problem after using pure vpn

[konnic]

i'm new in foxpro and have situation in my company

My company about have ~30 clients, use novell as file server to store fpd 26 *.dbf files and using windows 2000 as vpn gateway to let other branch to retrieve the database. For some critical update to database, clients use dial-up pcanywhere to remote control the terminal computer to novell directly in physical lan environment.

After harddisk crash of novell, I decide to integrate the database with windows 2000 server and remove pcanywhere feature. However, sometimes the update process fails from vpn. There are also data integraty problem as some duplicate records appear in the database. Also the performance is rather slow using vpn.

I have following questions:

1. is it the programming problem (i.e. locking file, retrieving the whole database using select table command) to induce data integraty problem?

2. wat's the best method to solve the performance problem and data integraty progblem?
a. re-coding (i'm not sure using sql statement is better than using foxpro command as sql select only retrieve the result records)
b. upgrading the foxpro system (to fpw 3.0 or visual foxpro? my company only use the menu and print , no macro ,reporting)

 

[infinitelo]

backwords last to first:
b. upgrade, visual foxpro 7 should integrate into your windows 2000 operating system better. but i would not expect performance gains.
a. since you have multilple users i have to ask are two or more "on" at the same time and updating the same record. that would cause your data problems. you may have to set up a system where a record or file is checked out
2. data checking is a programming issue. speed is hardware/os issue. i respectfully submit that novell is better suited for you needs. {feel free to disagree}

1. its not clear to me weather you had the same issues before you switced operating systems. even so, if users can introduce such errors. a good look at the code is called for. you have to ask youself whatif another user is accessing this record right now.
since you are making a copy of data with sql your programs cannot tell weather a record has been updated since a "session" has started. that would cause dups, and wierd data. unless you check for dups before adding. and compare data before submission.

since computers can only count to one, the smallest details are the most significant.