We will soon be migrating to Red Hat Enterprise Linux (RHEL) version 7. This new OS is available to users now to detect issues ahead of the general release. Our latest GPU cluster, m9g, is also available through the RHEL 7 OS.


To access it, connect to the host rhel7ssh.fsl.byu.edu instead of ssh.fsl.byu.edu. When you submit jobs add -C rhel7 to your sbatch/salloc flags.

Be sure to submit jobs to rhel7 from rhel7ssh.fsl.byu.edu, and submit jobs to the old image from ssh.fsl.byu.edu. Submitting jobs to rhel7 from ssh.fsl.byu.edu will fail, as will submitting jobs to the old OS from rhel7ssh.fsl.byu.edu.


This OS is available as alpha quality, and we may need to cancel jobs and/or reboot nodes unexpectedly. Things may break or change in the future, but please let us know when you hit issues; we cannot fix issues we are not aware of.


Most software is missing. A lot of software built on the old OS will not work on the new image and will need to be recompiled. Please let us know if there is software you were using on the old image that is not available or does not work on the new image. We are prioritizing the installations of software that is specifically requested by users.


Initially, we will have approximately 40 nodes available with 4 Nvidia P100 GPUs each (we're calling them "m9g"). We will also have a few nodes of other types, converted from our main cluster. The portion of our cluster running the updated system image, will increase over time.