I had a situation like that many years ago, possibly Server 9 but mybe before that. The issue was a corrupted file. I had to find a clone and import the data from the corrupted file. That fixed the issue.
Think this is likley to be the isssue you have, sorry
We are running a pretty large number of FMS's at our own premises, for customers, and at our customers own servers.
We had some problems like the one you describe years ago. Quiet a few of them with Server 5.5 and FileMaker 5-5.5-6 - and a few problems with FileMaker 7-8.
Since fms 9 the problems seems to have died out. 10 and eleven seems to be extremely stable.
But if you believe that you have reason to fear this to happen again on you installation I would definetely reinstall the server - and maybe even the OS.
But if you want to ensure that this can't happen again you could set up a serverscript updating a log-entry with a timestamp every, lets say, 5 minutes. Then call this informatioin via http/xml-php and "ring the alarm" if the update is delayed more than 7-10 minutes.
Is this helpfull
Ps. we have never seen problems like this with FileMaker 9, 10 or 11 (client and server).
I'v just had this issue this morning. 9 days of data are gone .
maybe I missed that info in your post, but can you tell what FileMaker Server version is in question as well as what OS are you running it on?
Must say: Very important question. But it seems that this issue died out. But yes, lets hear about
OS, FIleMaker version and updates. Hardware and setup (Virtual/Physical etc).
Details are as follows:
HP Compaq ML350 G4p Server
Windows Server 2003 R2 SP2
Filemaker Server Advance 22.214.171.1249 (Now I dont think it has been updated since the issue)
54 Hosted Files
No Server Plugins
700Mb Cache Reserve
Physical setup with 3 x 72Gb U320 drives in RAID 5
What about the backups that were run while the data was not being written? Are those backups complete? Damaged? Do they contain the new data?
Also: when you backtrack through the Windows event log (application or system) are there any warning or error events logged around the time that it starts? If there are you can use a monitoring tool (there are some free ones) to send you an email when it happens.
Same applies with any dump files in the FMS logs folder. I've seen situations where FMS creates a DMP file in its log folder when the problem starts, so that could be an early warning that you can pick up on.
How much physical memory in the box?
How many processors or cores?
As an aside: RAID 5 is not the recommended RAID for anything that is disk intensive like databases. Go for RAID 1+0.
Dual 3Ghz Intel Xeon and 3Gb RAM
This is an old machine that was used to take the load off the main server and 1+0 wasnt an option.
Both are being replaced with an ML350 Gen 8, 24Gb Ram, 2 x 76gb 15k drives in a mirrored array for the OS and 4 x 300Gb 15k Drives in 1+0 for the data when we go to FM12 in December.
No backups as we perform a full backup overnight only.
Off the top of my head I am sure we checked the logs and found nothing, it was over a year ago now so nothing to look back at and not in a hurry to recreate the problem.