>I have a customer who is complaining that he can run a sales report at the end of the day and get one total, then without conducting any additional transactions on any of the workstations, re-run the report 10 or so minutes later and get a different (but correct) total. Suspecting caching, I had him uncheck the box that says “Enable write caching on the disk” on the server. He says the problem still exists. If other workstations are still in the app with tables still open with "un-flushed" data, I could see where this could cause the problem. Is there a setting on the workstations that affects caching that I need to change as well? The DBC and all tables are on the server. TIA!

Hi Mark,

you can enable disk cache at Server side ( if you have a UPS backup !),
because this is not a Server side problem, this is a Client side problem.

When a client open a net file, the Windows redirector cache it into the memory of the Client,
the master rule of the cache manager it is :

if the file size don't change then it is probable that it don't change ...

Because DBF files uses all fixed fields, and memo files use a block schema,
if you do a Update, the dbf size don't change ( only it'is file datetime change is updated).
But the big problem is this:
- Until you don't close the table file handle, VFP don't flush the updated index immediatelly.

Result,
- you can solve the problem of the redirector cache manager:
a link:
http://www.lexialearning.com/updates/OpportunisticLocking/FixOpportunisticLockingIssues.pdf
or search into the web for Windows Redirector Cache Write behind

but i think that you can solve the VFP problem only with the new FLUSH command of VFP9.

Fabio