I’m curious if there is a good pattern for creating a new messages once all
previous tasks have been properly executed.

So say I have 100 A messages which are consumed by a bunch of tasks in your
cluster.

When all the A tasks have executed, you want to run 1 B task to compute
their results.

So it’s sort of like map-reduce in a way.

I think what I could do is use JMS transactions, and then read the value of
an incrementing pointer, when it hits 100, I can create the B task, and
then commit() the entire session.


-- 

Founder/CEO Spinn3r.com
Location: *San Francisco, CA*
blog: http://burtonator.wordpress.com
… or check out my Google+ profile
<https://plus.google.com/102718274791889610666/posts>
<http://spinn3r.com>

Reply via email to