blob: 93b32a7a7c3e517088ded955850a6527d1bb8838 [file] [log] [blame]
This job completion plugin runs a program after the job is completed.
The JobCompType is jobcomp/script and the JobCompLoc is the pathname of the script to run.
The plugin sets up the environment for the program to contain the following variables:
ACCOUNT: Job's account value
ARRAYJOBID: The job ID of the first element of a job array
ARRAYTASKID: The array index value for this element of a job array
BATCH: Was the job submitted as a batch job ("yes" or "no")
CLUSTER: The name of the cluster
CONSTRAINTS: Node features the job is constrained to use
DB_FLAGS: Flags to indicate the scheduling cycle that scheduled the job
and whether the job was started
DEPENDENCY: Original list of job IDs on which this job depends
DERIVED_EC: Highest exit code of the job and after a semicolon ":" the signal number
ELIGIBLE: The eligible time of the job (UNIX timestamp)
END: The end time of the job (UNIX timestamp)
EXITCODE: The exit code of the job and after a semicolon ":" the signal number
GID: The GID of the group the job was run for
GROUPNAME: The name of the group the job was run for
HETJOBID: The job ID of the heterogeneous job leader
HETJOBOFFSET: The zero origin offset within a collection of heterogeneous job components
JOBID: The job ID of the slurm job
JOBNAME: The name of the job
JOBSTATE: The state of the job when it ended
LIMIT: The time limit of the job (minutes or "UNLIMITED")
NODECNT: The number of nodes the job ran on
NODES: The nodes the job ran on or an empty string if it wasn't assigned
PACKJOBID: See HETJOBID
PACKJOBOFFSET: See HETJOBOFFSET
PARTITION: The partition the job ran on
PATH: Sets to the standard path
PROCS: Number of processors allocated to the job
QOS: The QOS of the job
RESERVATION: The reservation name of the job
START: The start time of the job (UNIX timestamp)
STATEREASONPREV:The previous state of the job
STDERR: The job's STDERR path
STDIN: The job's STDIN path
STDOUT: The job's STDOUT path
SUBMIT: The submit time of the job (UNIX timestamp)
TZ: The time zone of the system the job ran on
UID: The UID of the user the job was run for
USERNAME: The username of the user the job was run for
WORK_DIR: The job's working directory