Not sure if anyone else has come across this issue as can't seem to find anything about it on the web.
We are running two FM Servers with numerous databases on each.
Twice in the last year (on the same server) we have had an incident where we will get a user crash out that causes the server to stop writing anything to disk, but critically shows no signs of the problem. Work will carry on for hours with the server just caching the data changes. Eventually something will happen that will call the error to light, in this instance another user crashed out and locked a record.
You cannot force disconnect any users, and you cannot close any databases that have changes outstanding in cache. They simply say "Closing" but never actually close.
The only choice you have is to restart the server, prompting file verifications upon restart and a loss of data to the point of the original crash.
An example of Monday's crash:
11.28am - User gets a spinning blue circle and eventually has to crash Filemaker on his machine.
3.30pm - User crashes locking a record brining attention to IT Dept.
3.45pm - After failure in closing databases and disconnecting users - Server restarted
After restart and file verifications ALL files are complete and acurate up until 11.28am - changes/additions made after this time are lost.