Don,

Maybe this hasn't so much to do with the original question/problem (but who knows), but I'd like to know a little more of this if you allow me. Something might be in there, and I never looked at it from this point of view.
The point is, we always seem to have some problems of these left, in the very end solved by the reset of the server. What server ? NT ...
BTW, yes, Novell seems more applicable for the problem in general, because of leaving records locked at switching off the PC, leaving you with an "UNKNOWN" user in "Connections", depending on the set timeout to reset that physical connection. So, in that case to be solved by connecting the PC to Novell again (of delete the connection). But never mind this.

When you say "We solved the problem by adding code to make sure each workstation logon was unique", what do you mean ? code where ? and what do you mean exactly by "workstation ID" ?

Again, I feel something's in there, looking at some other kind of problem with making logon's unique by our own app. I mean, it depends on the PC's domain name (and some more), and this appears not to be unique at all (in some environments). Now it is not at all about this, but maybe Novell (or NT or Citrix) make the same kind of mistakes ?

So again, if you could tell me what Workstation ID is in your case, and how this is influence by "code" ...

TIA,
Peter



>Hi
>
>The application was a legacy app originally written in DOS for Clipper. It was fairly common at one time for programmers to set up a semaphore system to assure multiuser access to records. When I became involved, the application had been converted to VFP3 with with the original architecture left pretty much intact. File and record locking were tied to the workstation which attempted to write/update records.
>
>We had unexplained workstation lockups and File Access denied. Not always the same machine and not reproducable. The problem would also seem to clear itself. After a lot of hair pulling we found that the Server was 'seeing' two different physical machines with the same workstation ID. Most of the time things would run just fine, but under the right circumstances the record or file lock was not being released. This only happened on Novell servers, no problems on NT servers.
>
>We solved the problem by adding code to make sure each workstation logon was unique.
>
>You probably don't have to deal with a legacy semaphore issue or even this exact issue. BUT the point here is that Novell treated record locking just differently enough from NT to cause the problem you described.
>
>
>- Don Lowrey