We have a 3gb HQ db with 6 stores. Whenever a 401 needs to send a large amount of data to HQ, the following happens:
HQ spikes all 8 CPUs at 100% for about 40 minutes then completes Store HQ client times out, never registers that the 401 completed Next time the store connects, same thing happens again until I delete the
401 in question.This is a major problem, because when we do a physical inventory, we need to do all sorts of gyrations to ensure that the data is uploaded and the offending 401 is then removed.
It's really not all that much data, what is it with the RMS db design that makes this SO INEFFICIENT?? I mean 50 thousand DB record updates is routine for many systems.
I have been working with MS support on this and the support tech claims it is "working" because it works when the store DB and HQ db reside on the same machine. This is true but that's not a real world scenario. They want to blame our net connection. Plenty of apps run just fine on DSL lines (this one has 0.00% packet loss)
I seem to have hit a dead end with MS support, can anyone shed some light on this?