Doug,

Thanks very much for your integring stories, nice to talk over a beer afterwards. BTW, the more of these things I hear, the more I start believing it (and as already "promised" in the earlier message, surely someone will going to test our cabling).

However ...
The exact same problem and resulting strange behaviour is at 4 different sites occurring. In addition, to my findings this is no real corruption in the sense of "mixed up data", shifted data, rubbish, etc.
Only when some general failure causes (e.g.) Novell to fail the same way always, this might indeed be it. But think of it : when this were the case, I'd say that Novell isn't dealing with something properly afterall (similar to allow the user to enter new items, the software not checking for it existing already and then blame the user).

Furthermore, supposed the cabling is the problem, how to clarify that it doesn't occur within NT server ? Remember, the one customer having it all the time like us, replaced all we both could think of (including the whole server like we did several time), but as soon as he moved to Novell, the problem vanished. Too much of a coincidence I'd say, and in addition I just can't tell how NT would be dealing with these kind of (cabling) anomalies better. But it could ...

During my 15 years of experience with the more heavy multi user (ERP) Fox environment, I think I've seen all kind of corruptions "available". For many years now we covered them all (lots of them are implied by Fox-bugs), and the only one remaining are lousy network IC cards (to my experience). But the last I encountered was some three years ago (couple of thousand users).

Thanks Doug,
Peter



>Hi Peter,
>
>Three scenario's I have come across over the years involving WIN95, WIN98 and WINNT4...
>
>1. A user running a multi-user FPD application (SBT) had irregular corrupted tables, indexes and loss of records. We pulled our hair out for 4 weeks. Finally got a cabling guy in to check the coax (no cat5 in those days) in the multi story office. Found the cable passed around a steel pole in the roof. When he questioned the owner it was found the steel pole was the 2 way radio antenna for the people occupying the lower floor. A shift of cable and everything ran OK for many years.
>
>2. Another user had similar problems always in winter. After close analysis we tracked the problem to days when the temperature was less than 5 degrees celcius and always around 9am (the time when all the offices in the neighbourhood started work and all turned their heating on). We tested the voltage on one of these mornings and found the office to be suffering a considerable brown out for about 20 minutes. The computers kept going, but the that's when the problems occurred. We installed UPS's on all computers and problem solved.
>
>3. And the last customer was in a multi storey corporate building in our capital city. Over a period of 12 months continually had niggling problems like above. Eventually got another cabling specialist in who tracked the problem to some cat5 being laid right over a fluoroescent light in the ceiling. A rewire corrected all problems.
>
>...I guess what I am saying, please don't neglet the cabling - in each case it was our last option, so now days I make it one of the first and only recommend certified phone cabling people to run network cabling. Clients like to use their electricians because they are cheap, if they insist, I walk away from the job.
>
>Regards
>Doug Johnston
>