0 Replies Latest reply on Sep 1, 2010 1:51 PM by obolabs

    Understanding Deployments & Executions

    obolabs

      Please read the following knowing that I am still only finding my way around JBPM as well as JBoss, although I've got some experience with weblogic and JEE in general.

      Here I'm trying to wrap my mind around how the deploy and execution scenarios work together.

       

      Consider this Ant-based deploy process:

      <taskdef

           name="jbpm-deploy"

           classname="org.jbpm.pvm.internal.ant.JbpmDeployTask"

           classpathref="jbpm.dependencies" />
      <jbpm-deploy file="ProcessTester.bar" />

      As far as I can tell, it parses the jbpm xml {for validation?} and  writes the .bar file contents into the JBPM4_LOB table, as well as creating entries for the process in other deployment related tables.

      This is done without needing anything from the JBoss+JBPM runtime, other than the support libraries.

       

      Once running, BPM Console checks the DB for available deployments.

      When a process is 'Start'-ed in the Console's Process Overview, its .bar is read from DB, reassembled and executed in place. In other words, the business archive does not have to be explicitly in the JBoss classpath.

       

      Is that more or less it or are there missing bits in this description?

       

      Also, I am fairly certain that you can also deploy processes from .bar files included in the classpath by simply putting the file into the deploy folder of the JBoss instance. However, I remember reading in another thread that JBPM 4.4 doesn't yet have a deployer, but one can use 4.1,2,3 deployers?

      This method of deployment would result in the same DB artifacts being generated and the execution would be same.