I
currently have a problem with a dynamic MQSeries send port, and I'm not sure how
to handle the issue. Whenever the MQ Server is down or unreachable my
orchestration has to detect the failure and do something else instead of sending
the message to the MQ server.
I
tried setting the "Delivery Notification" setting of the port to "Transmitted",
and catch the DeliveryFailureException.
Today, during testing, I was notified that work would be done on the MQ
server and that it may be unavailable. Testing on my orchestration continued
anyway, however it did not behave as expected. Instead of failing on the Send
shape, the orchestration instances would just get stuck on the Send operation.
All my send port instances are marked as "Active", and all the orchestrations
are now "dehydrated". My exception handling code in the scope where the Send
shape is located never got executed. Both the retry cound and retry interval for
the send port are set to 0, so I'm sure that no retries are taking
place.
Is there any way to set some timeout on the send ports? I am aware that I
could set a timeout in a long running Scope shape, however it looks like that if
I do that the Send port instances would still be running after the Timeout exception. Once the MQ comes back online, the messages would be sent (even after I executed the code for the Timeout exception) and I don't want that. I would never want the
messages to be sent to MQ once the Timeout or DeliveryFailureException exceptions occurred.
I
would appreciate any help or suggestion about this.
Thanks