Again looking at waits, and the average wait time for my tracewrites, when running Profiler interactively on my workstation monitoring the engine running on the server, server shows average waits for this of over 1000ms. Hey that's slow. Is it always that slow and I never noticed before, or does this indicate a problem?
Possibly a networking problem, which we suspect from time to time, but the asynch_network_io does NOT seem to be a problem, and in general running apps or queries in SSMS seems pretty snappy.
Also, when the profiler is generating these waits, besides the tracewrite it generates oledb waits, it looks like a couple per tracewrite totally almost the same time, although that is less per wait.
Another trace running by script on the server writing to a local file, generates no waits.
SQL 2008 R2 engine on Server 2008 R2 all 64 bit, workstation is Windows 7 Enterprise 64 bit, 1gbps LAN, but I'm not sure where the servers actually are so there must be slower links though tracert doesn't show anything but 1ms delays.
Thanks.
Josh
ps - run interactively ON the server, waits are about 25ms. very negligible waits in all other categories. also server is running under VM.