14.7 Problems with Individual Jobs

To determine why a particular job will not start, there are several helpful commands:

Checkjob evaluates the ability of a job to start immediately. Tests include resource access, node state, job constraints (such as startdate, taskspernode, and QoS). Additionally, command line flags may be specified to provide further information.

-l <POLICYLEVEL>: Evaluates impact of throttling policies on job feasibility.
-n <NODENAME>: Evaluates resource access on specific node.
-r <RESERVATION_LIST>: Evaluates access to specified reservations.

checknode

Displays detailed status of node.

mdiag -b

Displays various reasons job is considered blocked or non-queued.

mdiag -j

Displays high level summary of job attributes and performs sanity check on job attributes/state.

showbf -v

Determines general resource availability subject to specified constraints.

14.7 Problems with Individual Jobs

See Also