6:17 a.m.
Hi all,
Before I ask my question, just wanted to know if others have noticed
that the Hudson docs for Drools 5 is not available online. Or maybe
it's just me?
I'm a little bit of a beginner in this particular area on the Drools
flow and its persistence mechanism. I have a couple questions that
maybe someone can answer for me.
Say that I have a long-running workflow process. When it reaches a
human task (ie. take out trash), the workflow gets saved/persisted to
the DB. Once the trash is taken out, the user updates this status.
The workflow continues and finishes.
My questions are fairly simple:
1) How would I query for a list of human tasks that have not yet been
completed? I believe the answer for this one is simply in the demo/
test/example codes, but somehow I can't seem to find it. I see mostly
codes that sends complete/fail methods to indicate that the process is
completed. Is there an API for this that I can look at?
2) Is there an API that allows me to query completed items, failed
items, etc? It'd be great if someone can point to the right class.
3) What happens to old long-running processes currently still waiting
to be finished (and stored in the DB) when I decide to change the
workflow process (either manually or in Guvnor)? What would be a good
way in dealing with a situation with changing workflow processes?
Create a new workflow (with a new workflow id) and have all new
requests go through the new workflow? If that's the case, then how
would one "upgrade" the old workflow processes? I believe jBPM also
has an issue with this, along with all other workflow engines with
long-running processes. Is there a recommended way?
4) Is it a good idea to just have Drools store the workflow status
data (ie. task start and completion date) and use that to query and
calculate performance or is it better to create an additional workflow
step that updates my own database table containing these data? Is
there also a recommended way for this procedure?
I appreciate any insights and recommendations that you have.
Thanks,
Chris
6:51 a.m.
Good questions, here goes some answers:
1,2) As you know the human task service here is fully pluggable and
drools have their own implementation. This implementation is based on
Web Services Human Task standard and is implmented in MinaTaskClient
class using the TaskServiceSession class to do all that kind of stuff
with tasks. Like assignments, abort, skip, query, etc. You could take
a look at this class to see how it works or just use it. This class is
inside the Drools-Task module, and of course, this module have all the
functionality about tasks. Also you can take a look at
WSHumanTaskHandlerTest where all this kind of functionalities is
tested.
3) there is a lot of situation when you don't have to update the old
processes that start the execution with an old definition. there is
also a lot of situation when you want to updated. In jBPM and here i
think that the best approach will be some kind of migration procedure
for one version to another.. but is just my opinion.. and i think that
this need to be analyzed for each particular situation.
4) If you see the TaskData class have the following attributes:
private User createdBy;
private Date createdOn;
private Date activationTime;
private Date expirationTime;
that are persisted when a task is created.. this TaskData is an
embedded entity inside the Task entity. You also can take a lot at
that.. to see if all the attributes that this class have are enough
for you.
Let me know if i can help you with something else!
On Wed, Feb 25, 2009 at 9:17 AM, CK <chrisckc(a)gmail.com> wrote:
Hi all,
Before I ask my question, just wanted to know if others have noticed that
the Hudson docs for Drools 5 is not available online. Or maybe it's just
me?
I'm a little bit of a beginner in this particular area on the Drools flow
and its persistence mechanism. I have a couple questions that maybe someone
can answer for me.
Say that I have a long-running workflow process. When it reaches a human
task (ie. take out trash), the workflow gets saved/persisted to the DB.
Once the trash is taken out, the user updates this status. The workflow
continues and finishes.
My questions are fairly simple:
1) How would I query for a list of human tasks that have not yet been
completed? I believe the answer for this one is simply in the
demo/test/example codes, but somehow I can't seem to find it. I see mostly
codes that sends complete/fail methods to indicate that the process is
completed. Is there an API for this that I can look at?
2) Is there an API that allows me to query completed items, failed items,
etc? It'd be great if someone can point to the right class.
3) What happens to old long-running processes currently still waiting to be
finished (and stored in the DB) when I decide to change the workflow process
(either manually or in Guvnor)? What would be a good way in dealing with a
situation with changing workflow processes? Create a new workflow (with a
new workflow id) and have all new requests go through the new workflow? If
that's the case, then how would one "upgrade" the old workflow processes?
I
believe jBPM also has an issue with this, along with all other workflow
engines with long-running processes. Is there a recommended way?
4) Is it a good idea to just have Drools store the workflow status data (ie.
task start and completion date) and use that to query and calculate
performance or is it better to create an additional workflow step that
updates my own database table containing these data? Is there also a
recommended way for this procedure?
I appreciate any insights and recommendations that you have.
Thanks,
Chris
_______________________________________________
rules-users mailing list
rules-users(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
--
- Salatino Mauricio -
5753
days inactive
5753
days old
1 comments
2 participants
participants (2)
-
CK -
Mauricio Salatino