>>>We have a VFP9 database running on networks. In some case, even with only 2 or 3 users and less than 50,0000 records in largest file, this becomes painfully slow. When one logs into the same databse in single user mode, the same system runs like the clappers (same time of day on network - thus it's not the network per se. The offending network has PCs running on Windows 7 Enterprise with Service Pack 1. The network runs on Extreme Networks switches (1 Gbit). Storage is hosted on NetApp.
>>>
>>>When in multiuser mode, the VFP app runs the code:
>>>SET EXCLUSIVE OFF
>>>SET REFRESH TO 1
>>>SET REPROCESS TO AUTOMATIC
>>>
>>>There are no special commands in the config.fpw file. I have tried adding ...
>>>
>>>TMPFILES=c:\work
>>>EDITWORK=c:\work
>>>SORTWORK=c:\work
>>>PROGWORK=c:\work
>>>
>>>but I can't see any files appearing in c:\work so unsure that is working in any case or if it would help?
>>>
>>>Appreciate this is rarther general - but any other ideas to help speed things up?
>>>Many thanks.
>>
>>Denis,
>>
>>this sounds like the classical oplocks issue. See here:
>>
>>http://www.dataaccess.com/whitepapers/opportunlockingreadcaching.html
>
>technically, spot on. But with that tiny data set size vfp should run pretty well without caching if indexed correctly.
>So perhaps the Oplocks speed gain just masks the speed loss from not enough optimizing...

Optimising is important, but once it is running fast when only one user is in and then it slows down when more users join in it has always been fixed for me by making these adjustments to oplocks. I think the caching referred to there is just locally caching, whether it's all in RAM or swapped out to the hard drive. I don't think that the size of the data would determine if caching takes place or not.

Now, having said that, I know very little about the OS side of things so may be completely incorrect.