If you are using the Oracle Coherence cache provider, and you have configured the monitored project to discover cluster members using well-known addresses (WKA) then you must make some additional changes to the project CDD so that MM can monitor and manage the cluster.You must also configure the MM.cdd to work with the cluster to be monitored, but the MM cluster itself does not use well-known address discovery.
Use unique ports — including for MM Server engine If any monitored cluster engines will run on the same machine as the MM server, ensure that all localport values are unique across the MM server and monitored cluster localport values on that machine.This section assumes that the CDD for the cluster to be monitored has already been configured for WKA cluster discovery. The additional configuration below enables MM to communicate with the cluster to be monitored.For details see Configuring a Cache OM Cluster — Cluster Tab, especially Configuring Well-Known Address Cluster Discovery.
1. Open the project to be monitored in TIBCO BusinessEvents Studio and open its CDD file in the CDD editor.This CDD must be copied for use as the master CDD. See Project, Master, and Deployed Locations of CDD and EAR Files to understand the purpose of these copies of the CDD.
2. For each PU you will deploy to a WKA machine, add the be.engine.hostaddress property (in addition to the tangosol.coherence.localhost and tangosol.coherence.localport properties as needed for WKA configuration). Set it to the same value as the localhost property:Task B Configure the MM.CDD File
1.
2. In the CDD editor, click the Agent Classes tab and select mm-class agent.
3. The broker process must connect to the cluster to be monitored. In the mm>broker properties group, add the following properties:Set the localhost property to the IP address of one of the hosts used by the cluster to be monitored. Set the localport property to one of the ports specified for that host in the well-known addresses list, specified in the master CDD (see Task A step 2).
4. If the MM server runs on the same host as any monitored cluster engine specify the following properties in the Cluster > Properties sheet:As you did with the broker process properties, set localhost to the IP of the host where the MM server will run and set the localport property to a different port than any used by any monitored cluster engines on this host machine.For all host machines with multiple network cards (NIC), add the following property to the be-engine.tra file on that host:The default value is localhost.Set this property to the IP address of the desired NIC. This IP address must match the value specified in other properties where host IP is specified.This property is required so that the engine is discovered by the MM cluster and appears as an active predefined engine. It is also required so that remote users can connect to any agents running on this host using a JMX client like JConsole.
Copyright © TIBCO Software Inc. All Rights Reserved.