1slurmd(8) Slurm Daemon slurmd(8)
2
6 slurmd - The compute node daemon for Slurm.
7
10 slurmd [OPTIONS...]
11
14 slurmd is the compute node daemon of Slurm. It monitors all tasks run‐
15 ning on the compute node , accepts work (tasks), launches tasks, and
16 kills running tasks upon request.
17 OPTIONS
19 -b Report node rebooted when daemon restarted. Used for testing
21 purposes.
22 -c Clear system locks as needed. This may be required if slurmd
25 terminated abnormally.
26 -C Print actual hardware configuration and exit. The format of out‐
29 put is the same as used in slurm.conf to describe a node's con‐
30 figuration plus its uptime.
31 --conf-server <host>[:<port>]
34 Specify the host and (optionally) port number of the primary
35 slurmctld. This host is where slurmd will fetch the configura‐
36 tion from when running in "configless" mode.
37 -d <file>
40 Specify the fully qualified pathname to the slurmstepd program
41 to be used for shepherding user job steps. This can be useful
42 for testing purposes.
43 -D Run slurmd in the foreground. Error and debug messages will be
45 copied to stderr.
46 -f <file>
48 Read configuration from the specified file. See NOTES below.
49 -G Print Generic RESource (GRES) configuration (based upon
51 slurm.conf GRES merged with gres.conf contents for this node)
52 and exit.
53 -h Help; print a brief summary of command options.
55 -L <file>
57 Write log messages to the specified file.
58 -M Lock slurmd pages into system memory using mlockall (2) to dis‐
60 able paging of the slurmd process. This may help in cases where
61 nodes are marked DOWN during periods of heavy swap activity. If
62 the mlockall (2) system call is not available, an error will be
63 printed to the log and slurmd will continue as normal.
64 It is suggested to set LaunchParameters=slurmstepd_memlock in
66 slurm.conf(5) when setting -M.
67 -n <value>
70 Set the daemon's nice value to the specified value, typically a
71 negative number. Also note the PropagatePrioProcess configura‐
72 tion parameter.
73 -N <nodename>
76 Run the daemon with the given nodename. Used to emulate a larger
77 system with more than one slurmd daemon per node. Requires that
78 Slurm be built using the --enable-multiple-slurmd configure
79 option.
80 -v Verbose operation. Multiple -v's increase verbosity.
83 -V, --version
85 Print version information and exit.
86
89 The following environment variables can be used to override settings
90 compiled into slurmd.
91 SLURM_CONF The location of the Slurm configuration file. This
93 is overridden by explicitly naming a configuration
94 file on the command line.
95
98 SIGTERM SIGINT
99 slurmd will shutdown cleanly, waiting for in-progress rollups to
100 finish.
101 SIGHUP Reloads the slurm configuration files, similar to 'scontrol
103 reconfigure'.
104 SIGUSR2
106 Reread the log level from the configs, and then reopen the log
107 file. This should be used when setting up logrotate(8).
108 SIGPIPE
110 This signal is explicitly ignored.
111
114 If slurmd is started with the -D option then the core file will be
115 written to the current working directory. Otherwise if SlurmdLogFile
116 is a fully qualified path name (starting with a slash), the core file
117 will be written to the same directory as the log file. Otherwise the
118 core file will be written to the SlurmSpoolDir directory, or
119 "/var/tmp/" as a last resort. If none of the above directories can be
120 written, no core file will be produced.
121
124 It may be useful to experiment with different slurmd specific configu‐
125 ration parameters using a distinct configuration file (e.g. timeouts).
126 However, this special configuration file will not be used by the slurm‐
127 ctld daemon or the Slurm programs, unless you specifically tell each of
128 them to use it. If you desire changing communication ports, the loca‐
129 tion of the temporary file system, or other parameters used by other
130 Slurm components, change the common configuration file, slurm.conf.
131 If you are using configless mode with a login node that runs a lot of
133 client commands, you may consider running slurmd on that machine so it
134 can manage a cached version of the configuration files. Otherwise, each
135 client command will use the DNS record to contact the controller and
136 get the configuration information, which could place additional load on
137 the controller.
138
141 Copyright (C) 2002-2007 The Regents of the University of California.
142 Copyright (C) 2008-2010 Lawrence Livermore National Security. Copy‐
143 right (C) 2010-2016 SchedMD LLC. Produced at Lawrence Livermore
144 National Laboratory (cf, DISCLAIMER).
145 This file is part of Slurm, a resource management program. For
147 details, see <https://slurm.schedmd.com/>.
148 Slurm is free software; you can redistribute it and/or modify it under
150 the terms of the GNU General Public License as published by the Free
151 Software Foundation; either version 2 of the License, or (at your
152 option) any later version.
153 Slurm is distributed in the hope that it will be useful, but WITHOUT
155 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
156 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
157 for more details.
158
161 /etc/slurm.conf
162
165 slurm.conf(5), slurmctld(8)
166September 2020 Slurm Daemon slurmd(8)