Hi!

FLUSH guarantee write all data from buffers to disk (even in network), so the risk you will lose data during crash is minimal, but still exist.
I had such problem even without crash. 2-3 times at month I get the same report from client that table was corrupted. And still the same problem - corrupted header (that does not allow to add new records in run-time). The most interesting thing is that they say that no any crash occurred. So this is not a crash problem. This is something internal in communication across local network between VFP and files on file server.

>>Right, I know this. I was hoping for something preventative...:)
>>
>>>>This is a real problem. I posted this issue a long time ago, see thread #236192. Never did get a solution, I ended up having to write a stored
>database procedure to trap when the error occurred.
>
>Eric,
> I believe that the error happens during a PC or network crash while writting. I keep log files of when a user logs into the system and out. If they don't log out before they log in it is flagged. My logs show that the errors happened about the time a users PC crashed. I'm not sure what you can do for this except go to back office which I will never do because of the slow perfomance and excessive cost relitive to the size of the businesses that I deal with. Namely 2-10 station networks in auto body shops.