Hi,
We have a situation which cause us a lot of problem.
Until now it was not a big problem but now we use UM often, it begins to be very embarassing.

We use some topic to feed several data destination.
But sometimes, for a reason I ignore, the trigger is not triggered. It is not systematic but I think it is a question a data amount. I know that my messages go to bus because I see there in snoop. (Entreprise Manage). But I know that my messages doesn’t go to my output services which is triggered by a tJMS trigger.

Could you say me why I have this behaviour.

Thanks for your reply.

wM 9.10

Hi Vital,
Are you using BPMS? What fix level are you on?
Thanks, Jane.

No we just use Integration Server (flow service) and Universal Messaging.
Our fixes are up to date

Hello Vital,

Is the trigger stuck at the moment?
If so, could you collect a certain set of diagnostic information for us?

Thanks,
Stefan

Actually now that’s work.

If you happen to stumble upon the issue again, could you collect the following diagnostic information and open a support incident with it:

• Data directory from UM server(s).
• Heap dump from UM(s).
• 2 thread dumps from IS instances - 3 minutes apart from one another.
• Heap dumps from IS instances.
• IS logs.
• List of Trigger(s) stuck. Detailed information for each stuck trigger - native, JMS, concurrent, serial.
• UM resources (channels or queues) those stuck trigger(s) use.
• Use Enterprise Manager and navigate to the named object corresponding to the stuck trigger. Take a screenshot and write down the number of pending events. If the number of pending events is equal to “ClientQueueSize”, we know UM will not deliver more events by specification. This would mean processing on IS side has lead to too many events being unacknowledged.
• Screenshot the IS trigger (on all IS nodes) in order to confirm if its processing queue is full

The list is pretty exhaustive and with that information the chance is quite high that we will be able to identify the problem. Unfortunately the root cause for this type of issues can vary significantly in their nature, so we really need all of this diagnostic data in order to track it down.

Thanks,
Stefan

Ok that’s happening again : what I have to do ?

It seems to be a problem by trigger and not by Topic because some trigger are trigerred and some other trigger are not trigerred.

So I have questions : If a Topic has two destinations (two triggers) :

• The Processing mode can be concurrent ? (If it is the case, were the two triggers systematically triggered ?)
• Durable Suscribe name should’nt be the same ? (It is my cas, I’m not sur it should be like that)

Hi Vital,

Please collect the information in Stefan’s post and raise a support incident.

Thanks, Jane.