @.> Subject: Re: [ml-tooling/ml-workspace] Update for the R flavor (#96)
@wilsongishttps://github.com/wilsongis Thanks for your interest! Unfortunately, I have removed the R-flavor in the latest version (0.13). The reason is that the newer versions of RStudio Server have some unresolved problems, making it complicated to update the R-Tooling. However, if there is interest, I might bring it back and try to figure out a way to deal with the RStudio Problem. It would help for deciding the next steps if you could answer the following questions:
Btw. you can find the r-flavor resources of the 0.12.1 version here: https://github.com/ml-tooling/ml-workspace/tree/v0.12.1/r-flavor
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/ml-tooling/ml-workspace/issues/96#issuecomment-874770824, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AARMKMNQ6RSYE4VU42C6T4TTWMBZ5ANCNFSM474SLSYQ.
Since I need CUDA, R and R-StudioServer on my image I used the 0.12.1 CUDA-version (mltooling/ml-workspace-gpu
) with an selfmade script to install the R-dependencies which always felt kinda whaky. As @LukasMasuch pointed out I can't seem to make the same script working under the new 0.13.2.
So in conclusion: I have interest in an ml-learning image with R/R-Studio and CUDA and I really love your work here. Your ml-tool is a really great idea and I hope you keep improving it :-) I use it quite extensively.
For reference here is the script:
#Set all rights to everyone on /tmp, otherwise the apt-get update runs into trouble:
chmod 1777 /tmp
# Fix bad permissions on the /usr folder
#chown -R root:root /usr # bad since netdata does not work
find /usr -uid 4011 -exec chown root:root {} +
#Add certs:
sudo update-ca-certificates
#Add the key and repo for r-base:
apt-key adv --keyserver keyserver.ubuntu.com --recv-keys E298A3A825C0D65DFD57CBB651716619E084DAB9
add-apt-repository 'deb https://cloud.r-project.org/bin/linux/ubuntu bionic-cran40/'
#Add the key for the MS SQL-Server driver:
curl https://packages.microsoft.com/keys/microsoft.asc | apt-key add -
curl https://packages.microsoft.com/config/ubuntu/18.04/prod.list > /etc/apt/sources.list.d/mssql-release.list
#Install r,
# r-devel (to be able to install source pakages),
# a curl dependency,
# a DB driver dependency,
# a R-Studio-Server dependency:
apt-get update
apt-get install \
r-base=4.0.5-1.1804.0 \
r-base-dev=4.0.5-1.1804.0 \
r-recommended=4.0.5-1.1804.0 \
r-base-core=4.0.5-1.1804.0 \
libcurl4-openssl-dev \
unixodbc-dev \
libclang-dev \
#Install MS SQL-Server driver:
ACCEPT_EULA=Y apt-get install msodbcsql17 mssql-tools
echo 'export PATH="$PATH:/opt/mssql-tools/bin"' >> ~/.bash_profile
echo 'export PATH="$PATH:/opt/mssql-tools/bin"' >> ~/.bashrc
#Download and install RStudio-Desktop:
#wget -nc --no-check-certificate https://download1.rstudio.org/desktop/bionic/amd64/rstudio-1.4.1103-amd64.deb
#gdebi -n rstudio-1.4.1103-amd64.deb
#rm rstudio-1.4.1103-amd64.deb
#Download and install R studio server:
# Other possible URLs, be carefull 1.3 and 1.4 seems to be buggy:
#https://download2.rstudio.org/server/bionic/amd64/rstudio-server-1.4.1717-amd64.deb
wget -nc --no-check-certificate https://download2.rstudio.org/server/bionic/amd64/rstudio-server-1.2.5042-amd64.deb
gdebi -n rstudio-server-1.2.5042-amd64.deb
rm rstudio-server-1.2.5042-amd64.deb
#Add a rstudio user to run the server application:
useradd -m -d /home/rstudio rstudio
usermod -aG sudo rstudio
echo "rstudio ALL=(ALL) NOPASSWD:ALL" >> /etc/sudoers
# Setup R-Studio-Server config:
printf 'rsession-ld-library-path="/usr/local/lib/R/site-library/"' > /etc/rstudio/rserver.conf
printf "session-default-working-dir="$WORKSPACE_HOME"\nsession-default-new-project-dir="$WORKSPACE_HOME > /etc/rstudio/rsession.conf
printf "setwd ('"$WORKSPACE_HOME"')" > /home/rstudio/.Rprofile
# Create an new entry in the tools list:
echo '{"id": "rstudio-link", "name": "RStudio", "url_path": "/tools/8071/", "description": "Development environment for R"}' > $HOME/.workspace/tools/rstudio-server.json
# Download the supervisor config script to restart the R-Studio-Server after every reboot:
wget -nc --no-check-certificate https://raw.githubusercontent.com/ml-tooling/ml-workspace/main/r-flavor/resources/rstudio-service.conf -O /etc/supervisor/conf.d/rstudio-service.conf
# Start the R-Studio-Server:
LD_LIBRARY_PATH="" LD_PRELOAD="" USER=rstudio /usr/lib/rstudio-server/bin/rserver --server-working-dir=$WORKSPACE_HOME --server-daemonize=0 --auth-none=1 --auth-validate-users=0 --www-port 8071 &
As it turns out the https://raw.githubusercontent.com/ml-tooling/ml-workspace/main/r-flavor/resources/rstudio-service.conf
file is 404, i need to find a replacement for it.
This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 14 days
Feature description:
I am interested in updating the R version of the container. It does not look like R is a branch so I am not sure what I need to do to make the update happen.
Problem and motivation:
I am using the container extensively and I would like an update to the R version
Is this something you're interested in working on?
Yes