Gary tse [
http://community.jboss.org/people/garytse] created the discussion
"Re: JBPM 5 and clustering ?"
To view the discussion, visit:
http://community.jboss.org/message/614802#614802
--------------------------------------------------------------
I'm working on similar thing:
http://community.jboss.org/thread/169261?tstart=0
http://community.jboss.org/thread/169261?tstart=0
Short answer, yes, it can be -MADE- to run on a cluster. The solution is not "off
the box". In any case, I'm fairly certain you want to config jbpm5 to persist
states to DB (ie. setup persistence according to the user guide).
For the jbpm-console, I've spent a couple of days attempting to port it to Websphere
but finally given up. It's going to take too long for my (lack of) ui skill :(
As for drools-guvnor (storage of the processes) in the installer-example, it's more of
an optional thing. I'm loading the bpmn2 files directly from the file system. It
took minimal effort to port it into Websphere in my case.
The most important matter is the strategy of the stateful knowledge session.
Options:
1. You can have one stateful knowledge session per cluster-node (essentially means each
server working on their own, will need some sort of "lookup" to match the
process to the server... )
2. Share a single stateful knowledge session for all cluster-nodes.
My post is asking about the 2nd strategy and how that works... hopefully a guru will
come into rescue.
--------------------------------------------------------------
Reply to this message by going to Community
[
http://community.jboss.org/message/614802#614802]
Start a new discussion in jBPM at Community
[
http://community.jboss.org/choose-container!input.jspa?contentType=1&...]