We are creating a trigger service that will update a back end system based on a specific document being published to the broker.
Michael,
ART 6.5 adds two new services: pub.art.notification:suspendPollingNotification and pub.art.notification:resumePollingNotification, which allow you to suspend or resume polling notifications while leaving the buffer tables in tact, whatever the type of notification.
If you are pre-6.5 and using basic polling notifications, then suspending the notification is not a problem. You can do it from Administrator and it will leave all the DB objects in tact. If, however, you are using an insert/update/delete notification pre-6.5, then you have a problem, as suspending the trigger from Administrator also destroys the DB objects and you lose notifications in the mean time.
For Broker, you can use the pub.trigger:suspendRetrieval and pub.trigger:resumeRetrieval services introduced in one of the 6.1 fixes to suspend trigger retrieval and allow docs to pile up on the Broker. From 6.5 onwards you can also access these functions in Administrator from Settings > Resources > Trigger Management.
Hope this helps.
Jonathan Heywood
webMethods Professional Services
We are doing something similar to this. When the target system is down, we suspend the notifier and send a page via email. We have to resume the trigger manually.
We also suspend the trigger every night when we know the target system is down. We do this by scheduling the suspend and resume services.
But you want to suspend the notifier for a period of time and then try it again after some time. To do this I think you would have to suspend the trigger when the target system is down. Then schedule the resume service every 30 minutes (whatever interval you want). If the notifier is actively poling, the resume service shouldn’t cause any problem (you might want to test this). But if it is suspended it will start it back up and try again. If the target system is still down the notifier will be suspended again.
To suspend the notifier when the broker is down, you might need to run a job to connect to the broker and wait for a document. If the broker goes down, this will cause an exception. Catch the exception and suspend the trigger. I do something similar to this using the broker client API in a java service on the IS. So the service should resume the notifier when it gets connected and suspend if an exception is thrown. Schedule the service to run every 10 seconds with the “repeat from end of invocation” checked. This will restart the service every 10 seconds whenever it fails to connect to the broker.
Thanks,
Steve
Thanks for the suggestions. Suspending the notification is certainly something I will look into. Is there an option for us if we are not on IS 6.5? We are currently only on 6.1.
I’m sorry, all of my reply was referring to suspending the trigger, but I’m sure you can do the same with a notifier. In 6.0.1 you can call the wm.art.admin:setNotificationStatus. The actions are enableNotification and disableNotification.
Thanks,
Steve
Hi Steve,
Can you send me that jaav service which you developed for broker client. I have developed a similar functionality but it doesn’t cover like what you said.
my mail id:satish.vutti@gmail.com
Thanks in advance,
Regards,
Satish