Copyright © TIBCO Software Inc. All Rights Reserved


Chapter 6 Basic MM Configuration : Site Topology Overview

Site Topology Overview
The topology file contains deploytime information such as what processing units to deploy to specific machines in your environment. You need to know information about the machines that will host the agents you plan to deploy, for example information about the machines’ operating system and IP address.
You also need to know what remote invocation software you’ll use to start remote processes on these machines: TIBCO Hawk, PSTools, or SSH.
Changes to the EAR file do not affect the topology configuration. However if the cluster, processing unit, or agent definitions in the CDD file change, you must recreate the site topology file using the updated CDD.
Editing the Site Topology File in a Text Editor or in the Graphical Editor
When possible, use the graphical site topology file editor in TIBCO BusinessEvents Studio. It provides validation and structure that are helpful.
When working on runtime installations, however, it may not be possible to use TIBCO BusinessEvents Studio. An annotated site topology file template is available, so you can edit the XML-based topology file in a text editor. It is located here:
BE_HOME/mm/config/site_topology_template.st
If you are editing the file directly, adapt the GUI-based instructions accordingly. The configuration requirements are the same in both cases.
Summary of Site Topology Configuration
Using the canvas-based editor in TIBCO BusinessEvents Studio, you create a visual representation of the desired site topology. Property sheets let you configure each item represented by the topology diagram icons: the cluster, hosts, deployment units, and processing unit configurations.
The output of this activity is an XML file used in MM. Summary steps are shown below. Detailed steps are provided in Configure the Site Topology in TIBCO BusinessEvents Studio.
1. Configure Cluster Properties  In the Cluster Properties tab, reference the fully configured CDD and EAR files for your project. See Project, Master, and Deployed Locations of CDD and EAR Files for more on the use of these files.
2. Add Deployment Units (DUs)  Add DUs to the canvas as needed. For each DU, specify the following:
See the note in Project, Master, and Deployed Locations of CDD and EAR Files for an important limitation when deploying multiple DUs on one machine.
One or more processing unit configurations (PUCs). You’ll configure the PUCs in the next step.
3. Add Processing Unit Configurations (PUCs) to DUs  For each PUC, select one processing unit (PU) from the list of PUs defined in the CDD file. Set deploytime properties such as the JMX ports used by MM to communicate with the deployed engine.
4. Add Hosts  Here you specify the machine configuration, including the software used on the remote machines to start remote machines. Connect hosts to the DUs you want to deploy on them. Multiple hosts can use the same deployment unit, as long as the configuration is the same in each case.
Project, Master, and Deployed Locations of CDD and EAR Files
In the topology file, you reference three locations for the CDD and EAR files. The files in each location must be exact copies:
Project CDD file: In the cluster configuration tab, you specify a locally available copy of the project CDD, used only at design-time for configuring the topology file in TIBCO BusinessEvents Studio.
Master CDD and EAR files: Also in the cluster configuration tab, you specify the location of the master CDD and EAR files. These copies must be manually copied to the specified location on the MM server, for use in deployment.
Deployed CDD and EAR files: In the Deployment Unit settings, you specify where MM will place the CDD and EAR files when it performs deployment.
The project and master CDD can be in the same location if you are using one machine to configure the topology file and to run MM server. These two sets of fields are available in case you are configuring the topology on a different machine from the MM server machine.
Limitation: One CDD and EAR file per Cluster Machine: Currently deployment is at the Machine level and each machine can have only one copy of the deployed CDD and EAR files. If you specify multiple DUs for the same host, problems may occur because CDD and EAR files are copied only to the first DU’s deployed files location.
Deployment-Specific Processing Units
In general, you can reference one processing unit multiple times to create different processing unit configurations (PUCs). However processing units that have deployment-specific settings cannot be used in this flexible manner.
Agent-Instance-Specific Properties
If a processing unit contains agent-instance-specific properties (such as agent key and priority settings), you must use it in only one PUC, which is used in only one DU, that is itself used only once in the deployment.
Host-Specific Processing Units
Processing units can host-specific settings. If a deployment unit contains a PUC that references such a processing unit, you must link it only to the appropriate host for deployment. For example, the Coherence cache provider property tangosol.coherence.localhost property is a host-specific setting, and so is the TIBCO BusinessEvents DataGrid property be.engine.cluster.as.listen.url.
Global Variables
Note that global variable overrides (if any) set in the master CDD are used for command-line deployment. They can be overridden by TIBCO BusinessEvents Monitoring and Management settings. If you plan to deploy using MM, override global variables using MM, instead of in the CDD file.

Copyright © TIBCO Software Inc. All Rights Reserved