slurm(1)

📦
version 26.05.5.1 · 6 versions
, in arch rolling, epel 10, nixpkgs unstable

💖 slurm(1) — the Slurm Workload Manager overview, bestie edition 💖

okay so this is slurm(1), a page from the Slurm Commands manual set, and it's basically the "what even is Slurm" page. Sit down, grab a boba, we're doing this. 🧋✨

🌸 NAME 🌸

Slurm — Slurm Workload Manager overview. that's it, that's the name, she's iconic like that. 👑

✨ DESCRIPTION ✨

The Slurm Workload Manager is an open source, fault-tolerant, and highly scalable cluster management and job scheduling system for large and small Linux clusters. like, she scales. she does not care about your cluster's size, she's serving either way. 💅

Slurm requires no kernel modifications for its operation and is relatively self-contained. no cap — she just moves in and gets to work. 🎀

👑 the three key functions (the holy trinity, if you will)

As a cluster resource manager, Slurm has three key functions:

  1. Allocation era — it allocates exclusive and/or non-exclusive access to resources (compute nodes) to users for some duration of time so they can perform work. your nodes, your time slot, your moment. ⭐
  2. The framework — it provides a framework for starting, executing, and monitoring work (normally a parallel job) on the set of allocated nodes. she starts it, she runs it, she watches it. a full-service icon. 🦋
  3. The referee — it arbitrates contention for resources by managing a queue of pending work. somebody has to keep the peace and it's her. 😤
and then there's the optional plugins — the accessories, the little extras that ate, no crumbs: - accounting - advanced reservation - gang scheduling (time sharing for parallel jobs) - backfill scheduling - resource limits by user or account - sophisticated multifactor job prioritization algorithms

🧚 the daemons (the cast)

Slurm has a centralized manager, slurmctld, to monitor resources and work. there may also be a backup manager to assume those responsibilities in the event of failure — like a cover artist who knows the whole setlist just in case. 🎤

Each compute server (node) has a slurmd daemon, which can be compared to a remote shell: it waits for work, executes that work, returns status, and waits for more work. the way she just sits there, patient, ready, employed. 💀✨

An optional slurmdbd (Slurm DataBase Daemon) can be used for accounting purposes and to maintain resource limit information. optional but honestly? she's giving main character energy in any cluster with a budget. 💜

🛠️ the basic user tools (your starter pack)

  • srun — to initiate jobs
  • scancel — to terminate queued or running jobs
  • sinfo — to report system status
  • squeue — to report the status of jobs

There's also the administrative tool scontrol, available to monitor and/or modify configuration and state information. and APIs are available for all functions — every single one, periodt. 🫶

📁 where the config lives

Slurm configuration is maintained in the slurm.conf file. that's the one. that's the file. guard it with your life. 🔥

📚 more reading (iykyk)

Man pages are available for all Slurm commands, daemons, APIs, plus the slurm.conf file. Extensive documentation is also available on the internet at https://slurm.schedmd.com/. go read it, bestie, she's thorough. 🌈

🎀 COPYING 🎀

Copyright (C) 2005-2007 The Regents of the University of California. Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER). Copyright (C) 2008-2009 Lawrence Livermore National Security. Copyright (C) 2010-2022 SchedMD LLC.

This file is part of Slurm, a resource management program. For details, see https://slurm.schedmd.com/.

Slurm is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.

Slurm is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

🫶 real talk: free software, no promises, read the license. she's generous but she's not your mom.

🦋 SEE ALSO 🦋

the whole friend group, in order, exactly as listed:

sacct(1), sacctmgr(1), salloc(1), sattach(1), sbatch(1), sbcast(1), scancel(1), scontrol(1), sinfo(1), squeue(1), sreport(1), srun(1), sshare(1), sstat(1), strigger(1), sview(1), slurm.conf(5), slurmdbd.conf(5), slurmctld(8), slurmd(8), slurmdbd(8), slurmstepd(8), spank(8)

🌸 okay that's the whole page!! this one's Slurm 26.05, part of the Slurm Commands manual set — and yes she's dated, every icon has a timestamp. love her, run her on your cluster, bye bestie!! 💖✨🎀

📦 where this page comes from