I'm hoping that somebody may be able to give me some insight as to some Filemaker Server problems we keep running into.
Our Filemaker solution consists of about 8 databases that equal out to about 30+ GB. At least 4 or 5 of these databases have 2 -3 million records and one has 19 million records.
We have on average about 22 - 23 users connected at a time.
The server we have right now is a Dell Poweredge R810 with 4 Xeon e7520 4 core processors, 16GB of RAM, a 400GB RAID 10 array for the database files and running Windows Server 2008 R2 and Filemaker Server 12v4.
Our users connect using Filemaker Pro 12v4 and our database developer uses Filemaker Pro Advanced 12v4 remotely from his home.
The problem is a tricky one to explain. What normally happens is that when we try and run a few different scripts or different finds, the server RAM usage will dramatically rise from say, 8GB to around 13GB and just stay there. The more scripts we run the higher the RAM usage gets until it completely uses 99% of the memory and Filemaker Server hangs up until I close the Filemaker Server service. Then the RAM will shoot all the way back down to around 2GB.
The second problem is that when the Filemaker Server gets hung up, the Elapsed Time Call will go up to around 14 million and stay there until I close Filemaker server. When that happens, everyone will get the dreaded coffee cup and I will have to forcefully disconnect everyone and even then the Elapsed Time Call will still be in the millions even when no one is connected.
This doesn't happen at a fixed inteval either. It could be fine for 3 days then the 4th day it could hang up. Or it could happen twice in one day. This makes it very hard to pinpoint what the problem is.
I will be upgrading the RAM for our server to 32GBs sometime next week to see if that helps anything but there seems to be an underlying problem.
Could the problem be in the scripting? A network issue? Hard drive problem?
I'm not really sure what to try first, and I have to run those scripts every day as it puts important information into our databases.
Any help you could give would be wonderful.