We’re in the process of performing this migration and would like to discuss issues that we’re having with someone who has done this successfully already.

Our understanding is that, unlike with 2.0, physical connections to MQ are logically shared. This makes the use of commits and rollbacks mandatory, as opposed to “best practice” behavior. In our case this is a big deal.

Also, our understanding of the behavior of the product when there is a failure in the physical connection is different than before, again pointing to the need to carefully and consistently perform commits.

Are we on the right track? Have you been through this? If so, I want to hear from you.

Thanks,

Paul,

Why does using connection pools increase your concern about commits and rollbacks? What is “different” about how physical connection failures are handled?

Mark