is this standard behavior?

[Rob Tito]

Hi Forum

I have a FMP 8 Database containing 6.7 million records, about 46 GB of space using MAC OS X (10.4.5).

I recently had a power failure while I was importing records from an external source.

Normally due to the size I replace the fault/corrupted database by its latest back-up and redo the action, leaving me with the least influence in lost records and so on.

This time I chose to let it "look for unused space"

It went through without needing to recover the database (nearly impossible on the size and failure is nearly guaranteed).

After completion it started up again yet to my amasement in stead of the 6.7 million records I found out the process deleted some 1.3 million records for no apparent reason.

Is anybody familiar with this rather fuzzy behavior and give me some causalities and sources or insight on what might be the reason for this rather distressing behavior.

And no, no records have been deleted only added.

TIA

Rob

[overrider]

i am not sure if you are looking at file corruption, but i recommend you replace it with your last good backup as you are used to. i wonder also what would happen if you take a good copy of your file and run the compact and optimize file routine on those. if i was crashing in the middle of any operation on my files, i would not dare to continue running them. as a matter of fact i have it sometimes that my filemaker server 5.5 crashes. always always i take an old copy and replace the files with it. even loosing two hours of work is better then to risk serving a corrupted file which gets backed up over and over again...

by the way, 6.7 million records is the largest filemaker database i have yet heard of. i personally have once tested one until 2 million records, but not more than that.

[Rob Tito]

The file replacement is my standard action ever since FM 1.

In 5 and 6 the filesize was limited to 2 GB and crossing that limit did pose problems without FM being able to cope with it, Recover seemed unable to help out there.

Since I noticed the recovery tools were rather meant for small databases I fled away from them since they had no real use: they simply didnot work or after over 24 hours of working a message: recovery failed. Thanks did I wait over 1 day for that message? Not likely. Ever since that emssage popped up I keep at least a double backup of the databases I use intensively.

Right now this large database nearing 7 Mrec has a backup in a Retrospect tape and two bloody copies on two different disks.

The file size is right now over 47 GB, and due to the really great indexing technology deployed by FM it is as fast as if there were only one million records present.

I must honor Filemaker Inc there, they did a great job by implementing the Rushmore technology into the database.

I wonder who else has experiences with large FM-databases and what are their experiences?

This database will be published on line when the target of 10 Mrec is reached, amounting to some 55-60 GB of data.

Hope to hear from anyone about large FM databases

Regards

Rob