How to configure the agent to run in a clustered environment.

Document ID : KB000010827
Last Modified Date : 14/02/2018
Show Technical Document Details
Introduction:

You can configure the agent to run in a clustered environment. To do this, each node in the cluster must have a physical agent installed, and the agents must be configured with a common alias to set up failover.

 

The following example shows how to configure the parameters (agentname, communication.alias_# and persistence.coldstart, and spooldir) in the agentparm.txt file for each node in a clustered environment:

Parameter

Node A

Node B

Node C

agentname

AGENT1

AGENT2

AGENT3

communication.alias_1

CLUSTER1

CLUSTER1

CLUSTER1

persistence.coldstart

FALSE

FALSE

FALSE

spooldir

export/dir/spool

export/dir/spool

export/dir/spool

Instructions:

Follow these steps:

  1. Install the agent on a partition that is mounted locally on each node of the cluster.
  2. Verify that each agent has its own copy of the agentparm.txt file, log directory, and log files. These files must reside on a locally mounted partition.
  3. Configure the spooldir parameter in the agentparm.txt file, so that the agents of the cluster share a common spool directory.

Note: When failover occurs and a workload object restarts on another node, a common spool directory ensures that the agent can retrieve the spool file and can continue updating it as required.

  1. Configure the communication.alias_n parameter on the agent of each node to enable each agent to respond to AFMs for all alias agents in the cluster.
  2. Configure each application package as an alias agent in addition to configuring each agent on the scheduling manager.
Additional Information:

agentcluster.png