LSF Version 7.3 - Administering Platform LSF

Administering Platform LSF 359
Goal-Oriented SLA-Driven Scheduling
Configure service classes for multiple SLAs that map to different EGO consumers.
1 Log in as the LSF administrator on any host in the cluster.
2 Edit
lsb.serviceclasses to add the service class definition for EGO-enabled
SLA.
Begin ServiceClass
NAME = Venezia
CONSUMER = Venezia
PRIORITY = 10
GOALS = [VELOCITY 10 timeWindow (8:00-18:00)]
MAX_HOST_IDLE_TIME = 10
DESCRIPTION = Daytime SLA
End ServiceClass
You can copy another service class definition from this file as a starting point;
remember to change the
NAME and CONSUMER of the copied profile.
Configure the additional parameters for an EGO-enabled SLA:
◆ CONSUMER=ego_consumer_name—enables an SLA to get resources
from EGO. Specify the name of the EGO consumer from which hosts are
allocated to the SLA. This parameter is mandatory for EGO-enabled SLA.
TIP: You should make the service class name the same as the consumer name.
3 Save the changes to lsb.serviceclasses.
4 Run badmin reconfig to reconfigure mbatchd.
Configure EGO
consumers and
resource groups
Prerequisites: You must be logged on as an EGO cluster administrator and you
should have already added hosts to your cluster. To create consumers, you must
either be a cluster administrator or a consumer administrator for the branch on
which you are creating the consumer.
1 Log on to the Platform Management Console.
2 Build the consumer tree for SLA jobs.
See “Create a consumer” in Administering and Using Platform EGO for detailed
steps.
The Platform EGO consumer tree organizes consumers into a structure that
makes it easy to apply resource plans. The plan cannot be defined without the
consumer tree. The tree defines organizational relationships among
consumers. The plan defines the resource allocation.
3 Configure resource groups for SLA jobs.
See “Create a resource group by host names” or “Create a resource group by
resource requirement” in Administering and Using Platform EGO for detailed
steps.
Create resource groups from the Platform Management Console to ensure your
consumers have the appropriate group of hosts available to them according to
the requirements of the workload.