Since updating to FileMaker Server 17 (Windows*), we have experienced a number of instances were we lose control of the server and need to reboot the hardware. We have a support ticket in with FileMaker, but have not to date been able to isolate the problem. It seems like memory usage is normal (2.5-3.0Gb) then suddenly spikes up to 8Gb, at which point users start reporting issues and we lose remote access to the Server.
With this occurring as recent as last Thursday, and then again today (Monday), I have desperately been digging into the access and event logs, and think I have isolated a pattern.
I noticed that leading up to the last crash (at 8.48pm) last Thursday night) one user had 33,425 "single sign-on authentication failed on database" entries in the access log, starting at 4.06pm and continuing right up until the crash at 8.48pm:
SECURITY: Client "Xxxx Yyyyy (87ZMJ12) [10.22.130.154]" single sign-on authentication failed on database "HR Database Menu.fmp12" using "NurkF [fmapp]".
With this going on, a scheduled script that runs at 8.45pm each week day, failed part way through.
It seems that we have 5 or 6 users (of around 50) for whom the logs frequency show a spate of "single sign-on authentication failed on database", in the lead-up to a crash - in the tens of thousands of entries.
This has only occurred since we updated from FileMaker Server 16 to 17. We have been using external authentication for many years with no issue, and we have not changed any of the security models or AD Security Groups. The scheduled scripts that run run without issue for years without issue.
And I cannot identify anything unique about the users for whom this is happening.
Would anyone have any idea why this would be happening?
* Microsoft Windows Server 2016 Standard, running on VMware: Intel Xeon CPU E5-2660 v3 @2.60GHz, 8GB RAM