LSB_ENABLE_HPC_ALLOCATION

Syntax

LSB_ENABLE_HPC_ALLOCATION=Y|y|N|n

Description

When set to Y|y, this parameter changes concept of the required number of slots for a job to the required number of tasks for a job. The specified numbers of tasks will be the number of tasks to launch on execution hosts. The allocated slots will change to all slots on the allocated execution hosts for an exclusive job in order to reflect the actual slot allocation.

This improves job scheduling, job accounting, and resource utilization.

When LSB_ENABLE_HPC_ALLOCATION is not set or is set to N|n, the following behavior change will still take effect:

Pending reasons in bjobs output keep task concept
Changes for PROCLIMIT, PER_SLOT and IMPT_SLOTBKLG
Event and API changes for task concept
Field "alloc_slot nalloc_slot" for bjobs –o
-alloc option in bqueues, bhosts, busers, and bapp remains
Command help messages change to task concept
Error messages in logs change to task concept

The following behavior changes do NOT take effect:

Command output for bjobs, bhist, and bacct
Exclusive job slot allocation change

Important:

After changing the LSB_ENABLE_HPC_ALLOCATION setting, the cluster admin must run badmin mbdrestart to restart mbatchd and make it take effect in mbatchd.
Changing LSB_ENABLE_HPC_ALLOCATION takes effect immediately for the affected commands, (that is, bjobs, bhist, bacct, bapp, bhosts, bmod, bqueues, bsub, and busers). Therefore, it is not required to restart mbatchd to enable the change for b* commands.
For a non-shared LSF installation, b* commands take effect on slave hosts according to the local lsf.conf setting. If LSB_ENABLE_HPC_ALLOCATION is not defined in the local lsf.conf, slave b* commands will call the master lim and take the configuration from the master host. To allow a slave take the latest configuration from the master, master lim must do a reconfig after configuration changes on the master.

Default

N

For new installations of LSF, LSB_ENABLE_HPC_ALLOCATION is set to Y automatically.