Has anyone else experienced files that pass the consistency check on some computers but not others?
Yes. We discovered that different OS will give different recovery results. It was four years ago but at that time FMI recommended try as many different OS as possible. In the end we got a complete recovery using Vista.
If it's repeatably different on each machine, or if the live file fails and backup passes, then it's a bug, and should be reported
It would also pay to run a disk utility to check each hard drive
Since the Consistency Check is just looking at blocks and pointers, the OS ( hopefully ) should not matter
( If the Recovery won't finish, THEN it would pay to try another OS. That has always been true )
Try a Save As Compacted. It rebuilds the file's block index, and so can fix consistency problems
> Has anyone else experienced files that pass the consistency check on some computers but not others?
> FMI recommended try as many different OS as possible. In the end we got a complete recovery using Vista.
Thanks for the tips. Unfortunately, they didn't provide the results I was looking for. Good try though.
Well, after quite a bit of running about, I was able to hit on a solution. Depending on how the file was copied to which machine, and then what activity I may have done to the file after arriving on the new machine (thanks again for the compacted tip), I was able to get file recovery to fail on every system. No, I don't consider this a failure.
What did end up happening though was I was getting different recovery logs depending on the computer I was running recovery on, still not sure why, but the one system that was passing then finally failed after saving a compacted copy and then running recovery pointed to a error in a calculated field. Since I was playing with multiple copies at that point anyway, I deleted the calculation and with that calc gone, the file past recovery on every machine, repeatedly, even after saving copies, saving compacted copies, uploading to server and back down again. Everything in the file appears fine now.
It seems that failing recovery for a bad calc is a bit harsh. I've kept the older backup version online, but I'm building a new solution currently for the client with a bunch of new features for them. I'll migrate their data over when they're ready for the big move.
Thanks everyone for your ideas and support. Hopefully this gives others the encouragement to not give up and keep digging a little deeper into the recovery logs for possible answers.
The Recover Log was originally an internal tool, that FileMaker fortunately decided to make public
But without proper documentation, it just ends up scaring the hell out of people
They need to publish a list of errors ( like this ) that are of "low severity"
> It seems that failing recovery for a bad calc is a bit harsh