Difference between revisions of "Rstudio"
Jump to navigation
Jump to search
Line 1: | Line 1: | ||
Rstudio Server is installed on the GPU cluster and is available at [http://hpc.radiology.hku.hk/rstudio hpc.radiology.hku.hk/rstudio]. As the software doesn't offer an account creation page, prospective users need to request an account from [[Admin]]. | Rstudio Server is installed on the GPU cluster and is available at [http://hpc.radiology.hku.hk/rstudio hpc.radiology.hku.hk/rstudio]. As the software doesn't offer an account creation page, prospective users need to request an account from [[Admin]]. | ||
− | [[File:Rstudio browser.png|thumb|The browser version looks similar to the desktop version of the Rstudio IDE]] | + | [[File:Rstudio browser.png|thumb|The browser version looks similar to the desktop version of the Rstudio IDE. With a parallel package you can use most of he cores on the `cpu1` node]] |
== Features == | == Features == |
Latest revision as of 12:58, 14 September 2020
Rstudio Server is installed on the GPU cluster and is available at hpc.radiology.hku.hk/rstudio. As the software doesn't offer an account creation page, prospective users need to request an account from Admin.
Features
- Rstudio has a built-in bash terminal
- Git is available
- Users can install R packages individually i.e. packages are not shared system-wide with other users.
- The Rstudio interface allows uploading and downloading files.
Rstudio runs in a new sandboxed environment, so don't be alarmed if you see an empty home folder
This is the free version (not professional). Users can't run multiple sessions in multiple tabs. It also lacks advanced monitoring and login options
Administration details
- We are running a recent docker image from the Rocker project.
- It is run by 3 prolific data scientists. Hosts multiple images, including some with cuda libraries.
- The container is on cpu1, it is behind our reverse proxy.
- Rstudio Server has a login system based on Linux users with home folders
- Respinning the container wipes user accounts but preserves their home data in a named volume
rstudiovol
.- So just recreate the user and they can access their packages like before.