2016-11-11 21:25:24 +01:00
|
|
|
slurm (workload manager)
|
2015-08-19 01:25:23 +02:00
|
|
|
|
2016-11-11 21:25:24 +01:00
|
|
|
Slurm is an open-source workload manager designed for Linux clusters
|
|
|
|
|
of all sizes. It provides three key functions. First it allocates
|
|
|
|
|
exclusive and/or non-exclusive access to resources (computer nodes)
|
|
|
|
|
to users for some duration of time so they can perform work. Second,
|
|
|
|
|
it provides a framework for starting, executing, and monitoring work
|
|
|
|
|
(typically a parallel job) on a set of allocated nodes. Finally, it
|
|
|
|
|
arbitrates contention for resources by managing a queue of pending work.
|
2015-08-19 01:25:23 +02:00
|
|
|
|
2016-11-11 21:25:24 +01:00
|
|
|
The SLURM controller (slurmctld) can run without elevated privileges,
|
|
|
|
|
so it is recommended that a user "slurm" be created for it before Slurm
|
|
|
|
|
is executed.
|
2015-08-19 01:25:23 +02:00
|
|
|
|
|
|
|
|
# groupadd -g 311 slurm
|
|
|
|
|
# useradd -u 311 -d /var/lib/slurm -s /bin/false -g slurm slurm
|
|
|
|
|
|
2016-11-11 21:25:24 +01:00
|
|
|
Next, a configuration file can be build using your favorite web browser
|
|
|
|
|
and the file /usr/doc/slurm-14.11.8/html/configurator.html.
|
2015-08-19 01:25:23 +02:00
|
|
|
|
|
|
|
|
Optional dependencies:
|
|
|
|
|
HWLOC=yes|no (default: no), requires hwloc
|
|
|
|
|
RRDTOOL=yes|no (default: no), requires rrdtool
|
|
|
|
|
NUMA auto-detected, requires numactl
|
