slurmd
Slurm compute node daemon
TLDR
Report node rebooted when daemon restarts (for testing)
$ slurmd -b
Run the daemon with a given nodename$ slurmd -N [nodename]
Write log messages to a specified file$ slurmd -L [path/to/output_file]
Read configuration from a specified file$ slurmd -f [path/to/file]
SYNOPSIS
slurmd [options]
DESCRIPTION
slurmd is the Slurm compute node daemon. It monitors all tasks running on the compute node, accepts new tasks from the central controller, launches tasks, and terminates running tasks upon request.
Each compute node in a Slurm cluster runs slurmd to participate in workload scheduling and execution. The daemon communicates with slurmctld (the central controller) for job management.
PARAMETERS
-b
Report node rebooted when daemon restarts-N _nodename_
Run with specified node name-L _logfile_
Write log to specified file-f _config_
Read configuration from specified file-c
Clear previous node state-D
Run in foreground (don't daemonize)-v
Verbose logging-h
Display help information-V
Display version information
CAVEATS
Must be run as root. Requires proper Slurm configuration (slurm.conf). The node must be registered with the Slurm controller. Firewall must allow communication with slurmctld.