1Slurm(1) Slurm System Slurm(1)
2
3
4
6 Slurm - Slurm Workload Manager overview.
7
8
10 The Slurm Workload Manager is an open source, fault-tolerant, and
11 highly scalable cluster management and job scheduling system for large
12 and small Linux clusters. Slurm requires no kernel modifications for
13 its operation and is relatively self-contained. As a cluster resource
14 manager, Slurm has three key functions. First, it allocates exclusive
15 and/or non-exclusive access to resources (compute nodes) to users for
16 some duration of time so they can perform work. Second, it provides a
17 framework for starting, executing, and monitoring work (normally a par‐
18 allel job) on the set of allocated nodes. Finally, it arbitrates con‐
19 tention for resources by managing a queue of pending work. Optional
20 plugins can be used for accounting, advanced reservation, gang schedul‐
21 ing (time sharing for parallel jobs), backfill scheduling, resource
22 limits by user or bank account, and sophisticated multifactor job pri‐
23 oritization algorithms.
24
25 Slurm has a centralized manager, slurmctld, to monitor resources and
26 work. There may also be a backup manager to assume those responsibili‐
27 ties in the event of failure. Each compute server (node) has a slurmd
28 daemon, which can be compared to a remote shell: it waits for work,
29 executes that work, returns status, and waits for more work. An
30 optional slurmdbd (Slurm DataBase Daemon) can be used for accounting
31 purposes and to maintain resource limit information.
32
33 Basic user tools include srun to initiate jobs, scancel to terminate
34 queued or running jobs, sinfo to report system status, and squeue to
35 report the status of jobs. There is also an administrative tool scon‐
36 trol available to monitor and/or modify configuration and state infor‐
37 mation. APIs are available for all functions.
38
39 Slurm configuration is maintained in the slurm.conf file.
40
41 Man pages are available for all Slurm commands, daemons, APIs, plus the
42 slurm.conf file. Extensive documentation is also available on the
43 internet at <https://slurm.schedmd.com/>.
44
45
47 Copyright (C) 2005-2007 The Regents of the University of California.
48 Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
49 Copyright (C) 2008-2009 Lawrence Livermore National Security.
50 Copyright (C) 2010-2013 SchedMD LLC.
51
52 This file is part of Slurm, a resource management program. For
53 details, see <https://slurm.schedmd.com/>.
54
55 Slurm is free software; you can redistribute it and/or modify it under
56 the terms of the GNU General Public License as published by the Free
57 Software Foundation; either version 2 of the License, or (at your
58 option) any later version.
59
60 Slurm is distributed in the hope that it will be useful, but WITHOUT
61 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
62 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
63 for more details.
64
65
67 sacct(1), sacctmgr(1), salloc(1), sattach(1), sbatch(1), sbcast(1),
68 scancel(1), scontrol(1), sinfo(1), squeue(1), sreport(1), srun(1),
69 sshare(1), sstat(1), strigger(1), sview(1), slurm.conf(5), slur‐
70 mdbd.conf(5), slurmctld(8), slurmd(8), slurmdbd(8), slurmstepd(8),
71 spank(8)
72
73
74
75
76June 2018 Slurm System Slurm(1)