| Table of Contents |
|---|
R version updates August 2023
After the August 15, 2023 maintenance the default R version is R 4.3.1 both for RStudio Server on all compute servers, and on the command line (typing R or Rscript on the command line will use R 4.3.1). The RStudio Server application version itself was also upgraded.
To make sure your RStudio Server session picks up the version change, do both of the following:
- Start a new session (red/orange "power on" button
at the top right - Logout then log back in to the RStudio web application (arrow on paper icon
at the top right)
| Warning |
|---|
The R version upgrade may cause issues if you have installed many R packages locally. This can happen because your local package version is no longer compatible with the latest version of some (e.g. Bioconductor) program we have installed, but since your local installs take precedence they shadow the more up to date globally installed version. See Local/Global package installation conflicts below for more information. |
R and RStudio Server R versions
The issue of R versions is a difficult one, especially now that many important single-cell packages are only available in newer R versions, but not all older , (but still popular) R packages are. This section describes the versioning issues in both the system R and in the RStudio Server web application.
The "default" system version of R on Ubuntu 18.04 compute servers is R 3.6.1 , and the default for Ubuntu 20.04 is R 3.6.3 – this is the version that is invoked if you type R from the command line. (To see what OS version your compute server is running, type lsb_release -a on the command line.)After the August 2023 maintenance the default R version is R 4.3.1 both for RStudio Server on all compute servers, and on the command line (typing R or Rscript on the command line will use R 4.3.1).
We also have other versions of R installed "side by side" on the command line – R-4.0.3 (and others in the future)and R-3.6.1 – which can be accessed by typing the specific version from the command line (e.g. R-4.0.3). , Rscript-3.6.1) However multiple R versions are not available in RStudio Server because its R version setting can only be set to one value system-wide and cannot be specified per-user.
Most All POD compute servers now use R 4.3.6 1 as the R version in their RStudio Server web application but some use R 4.0.3 – see RStudio Server and Python JupyterHub web applications for details. If your POD has multiple compute servers and you would like one to run a different R version, please contact us at rctf-support@utexas.edu.
We have also installed many popular add-on packages in all the R versions (e.g. tidyverse, ggplot2, DESeq2); however be aware that not all packages are available in all R versions.
Other R versions
If you need a GUI environment to access versions of R other than 3.6 or 4.03.3, 1 an option that provides maximum per-user flexibility is as follows. Use the RStudio Server web application for R 3.6- or R 4.03.1-compatible workflows. For workflows requiring other R versions, users can install a different version of R on their own desktop/laptop computers along with the RStudio desktop application. Then users can access files on shared storage by mounting their Work area file system via Samba (see Samba remote file system access for more information). The main drawback to this workflow is that typical personal computers do not have as much RAM as POD compute servers, and some R tasks can be memory intensive. What users can do in such cases is test the code in RStudio on their desktop computer, using smaller data sets if necessary. Then run the "full" workflow from the POD compute server command line using the appropriate R version (you can locally install newer R versions yourself).
Understanding R add-on packages
...
To see what packages are installed system-wide for a given R version, users can look at the version's package installation directories:
- R 3 3.4.4 (Ubuntu 18.04 only)6.3
- /usr/lib/R/library
- /usr/lib/R/site-library
- R 34.6.1 (Ubuntu 18.04 only) 0.3 – /stor/system/opt/R/R-3.6.1/lib/R/libraryR 3.6.3 (Ubuntu 20.04 only)/usr4.0.3/lib/R/library/usr/lib
- /R /site-librarR 4.03.31 – /stor/system/opt/R/R-4.03.31/lib/R/library
Local/Global package installation conflicts
...
If problems persist, please email our rctf-support@utexas.edu support email, including a description of what error or issue you are experiencing.
EDU pod connection issue
An issue on the EDU pod may arise when students use the edupod.cns.utexas.edu virtual host to access RStudio server. The edupod.cns.utexas.edu virtual host acts as a front end load balancer for the request, and forwards it to one of the back-end compute servers. If this occurs, try accessing individual servers specifically:
...
If accessing a specific server works when using the virtual host does not, please let us know by emailing our rctf-support@utexas.edu support email.
Zombie processes
A "zombie" RStudio process may be preventing you from successfully logging in to an RStudio session. If your POD has more than one compute server, try logging in to one of the others. If that works, the issue is related to something on the compute server you can't login to.
You can get rid of any zombie RStudio processes by using ps -ef | grep <user_name> to see processes associated with your account, then kill any RStudio processes using kill -9 <pid>, where <pid> is the 1st number after your account name in the ps list. E.g.:
| Code Block | ||
|---|---|---|
| ||
ps -ef | grep abattenh |
Can report information like this:
| Code Block |
|---|
# output from ps -ef | grep abattenh:
abattenh 65327 5968 0 Aug01 ? 00:00:49 /usr/lib/rstudio-server/bin/rsession -u abattenh --session-use-secure-cookies 0 --session-root-path / --session-same-site 0 --session-use-file-storage 1 --launcher-token F62AEC5D --r-restore-workspace 2 --r-run-rprofile 2 |
Here the pid is 65327, and the process can be killed like this
| Code Block | ||
|---|---|---|
| ||
kill -9 65327 |
File upload issues
Sometimes users will get an "Unexpected response from server" error when attempting to upload a file. Here are some troubleshooting tips:
- Try the Browser issues troubleshooting tips, especially:
- Make sure your browser is up to date with the most current version
- Try the operation in at least 2 different up-to-date browsers (e.g. Chrome and Firefox)
- Try the operation in a private/incognito/safe browser session
- Try the upload operation from the JupyterHub web application instead of from RStudio
- Try connecting to a specific back-end compute server as described at EDU pod connection issue
Disk quota exceeded
Another type of problem can arise when a user's 100 GB Home directory quota has been exceeded (not applicable on the EDU pod, which does not have Home directory quotas). This can produce errors when trying to start RStudio Server or R, perform work in R, or even install additional packages. For example, you may see a "Cannot connect to service" message after logging in to RStudio Server. Or, if an R session has been established and saving a new file would exceed the Home directory quota, users will often (but not always) see an error like the following:
...