WebSphere ND: Client reliance on Deployment Mngr
Chapter 12, page 535 suggests that J2EE clients should do the following:
"Bootstrap into the Deployment Manager and navigate the namespace to the cluster hosting the target application. By always keeping the Deployment Manager up, you have a reliable point of entry into the namespace."
But doesn't this make the Deployment Manager a single point of failure? I haven't heard (or learned) how to "cluster" DMs. Additionally, I have read and heard elsewhere that only your administrative/configuration tasks should be reliant on the Deployment Manager; in general, the advice seems to be that client apps should not be affected by the Deployment Manager being taken out of service.
Interestingly, the following page 536, says this:
"Each node and server has its own local, read-only view of the repository data. This is important, because it eliminates a single point of failure by not requiring app serverrs to connect to the Deployment Manager to read their current configuration."
So if app servers avoid the DM, why should clients use the DM namespace?
I could well have misunderstood the point of the argument on these pages, and any elucidation (particularly from the author) would be very gratefully received.
David (Leeds, UK).