Slurm batch queueing system
These pages constitute a HowTo guide for setting up a Slurm workload manager software installation based on RHEL 8 Linux and clones such as AlmaLinux and RockyLinux. It is strongly recommended to first read the official Slurm_documentation pages.
These pages are for Slurm >= 23.02 and EL8/EL9 Linux. Older archived versions of these pages with information related to pre-23.02 Slurm and CentOS 7 Linux may be found in the archive or in the v24.06 release of these pages.
Much of the information should be relevant on other Linux versions as well. The present pages are not official documents from SchedMD but only a user’s contribution to the Slurm community.
The information has been subdivided into sub-pages for separate Slurm topics:
Slurm operations for daily management.
Slurm configuration setting up the services.
Slurm database for storing user and job data.
Slurm accounting defining user accounts.
Slurm job scheduler for prioritizing jobs.
Slurm cloud bursting extending the local cluster with nodes in the Cloud.
Some Slurm tools at https://github.com/OleHolmNielsen/Slurm_tools
Slurm documentation
Documentation from SchedMD about Slurm:
The Slurm_homepage.
The Slurm_documentation pages.
Slurm_Quick_Start admin guide.
Commercial Support by SchedMD®: https://www.schedmd.com/support.php
Command_Summary (2-page sheet).
Slurm configuration file slurm.conf
Slurm NEWS (changes in recent versions of Slurm) and RELEASE_NOTES (change to the relevant Tag in stead of the master branch) on changes in recent versions of Slurm.
Slurm_download with links to external tools under Download Related Software.
Subscribe to Slurm_mailing_lists.
The slurm_devel_archive.
Slurm_publications and presentations.
Citing academic work involving Slurm from JSSPP 2023: Architecture of the Slurm Workload Manager, see https://slurm.schedmd.com/faq.html#cite
Youtube videos about Slurm.
Slurm_man_pages overview of man-pages, configuration files, and daemons.
Slurm_bugs tracking system.
Large Cluster Administration Guide (clusters containing 1024 nodes or more).
Slurm Elastic Computing (Cloud Bursting) (Google Cloud, Amazon EC2 etc.)
Documentation from other sources: