We use clustering throughout our production environment. This at the moment isn't very big as we only have 20 Qmgrs and we've experienced so many problem it's untrue.
My experience is that the concept of clustering is excellent however implementing it on versions of MQ up to 5.2 and 5.2.1 CSD03 is somewhat flaky. The fact IBM have improved the management of clusters in v 5.3 shows that clustering problems are inevitable. I don't understand how the cogs all wirr in clustering but it's a shame they can't improve the reliability of clustering rather that just improve the tools to fix the problems.
All the major problems we've had all revolve around a qmgr not being able to see a clustered object with no propagation of changes to clustered objects second on our fault hit list.
I don't think we've had one clustered environment where we haven't encountered a 2189 error (MQRC_CLUSTER_RESOLUTION_ERROR) at some stage (hence my nickname). I've pretty much got to grips with fixing them now in most cases.
This probably sounds like I'm slagging off clustering and I don't mean to. It's got it's probelms but I'd stick with it as I now have experience of most of the problems, plus the potential of reduced overhead in creating channels and remote queues in a large configuration is enormous.
One final thing though, we have had to introduce a non-clustered solution in production due to our clustering problems (unfortunately for us the projects using MQ have put a change freeze on MQ so we can't upgrade to 5.3.). So I'd recommend it's worth having non-clustered scripts ready as a backup plan on critical systems in case of a major clustering failure. _________________ 2189 errors make me
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum