Wanderland

Tips of setting up BPELSE in a glassfish cluster

Tips of setting up Glassfish Cluster

#1 Walk-through guides

1. App Server Guides
   
2. Nice Blogs
   

#2 Watch for caveats : Multi-machine High Availability cluster (default)

1. multicast is supported by the hosts
   
2. machines must be located in the same subnet
   
3. System time on the machines should be as close as possible

 Tips of setting up BPELSE on glassfish cluster

#1 Walk-through guides

Install guide for JBI Runtime

BPELSE cluster Walk-through with Extensive examples 

 #2 Create DB schema for Persistence and/or Monitoring

Currently, unlike running in a non-clustered glassfish,  when BPELSE (persistence/monitoring on) is enabled on a cluster target,  DB schema will not be automatically created . You can try one of the 2 workarounds below to address it:

1. Disable BPELSE on the cluster and Enable it on the Server target with persistence/clustering, after verifying DB schema is created, disable the server target and enable cluster target
   
2. Use db scripts (Only derby and oracle are supported for now)
   

• Note: You can not share the schema among clusters or between cluster and server target

#3 Simple JMSIn --> Clustered BPELSE --> JMSOut

Download Project

• Note:

    The example can be simply changed to connect to a cluster of MQ brokers by specifying the list on connectionURL

#4 You can run BPEL Monitor Tool to monitor BPELSE in a cluster (http://wiki.open-esb.java.net/Wiki.jsp?page=BPELMonitor#SupportingClustering)

#5 You can test BPELSE settings on a cluster programmatically (Useful if you are using BPELMonitor API and want to add cluster validation logic)

TestCustomExtensionMbeans.java

Classpath: appserv-rt.jar ($GLASSFISH_HOME/lib), j2ee.jar ($GLASSFISH_HOME/lib),  bpelmonitor-api.jar (bpelMonitorTool.zip), jbi-admin-common.jar ($GLASSFISH_HOME/jbi/lib)


 
 

Posted at 12:15AM May 20, 2008 by Mei Wu in Sun  |  Comments[0]