>
>That's fine. I know enough about doing support to realize that I can't afford the cost of keeping a fully-qualified CNE on staff for my small shop (Weatherhill) so it's a whole lot cheaper to pay people who know what they're doing when the need arises. You may find that stiumbling about with a production problem for months on end and avoiding paying the cost of bringing in a knowledgable and competent engineer is cost and time-effective. I know that we lose money when our production environment doesn't work right, and the cost far exceeds what getting things fixed will cost.
>

well fortunately, the problem seems to occur only once every couple of weeks and the quick fix is a 2 minute restore of the knackered file. If it escalates, then costly intervention may be warranted.

Having said all that, we seem to have narrowed the problem down to a single machine (although I need another couple of instances to confirm it and that might take a month or so). The workstation in question has on the last two occasions at least been the first to experience the problem and until it has logged on the other stations have been happily beavering away in the data. The offending machine exhibits some unconnected symptom, following which it appears to corrupt the index... Once I confirm that it is that machine, I'll give it the wire brush and bleach treatment, rebuild it if necessary and see if that sorts it out. My experience, so far, with networks, is that 99 times out of a hundred its a problem with a workstation. Especially if they're workstations running any version of windows!

Thanks for your help and guidance - if and when we have to reinstall the server I'll be sure to make use of it!

Harry