AnsweredAssumed Answered

Filemaker Server 11 Stalls

Question asked by DaveBlackhurst on Oct 21, 2010


Filemaker Server 11 Stalls


FileMaker Server



Operating system version

Server 2003 or Server 2008

Description of the issue

Hi, we have developed a large filemaker platform for a client, hosted on a dedicated 2003 server with Rackspace.

This has worked perfectly for years until we upgraded them to Server 11 (from server 8). Now we experience once or twice a week server 'crashes'. From the client end its all fine until you try to access the stalled database - then you get a white screen and eventually have to end task to get out of it.

On the server side, that user stays connected in the list of Clients (even though they are not) and if they reload the system they appear in the Client list twice.

Most of the time I have to kick all the clients out with limited success, and end up having to end task the FM services to then reload the system. Not ideal and some data has been lost.

This problem has appeared to effect different files within the setup, so not pointing to any one problem database.

Some more information to add, we decided to upgrade the server to an 8 core, 4gig, 15k raid disks, 100meg connection to extract as much performance as we could from the FM server. This is running server 2008 64 bit. During the first week of roll out, FM server got stuck in exactly the same way on the new equipment. However on Server 2008 there are some better tools to see what is happening with the resources on the server.

It appeared that FMserver had eaten up almost all the 4gig of RAM - as when restarted all the server Ram became free again, we have the server cache set to 800meg. When all the ram was used we could see the server heavily accessing the windows swap file to keep up with demands for more memory, and am wondering if this leads to the eventual server crash.

Lots to take in, please ask as many questions as you think may help. It really has become an issue, as there are 130 users that use the system (normally 60 at any one time).



Steps to reproduce the problem

Server 11 on a hosted platform at a push

Expected result

memory leakage?

Actual result

memory leakage!