Problem: FileMaker 15 WebDirect clients get a blank screen on login after the server has been running a while. There are up to 30 concurrent users from different countries logging on to the system.
Setup: We are running FMS 184.108.40.2068 on an AWS t2.Xlarge installation running Windows Server 2012 R2 base (4 cores, 16GB ram).
A second AWS server (t2.medium) runs FileMaker Pro Advanced 220.127.116.11 and is logged onto the server based file. This machine uses 360works RemoteScripter to listen to a port for a call to run a pdf generation script that takes about 40 seconds to process per request.
Since this was setup last week, it ran fine for 3 days, then we had problems with WebDirect clients logging on and seeing a blank screen. One quick fix was to turn off the Web Publishing Engine, then turning it back on. That works for about 90 minutes, then the problem re-occurs. Restarting the AWS server solves the problem, too, but I've only done this once since the problem just started in the past 24 hours. The robot client does not have any connection issues - it stays logged on all the time.
The FileMaker server logging system does not throw any errors when this happens. Windows Event Viewer does throw an error while this is happening:
A fatal alert was generated and sent to the remote endpoint. This may result in termination of the connection. The TLS protocol defined fatal error code is 10. The Windows SChannel error state is 120
I tried researching this, but was not able to find an answer to the cause.
Background: This AWS server is located in the AWS London region. The users are in the London area as well as in South Africa. The owner was using a MacinCloud server up to last week, but having a lot of bandwidth issues causing users to have very slow response times to the point the system was unusable. That setup had FileMaker Server 15 with a FMPA 15 both on the same machine. When the system was responsive, the robot successfully saved the pdf and uploaded it to a container field. This process took several minutes, and during that time all the other WebDirect Clients were able to continue accessing the database normally.
In an effort to fix the slowdown problem, the client upgraded to FMS 16 and FMPA 16 on the same machine. The problems persisted, so the decision was made to switch to AWS.
The printing still took the same amount of time, but the big difference was that the WebDirect clients are now locked out while the robot generates the pdf and are disconnected, as well. Moving the Robot to an different AWS server (the current configuration) improved the PDF generation time from several minutes to about 40 seconds, and the WebDirect clients are still locked out, but they are not disconnected.
To summarize, when the system ran on the Mac server, the WebDirect users are able to continue using the software during pdf generation. When running on Windows on AWS, the WebDirect clients are locked out from using the software during pdf generation.
So, two issues: logging into a blank WebDirect screen and users get locked out during pdf generation. They may be related.
Any help would be appreciated. This is a live system with unhappy clients, so time is of the essence.