IBM LoadLeveler for AIX 5L: Using and Administering

Step 6: Prioritize the queue maintained by the negotiator

Each job submitted to LoadLeveler is assigned a system priority number, based on the evaluation of the SYSPRIO keyword expression in the configuration file of the central manager. The LoadLeveler system priority number is assigned when the central manager adds the new job to the queue of jobs eligible for dispatch. Once assigned, the system priority number for a job is never changed (unless jobs for a user swap their SYSPRIO, or NEGOTIATOR_RECALCULATE_SYSPRIO_INTERVAL is not zero). Jobs assigned higher SYSPRIO numbers are considered for dispatch before jobs with lower numbers. See How does a job's priority affect dispatching order? for more information on job priorities.

You can use the following LoadLeveler variables to define the SYSPRIO expression:

ClassSysprio

The priority for the class of the job step, defined in the class stanza in the administration file. The default is 0.

GroupQueuedJobs

The number of job steps associated with a LoadLeveler group which are either running or queued. (That is, job steps which are in one of these states: Checkpointing, Preempted, Preempt Pending, Resume Pending, Running, Starting, Pending, or Idle.)

GroupRunningJobs

The number of job steps for the LoadLeveler group which are in one of these states: Checkpointing, Preempted, Preempt Pending, Resume Pending, Running, Starting, or Pending.

GroupSysprio

The priority for the group of the job step, defined in the group stanza in the administration file. The default is 0.

GroupTotalJobs

The total number of job steps associated with this LoadLeveler group. Total job steps are all job steps reported by the llq command.

QDate

The difference in the UNIX date when the job step enters the queue and the UNIX date when the negotiator starts up.

UserPrio

The user-defined priority of the job step, specified in the job command file with the user_priority keyword. The default is 50.

UserQueuedJobs

The number of job steps either running or queued for the user. (That is, job steps which are in one of these states: Checkpointing, Preempted, Preempt Pending, Resume Pending, Running, Starting, Pending, or Idle.)

UserRunningJobs

The number of job step steps for the user which are in one of these states: Checkpointing, Preempted, Preempt Pending, Resume Pending, Running, Starting, or Pending.

UserSysprio

The priority of the user who submitted the job step, defined in the user stanza in the administration file. The default is 0.

UserTotalJobs

The total number of job steps associated with this user. Total job steps are all job steps reported by the llq command.

Usage notes for the SYSPRIO keyword

• The SYSPRIO keyword is valid only on the machine where the central manager is running. Using this keyword in a local configuration file has no effect.
• It is recommended that you do not use UserPrio in the SYSPRIO expression, since user jobs are already ordered by UserPrio.
• You can use the UserRunningJobs, GroupRunningJobs, UserQueuedJobs, GroupQueuedJobs, UserQueuedJobs, GroupQueuedJobs UserTotalJobs, and GroupTotalJobs parameters to prioritize the queue based on current usage. You should also set NEGOTIATOR_RECALCULATE_SYSPRIO_INTERVAL so that the priorities are adjusted according to current usage rather than usage only at submission time.

Using the SYSPRIO keyword - examples

Example 1

This example creates a FIFO job queue based on submission time:

SYSPRIO : 0 - (QDate)

Example 2

This example accounts for Class, User, and Group system priorities:

SYSPRIO : (ClassSysprio * 100) + (UserSysprio * 10) + (GroupSysprio * 1) - (QDate)

Example 3

This example orders the queue based on the number of jobs a user is currently running. The user who has the fewest jobs running is first in the queue. You should set NEGOTIATOR_RECALCULATE_SYSPRIO_INTERVAL in conjunction with this SYSPRIO expression.

SYSPRIO : 0 - UserRunningJobs