| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| FDL model conversion - best practices? |
« View previous topic :: View next topic » |
| Author |
Message
|
| callocam |
 Posted: Fri Mar 30, 2007 4:28 am Post subject: FDL model conversion - best practices? |
 |
|
Novice
Joined: 09 Jun 2006
Posts: 14
Location: Middletown, Ct
|
Greetings,
My workflow application is making the following changes to the FDL model:
Removal of a PES
Removal of NOOP
Removal of appropriate data and control connectors
We are not changing the data container nor are we adding any new program activities.
My issue is how do I convert my old model instances that are running? Are there certain design approaches that I can take that are better than others?
My thoughts on what I should do are as follows:
1. Read in all process instances that are in a ready state of any old process models.
2. Store their data containers and workitem name.
3. Terminate the old process instances.
4. Create and start the new model and load the data container from the deleted model.
Another one of my questions was whether the new instance that I create would automatically flow to a given activity based on the data container I load.
Thank you very much for any advice! 
Ettore |
|
| Back to top |
|
 |
| jmac |
| Posted: Fri Mar 30, 2007 5:42 am Post subject: Re: FDL model conversion - best practices? |
|
|
Jedi Knight
Joined: 27 Jun 2001
Posts: 3081
Location: EmeriCon, LLC
|
| callocam wrote: |
My issue is how do I convert my old model instances that are running? Are there certain design approaches that I can take that are better than others?
|
Ettore: This can not be done.... The best thing is to let them die naturally. All new instances will be created using the new template.
What you are suggesting, might be possible, however, it would certainly be non-trivial code to write. You would have to Re create every instance and "replay" it to the current point of execution. In other words, your comment about starting with a particular container, and the engine recognizing what activity to advance to, is not the way things work. If the natural death option is a non-starter, and you need to pursue this replay option, give me a call and we can talk... but like I said, it MAY not be possible, especially with the way your process models work.
_________________
John McDonald
RETIRED |
|
| Back to top |
|
|
| callocam |
| Posted: Fri Mar 30, 2007 8:01 am Post subject: |
|
|
Novice
Joined: 09 Jun 2006
Posts: 14
Location: Middletown, Ct
|
Hi John,
Hope your doing well.
We're attempting a proof of concept to see if this is feasible. But my understanding of the nature of the project, we will need to convert the unopened workitems in case they get 'team reassigned' to the PES which is supposed to be gone.
That would be the reason why we can't let these old instances just die off naturally.
Thanks for your advice!
Ettore |
|
| Back to top |
|
|
| callocam |
| Posted: Wed Apr 04, 2007 11:03 am Post subject: Re: FDL model conversion - best practices? |
|
|
Novice
Joined: 09 Jun 2006
Posts: 14
Location: Middletown, Ct
|
| jmac wrote: |
What you are suggesting, might be possible, however, it would certainly be non-trivial code to write. You would have to Re create every instance and "replay" it to the current point of execution. In other words, your comment about starting with a particular container, and the engine recognizing what activity to advance to, is not the way things work. If the natural death option is a non-starter, and you need to pursue this replay option, give me a call and we can talk... but like I said, it MAY not be possible, especially with the way your process models work. |
Hi John,
At this point, this is where we stand...
1. We query activity instance to get only the worktasks for a given state 'ready'.
2. We then capture the container and save it.
3. We then terminate the process instance.
4. We then do a create and start using the new template with the data container from the old PI.
But what we've found is that the 'replay' option you mentioned seems to not be necessary. When we dummied up a container with the values we needed to get the new started process instance to a specific activity it seemed to have advanced to that activity.
What do you think?
Ettore |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
