Hi there,

We have an FMSA11 server running on Windows 2008 64 R2 in a datacenter ( HP Proliant Intel Quad Xeon / 4GB / 1TB Raid Disks ).

We serve an IWP and Filemaker Go solution to education clients across the UK. Normal server loads are fairly light with around 10 - 30 concurrent users, all accessing the same database to upload images from filemaker go (current file size is around 800MB which is mostly pictures) and from IWP to view and evaluate the uploaded images and comments.

Just the last two days, the server seems to be stopping serving requests, clients are experiencing some or all of the following issues: -

- Cannot upload images to containers

- Cannot log in via IWP (Web server is OK but no response or login window)

- Server Statistics shows massive elapsed time / call, all other stats look good.

- Client Stats shows multiple clients with big elapsed time during the 'lag outs'.

- Scheduled hourly backups sometimes giving 'Backup failed due to lock conflict with database users or other error (8003)'.

Stopping & Starting FM Server and Web Publishing engine gets things going again.

We have performed a consistency check on the File and it checks out ok.

Apart from the occasional backup failures (which also never happened before), there are no other errors in any of the FMS logs.

Has anyone seen this behaviour before or have any suggestions?

Any advice much appreciated.

are there any DMP files in the FMS logs folder?

Double-check that absolutely no other process is running on that server (virus scanning, indexing,...)

Are the physical files on the SAN as opposed to on the FMS box itself? If so, get the guys in the data center to troubleshoot any SAN issues.

Of the stats that you mention: which ones show the highest numbers? Is it I/O?

Hi, I found this while looking for why my hourly backups started failing - I do have DMP files in the FMS log folder. What do they mean and what do I do with them as they are not readable by notepad?

The DMP files signal that FMS encountered an internal error that it could not handle. It's consistent with your symptoms. It's something in FMS that manifests itself only under load (a number of users concurrently doing some taxing for FMS like performing a search that involves unindexed fields, large found sets,... this can include implicit searches like showing lots of related data in portals, multiple portals per layout)

Before you do anything, install the patch for FMS11v4 that was released two weeks ago. It specifically addresses a known issue in that area.

If the problem persists, check to make sure the machine and its deployment conform the the FMI requirements and best practices. With that info in hand contact FMI support.