>Thanks Neil,
>Now i remember that I've read that thread too - obviously my memory is shorter than yours :(
>We are in VFP 8, don't get the eof chars, but we get some very inconsistent errors:
>- broken indexes ( .cdx)

What OS r u using client & server. Do u have oplocking & read caching disabled.

If server downed is it being shutdown properly. Had case in last month where client just logged off and then turned power off, cache not being written back to disk = corrupted database. Required onsite visit to find out what was the problem. No problems since.

>- duplicate IDs (on a free table)

Are you using autoinc or pull the ID's from a control file ?
Caching can cause this.

>- IDs in reverse order (if you look at the date with no index set IDs should appear in ascending sequence, but sometimes they are reversed)

Check to see default sort order in DBC

>We update the table on a LAN from local and remote locations (the remote connected over FDDI). Updates come from users, but also from an unassisted computer.
>I could not make any connection between duplicates and the updates from the unassisted computer.
>It seems that the index corruptions happen after updates from remote locations...

Could also be caching.


Regards NMCD