MQ rejects

[AChrist]

Greetings,

DataStage: 8.7
MQ Connector Variant: 5.3

We are writing a utility to support our MQ team to move data from queue 1 to queue 2 on a scheduled cyclic basis. For this implementation, our strategy for guaranteed delivery of all messages includes writing messages that are rejects to a dataset. This dataset will then becomes input to a process that writes the rejected messages to the target queue.

We are hoping to deploy this pattern for many sets of queues and wanted to avoid creating an error queue for every set of queues.

One of our tests is to inhibit the target queue for input, to ensure that messages are written to the reject dataset and then processed to the target queue.

When we use the pattern with local queues messages are sent down the reject link as expected, but when we use clustered queues, messages are not written to the reject link and the job is aborting..

Cluster queue is enabled on the properties tab and we have opted to reject rows based on every condition on the reject tab.

Is there a way to reject on clustered queues?

A variant of this test was to limit the number of rows on the target queue, and process more rows than the limitation. This test also fails for clustered queues but works for local queues. With local queues, the number of rows down the queue link equals the limitation on the target queue and every other message goes down the reject link. With clustered queues DataStage reports that all messages were successfully written to the queue, but they weren't, only the number of rows written to the target equals the limitation, yet no rows go down the reject link.

I suspect that both errors are related, but I'm not sure what to do for a next step in troubleshooting.