>Hi. We had a little problem at the resort Sunday night. I got a call that my 24-7 reservation app was completely down.
>
>After I arrived on the scene it became obvious that the RESERV.DBF file had been reduced from 280 MB to 25 MB. The CDX and FPT files were not affected.
>
>We restored that table and an associated table from a Saturday night backup and got the system up but the front desk clerks had a good bit of work to redo the days check-ins and check-outs.
>
>1) Has anyone ever seen this happen before? We are using local tables on a Windows 2003 server. We had some corruption issues years ago but I have never seen most of a file simply dissappear.
>
>2) What experience do folks have with continuous backup possibilities (or hourly etc.)?
>
>3) Going to SQL backend data storage would be an obvious improvment. It's on the list. :-)
>
>Peter

Bad 'single event' corruption happens even in the SQL world and, yes, tables can be truncated during a failed write. I wouldn't worry unless it is a recurring event; although someone still has to do due diligence and check the system out before you chalk it up to an anomolous 'bad luck' crash during a file write.

Ideally, you find out exactly why a file write failed and caused corruption?. But there is no guarantee that you will find the exact answer (how much time do you have to search? and is the equipment going to be available to you during your search or will they have to keep working on it? etc, etc). It's always going to come down to the vigilence of the client site administrator to reduce the negative affect the great 'mixing bowl' (the specific mix of hardware and applications they have chosen to 'run together') can play on the data. Have they installed some neat new screensaver recently? Or is some piece of hardware going bad? What crashed? Did a workstation die? Was it the server? No machines went down it just stopped working? Details!

Perhaps the file size has reached some critical mass for the system in question - buffers are being overloaded and failed. Or maybe it is a hard drive going bad or a network card has intermittent problems. Again, any number of possible explanations - hardware or software based.

Anyone charged with administrating a database system has no doubt faced these issues at one point or another. It's a unique problem at that site - perhaps others have seen similar things but there is no single, simple, 'answer' to fix a site with recurring integrity problems that other sites running the same database system simply don't experience. Gotta look at the big picture and imagine all the possibilities - then start testing each and every one. In my book that goes beyond the database developers responsibilities unless you have some sort of contract to do site administration or you can (and do) dictate the specific hardware and software 'mix' your software runs within.

Steve