6.1 Configuration of PFM - Agent for Platform in a cluster system
This section describes the configuration in which PFM - Agent for Platform is applied to an HA cluster system. For an overview of cluster systems and details about the system configuration for operating Performance Management in a cluster system, see the chapter that describes construction and operations in a cluster system in the JP1/Performance Management User's Guide.
In this chapter, the term cluster system refers to an HA cluster system.
If you operate PFM - Agent for Platform in a cluster system, you can perform a failover if an error occurs, thus improving availability.
To operate PFM - Agent for Platform in a cluster system, you need to create an environment in which the same PFM - Agent for Platform can be executed on both the executing node and standby node. You also need to store a set of data, including data files, configuration files, and log files, on the shared disk.
Provided that you are monitoring processes that are running in cluster configuration, we recommend that you operate PFM -Agent for Platform by using the logical host.
To operate PFM - Agent for Platform in a cluster system, use the configuration in the following figure.
|
|
![[Figure]](GRAPHICS/ZU060001.GIF)
Necessary data such as definition information and performance information is stored on the shared disk, and is inherited in the event of a failover. If a single logical host contains multiple Performance Management programs, they use the same shared directory.
Multiple instances of PFM - Agent for Platform can be executed on a single node. For a configuration that contains multiple cluster configurations (an active-active configuration), execute PFM - Agent for Platform in each logical host environment. Each instance of PFM - Agent for Platform runs independently and can perform a failover separately.