slurmctld
Slurm central management daemon
TLDR
Clear all previous state from last checkpoint
$ slurmctld -c
Set the daemon's nice value$ slurmctld -n [value]
Write log messages to a specified file$ slurmctld -L [path/to/output_file]
SYNOPSIS
slurmctld [options]
DESCRIPTION
slurmctld is the central management daemon for Slurm. It monitors all other Slurm daemons and resources, accepts work (jobs) from users, and allocates computing resources to those jobs.
The controller maintains the state of the entire cluster including node availability, job queues, and resource allocations. It is typically run on a dedicated management node.
PARAMETERS
-c
Clear all previous state and start fresh-n _value_
Set nice value (typically negative for higher priority)-L _logfile_
Write log to specified file-f _config_
Read configuration from specified file-D
Run in foreground (don't daemonize)-v
Verbose logging (can be repeated)-h
Display help information-V
Display version information
CONFIGURATION
/etc/slurm/slurm.conf
Main Slurm configuration file defining cluster nodes, partitions, scheduling policies, and controller settings./etc/slurm/cgroup.conf
Control group configuration for resource isolation and accounting.
CAVEATS
Must be run as root or the configured SlurmUser. Requires slurm.conf configuration. For high availability, can be configured with a backup controller. Clearing state (-c) will lose all job history.