Hello Everyone, I am working at a shop that is revamping it's use of MQ. To date, my personal experience supporting MQ has not included using clustering. A proposal is under consideration to deploy hundreds of MQ servers as one cluster. This would include a handful of enterprise side QMGRs (including a few on z/OS) with the majority of QMGRs being widely dispersed. There is currently no need for the dispersed QMGRs to connect to each other so channels would exist between the enterprise QMGRs and the individual, dispersed QMGRs. Each dispersed QMGR would essentially be a clone of one another ... meaning something in the MQ object names on each QMGR would be unique but the number, type and attributes of the MQ objects from one QMGR to another would be uniform. There is also no use of client connections currently under consideration. The topology of the QMGRs will leave little, or no, opportunity to employ workload distribution. The enterprise side QMGRs will regularly be used to deliver messages to all of the dispersed QMGRs. While I crack the manuals and review the list's archives to learn more specifics about clustering benefits and pitfalls, I would appreciate hearing from others with experience using clustering, especially larger clusters. If the proposal is adopted, with what am I likely to be confronted? Will the single cluster transmit queue on each enterprise QMGR adequately service messages intended for many / all of the dispersed QMGRs in this cluster configuration? Without the benefits of workload distribution, I am concerned whether clustering in this situation is prudent.
Thanks for your thought, Bill Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive