Fixes to R Library Issues

Due to recent changes to the computing environment around the move to Rocky Linux 8, MSI has been having difficulty providing a stable set of R modules for use on our systems. This has resulted in a poor user experience for anyone wanting to use R on the Agate cluster, particularly those users who rely on installing complex or numerous libraries as part of their workflows.

Additionally, our attempts to address the issues with the existing R modules by patching them in-place have caused confusion and frustration by creating a 'moving target' for users trying to get their workflows up and running on Rocky.

MSI News

This created two distinct issues. The first issue impacts users who were unable to install necessary packages via R after the upgrade to Rocky, due to an inconsistency in the toolchain used to build the modules. The second issue impacts users who had installed some packages via R after the upgrade to Rocky, which then broke after the in-place patch created an inconsistency in the environments provided by the R modules. In both cases, impacted users were prevented from installing or loading certain libraries in R.

These issues impacted the following modules to varying degrees:

  • R/4.2.0-rocky8
  • R/4.2.0-openblas-rocky8
  • R/4.3.0-openblas-rocky8
  • R/4.4.0-openblas-rocky8

In an effort to clear up the situation, MSI has prepared a separate set of R modules that addresses both the original toolchain issue as well as removes the environment inconsistency introduced during patching.

Rather than attempt to replace the existing modules in-place again, we are providing the fixed modules alongside the existing modules to give everyone the chance to move their workflows over. The following table lists the current modules and their fixed equivalents. 

If you have already worked with MSI on R issues and were recommended the temporary "R/testing.donotuseyet" module, this module is superseded by "R/4.4.0-openblas-rocky8-fix".

Current Module

Fixed Module













The current and fixed modules will be available alongside each other under the current naming convention until the start of the June maintenance window on June 5, 2024, at which point the current modules will be removed, and the fixed versions will take their place. 

Transitioning to the fixed modules will likely involve reinstalling your R packages using that module, as packages installed using the current modules may have inherited issues from inconsistencies in the toolchain or environment that impacted the modules themselves.

Alongside the fixed modules, we are rolling out a change to the RStudio application offered via Open OnDemand. Previously, the launched application ignored most changes to the environment from both user configuration files and from the "Custom Environment" field in the resource request application. The updated version of the application will maintain these changes to the environment inside the RStudio app. As a result your RStudio session will reflect changes that you have made in your relevant config files, and you will now be able to load modules via the "Custom Environment" field and install packages that reference these dependencies from within the RStudio application.

Please reach out to with any questions or concerns about this transition.

Discover Advanced Computing and Data Solutions at MSI

Our Services