>>Actually FLUSH only causes VFP to commit any pending changes in its internal buffers to the OS for writing to the disk. If the OS has write behind caching enabled this may not be sufficient to eliminate the loss of data in the case of a power outage. The only reliable method to eliminate this problem is the use of a UPS and a controled shutdown in this event.
>
>
>Unfortunately, a UPS is not always an available option. We have data logging applications that write 3-5 rows per second for approximately 1 Gigabyte of data in environments where UPS's are explicitly forbidden (plant safety rules in the petro-chemical industry.) The write behind caching is a real problem, on Win2K I find the file length has been extended by the appropriate amount but filled with chr(0). This is a real pain because the table has candidate keys that are invalid. Rebuilding the indexes is the biggest problem because of the time.
>
>Do you know of any effective ways to dissable write behind caching on Win2K and WinNT? I've done quite a bit of research on the web and MSDN, but nothing has helped.

Albert,

From what I've read in "Inside SQL Server 7.0" it can be done. That book recommends that write-behind caching be disabled. It also notes that some controllers do not allow for it to be disabled and recommends at least one that does. If you like, I'll take a look Monday to find more details. Unfortunately, the book is at my office and I don't have a copy here at home.

Regards,