Hi Y'all,

Any suggestions on how to solve what appears to be a "data delay" across a network? Users here are on a combined Novell/NT network. A PC runs in their server room (W2K) that does a series of complex calculations for the user so as to offload the processing from their PC. The "processing" table is just a table that has all the parameters in one record and has a field in it that says "Calculate Me!". The PC in the server room looks down this column and calculates anything marked .T. and prints the result back to their printer.
When running totally on Novell (users and Calc PC), the user's would tag a record and every 5 seconds the Calc PC would do a search on this column (using an index) and so the delay was at most about 5 seconds (unless it was already processing something). Users got used to the fact that most times, the printout would come out quite quickly.

But the new Calc PC is now on a W2K box. "Loop" time on the box is still 5 seconds. There is a status screen on the Calc PC and after someone submits something, you can see that the Calc PC has done a search 3 or 4 times until it finds the tagged record (ie. 15 to 20 seconds has elapsed).

I suspect that it is an index caching issue - ie. the Calc PC is not refreshing this index right away and so it does not get the new flag until 15 or 20 seconds after submission (tagging).

How can I get something like a "re-read" on the data (and index) so that before I do a search, I can make sure the index is updated and thus I will get a hit on the very first search. FLUSH only works (I think) from the client machine to the server, not in reverse? If I write out a bit of data each time (eg. REPLACE ThisField WITH ThisField), would this refresh the indexes?

Note that the "loop" is done by a timer sitting on the status form on the Calc PC - the form is not closed between loops as I think in the old days this forced a re-read.

Thanks,
Albert