You are viewing a plain text version of this content. The canonical link for it is here.

Posted to users@camel.apache.org by StevenD <st...@gmail.com> on 2012/01/02 13:06:06 UTC

Re: Howto use delay message brokerage in a Camel Route using AMQ_SCHEDULED_DELAY

Thanks, this was really helpful. I noticed however that the delay only worked
for the first retry, all subsequent retries would be handled immediately. I
solved this by resetting the scheduledJobId back to null just before sending
it back to the queue. My exception handler now looks like:

...
.onException(Exception.class)
     .handled(true)
    .beanRef("dispatchRetryTagger")
    .choice()
        .when(header("dispatchRetryCount").isLessThan(3))
        .setHeader("AMQ_SCHEDULED_DELAY", constant("4000"))
        .setHeader("scheduledJobId", constant(null))
        .to("activemq:queue:myqueue")
    .otherwise()
        .removeHeader("AMQ_SCHEDULED_DELAY")
        .to("direct:dlq")
    .end()



--
View this message in context: http://camel.465427.n5.nabble.com/Howto-use-delay-message-brokerage-in-a-Camel-Route-using-AMQ-SCHEDULED-DELAY-tp4282763p5114568.html
Sent from the Camel - Users mailing list archive at Nabble.com.

Re: Howto use delay message brokerage in a Camel Route using AMQ_SCHEDULED_DELAY

Posted by garrydias <ga...@gmail.com>.

I found the problem. Is something between Eclipse and Maven Plugin
(M2Eclipse).

When I ran /mvn clean install -e/ my test cases were executed by Maven
lifecycle.

My testCases has a Thread.sleep(10000) at the last line. Sometimes, I did
not wait these 10 seconds and I killed Maven process through the red square
in Eclipse console view. When I did that, the process was killed within
Eclipse but not in Windows (7) OS. Through Windows Task Manager I could see
a "java" process running for every time I ran /mvn clean install -e/. And
each process was an activemq consumer. When I killed one of these "zombie"
process the consumer amount was decreased in the activemq web console.

So, before discover this, I had at least dozens of java processes running a
single activemq consumer for *activemq:destination* and each process was
represented by different versions of my program. Weird things could have
happened, like the exception I had posted before.

When I killed all java proccess via Task Manager my application consumed
scheduled messages succesffuly...

However, now scheduled messages (using AMQ_SCHEDULED_CRON) are succesfully
consumed... but once. In Activemq Web Console (/Scheduled/ menu) the field
/next scheduled time/ shows the next date to consume the message but when
the current time matches with this field, the message is not consumed but
/next scheduled time/ is recalculated.

- I used this cron for tests: *1,2,3,4,5,6,7,8 * * * **;
- I started my app at 4:00AM;
- 4:01AM the message WAS consumed and /next scheduled time/ was recalculated
to 4:02AM
- 4:02AM the message WAS NOT consumed and /next scheduled time/ was
recalculated to 4:03AM
- 4:03AM the message WAS NOT consumed and /next scheduled time/ was
recalculated to 4:04AM
and so on...

After 4:08AM the /next scheduled time/ was recalculated to 5:01AM. Normal
behavior! Except for the failure in consumption.

Consume the first scheduling is a normal behavior for AMQ scheduled
messages?

Thanx again

--
View this message in context: http://camel.465427.n5.nabble.com/Howto-use-delay-message-brokerage-in-a-Camel-Route-using-AMQ-SCHEDULED-DELAY-tp4282763p5724423.html
Sent from the Camel - Users mailing list archive at Nabble.com.