6:16 p.m.
Hello everyone,
I'm new to drools & I'm trying my hand at creating a rule-flow in which I can
back a couple of steps &
the KnowledgeSession is restored to the same state it was in earlier. I'm using Drools
5.1.0 M1 & Eclipse 3.5.
I have a rule-flow with the following nodes:
A->B->C
where A would be a node asking for an input (our fact), B would be a Rule Task that would
apply certain rules on our fact &
at C, I may decide that I need to back up to node A. Both A & C are custom work
items.
I've managed to rollback the transaction (I borrowed a lot of code from the test
classes in the source, thanks a ton for them).
But I just can't get my process to start again.
I'll paste a couple of lines of code from my test class below:
.....
ProcessInstance pi = ksession.startProcess("sampleflow");
ksession.fireAllRules();
ut = (UserTransaction) new InitialContext().lookup( "java:comp/UserTransaction"
);
ut.begin();//Is beginning the transaction boundary here correct?
.....
//Now I roll back the transaction depending on the input at node C
ut.rollback();
.....
//And recreate the knowledge session
ksession = JPAKnowledgeService.loadStatefulKnowledgeSession(id, kbase, null, env);
I expected the session to restart the process from node A once it was recreated. Am I
going in the right direction?
Or is my basic understanding of how session persistence works in Drools Flow flawed?
I'd really appreciate a push in the right direction.
Thanks in advance,
Mahesh.
6:24 p.m.
You don't need to demarcate the user transactions there.. Drools Flow will
be in charge of that.
you only need to keep the session id in order to be able to restore it when
you need it..
2009/10/19 Nair, Mahesh <Mahesh.Nair(a)qwest.com>
Hello everyone,
I'm new to drools & I'm trying my hand at creating a rule-flow in which
I can back a couple of steps &
the KnowledgeSession is restored to the same state it was in earlier. I'm
using Drools 5.1.0 M1 & Eclipse 3.5.
I have a rule-flow with the following nodes:
A->B->C
where A would be a node asking for an input (our fact), B would be a Rule
Task that would apply certain rules on our fact &
at C, I may decide that I need to back up to node A. Both A & C are custom
work items.
I've managed to rollback the transaction (I borrowed a lot of code from the
test classes in the source, thanks a ton for them).
But I just can't get my process to start again.
I'll paste a couple of lines of code from my test class below:
.....
ProcessInstance pi = ksession.startProcess("sampleflow");
ksession.fireAllRules();
ut = (UserTransaction) *new* InitialContext().lookup(
"java:comp/UserTransaction" );
ut.begin();*//Is beginning the transaction boundary here correct?*
.....
*//Now I roll back the transaction depending on the input at node C*
ut.rollback();
*.....*
*//And recreate the knowledge session*
ksession = JPAKnowledgeService.loadStatefulKnowledgeSession(id, kbase, *
null*, env);
I expected the session to restart the process from node A once it was
recreated. Am I going in the right direction?
Or is my basic understanding of how session persistence works in Drools
Flow flawed?
I'd really appreciate a push in the right direction.
Thanks in advance,
Mahesh.
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
--
- http://salaboy.wordpress.com
- http://www.jbug.com.ar
- Salatino "Salaboy" Mauricio -
2:48 a.m.
Quoting "Nair, Mahesh" <Mahesh.Nair(a)qwest.com>:
[...]
I expected the session to restart the process from node A once it was
recreated. Am I going in the right direction?
The process will be in the same state
as it ended in the last time.
In general, I would not advice to use database transactions like this to
roll back changes in your process (because you don't necessarily even
know the boundaries of that transaction, so you could have strange
side-effects).
Couldn't you model the same logic using for example a loop (draw a
connection back to a previous location in the process), exception
handlers, event signaling, or a combination of all of these? It would
make it much more visible for the user what's going on.
Kris
Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
5:55 p.m.
Thanks Kris for your suggestion.
What we are trying to model is similar to an interview process (for diganosing some kind
of trouble with a system). The user wants to be able to go back to any of the previously
answered question and change the answer and the restart the interview process from that
point (after which the flow would take a different route from the earlier route). if we
have a loop back it would mean that all nodes are connected to all other nodes !. Events
would definetley work but we are looking at some other approach which coul dbe mor
elegant
Mahesh
-----Original Message-----
From: Kris Verlaenen [mailto:kris.verlaenen@cs.kuleuven.be]
Sent: Monday, October 19, 2009 7:49 PM
To: Rules Users List; Nair, Mahesh
Cc: rules-users(a)lists.jboss.org
Subject: Re: [rules-users] Question about backing up a flow
Quoting "Nair, Mahesh" <Mahesh.Nair(a)qwest.com>:
[...]
I expected the session to restart the process from node A once it was
recreated. Am I going in the right direction?
The process will be in the same state
as it ended in the last time.
In general, I would not advice to use database transactions like this to roll back changes
in your process (because you don't necessarily even know the boundaries of that
transaction, so you could have strange side-effects).
Couldn't you model the same logic using for example a loop (draw a connection back to
a previous location in the process), exception handlers, event signaling, or a combination
of all of these? It would make it much more visible for the user what's going on.
Kris
Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
5579
days inactive
5580
days old
3 comments
3 participants
participants (3)
-
Kris Verlaenen -
Mauricio Salatino -
Nair, Mahesh