My desired flow is:
endpoint routes to bean
bean creates multiple child messages from source
split on body
child messages process in parallel, up to a max
aggregate results
reply w/ aggregated results
I've tried several different configurations. I've been able to make it all
process like a queue, or wide open parallel. Wide open parallel exhausts my
DB pool.
Here are the current route definitions:
main route:
from("stomp:request_queue").
bean(new MyGenerator()).
split(body(), new MyAggregationStrategy()).
parallelProcessing().timeout(0).
inOut("seda:child_queue?timeout=0").
end().
recipientList(simple("stomp:${headers[reply-to]}"));
child route:
from("seda:udm_kli_sql_queue?concurrentConsumers=4").
threads(1, 4).
bean(new MyRunner());
I thought that using parallelProcessing in the main route and
concurrentConsumers in the child route would give me the desired behavior,
but it doesn't.
Any suggestions?
--
View this message in context:
http://camel.465427.n5.nabble.com/Limiting-thread-pool-size-for-parallelProcessing-tp5755185.html
Sent from the Camel - Users mailing list archive at Nabble.com.
