Hi!

I have a few questions about JBoss BPM, regarding creating the environment for development and production,deploying processes, and user handling in bpm (task service)
Fort he job we need to create and start business processes.
I’m using JBoss EAP6.4, and jBoss BPM 6.3.
I’ve read the ducuments (install guide, development guide) i could find, but the information isn’t enough to get anywhere.

1) We need to create two environments: One developer (standalone), and a production (clustered). I already installed the JBoss EAP and BPM, and the setup of domain mode is done. However, i can’t figure out, how the clustered BPM should work.
a. I know there is a GIT, and a Maven repository int he whole picture. But can someone shed some light on how are these used? What is needed for clustered production?
b. Is the zookeeper+helix configuration needed for clustered BPM? What does it really do? Does it cluster the VFS-repo (the niogit folder)?
c. What’s needed for clustered runtime? Only the Maven repo? Or only the git? Or both?
d. How does the clustered BPM engine works?

2) How can we deploy our processes is the business-central is in server-exec mode? I know we can click on „Build & Deploy” in Business Central’s project authoring, but that deploys to the dev server (i dont see the way we can deploy EARs and WARs to an application server), not to the clustered production servers.

Here’s the deployment process i was able to achieve so far (but this is long, and for big project/lots of processes it’s no viable):
1. Make a project with one business process. Build it. Let’s say, the GAV (grouId, artifactId, version) is: testOrgUnit:testProcessArtifactId:1.0
2. We will find a JAR files under ${jboo.maven.dir}/testOrgUnit/testProcessArtifactId/1.0 called testOricessArtifactId-1.0.jar
3. I can upload this jar to the other BPM server’s business-central „->”
4. After that i go to the „ -> asd” and create a new definition, using the same groupId, artifactId, version.
5. Now i can start a process instance.
This way of deploying a process is just too long. It’s not maintainable if we’ll have 20-30 or more processes.
Is there an automated, or semi-automated way?

Thanks all,
Mark