Connection failure to host: [TCP]<server>:<port number>:NODELAY;STREAM

SOLVED

Hello all..We recently upgraded from Sage 2018 Advanced to Sage 2021.  As part of this process we also spun up a new Windows Server VM (one the same Hyper-V host as the 2018 installation).  The old server is running windows 2008 R2 and the new is server 2016.  

Everything seemed to be running ok but after a certain amount of time people start to get the error in the subject.  And also at this time we get Connection Timeout waiting for your application to start.  Then if I shut restart the Sage Advanced Service (Yes it is running as a service) then we can launch it again.  And after a certain about of time it starts again.  So I followed the instructions listed here Here .  The firewall part of this article is not valid since the windows firewall is turned off on this server for all profiles.  I did go through and refreshed the files as per the instructions.  I also went through all the users and click the option to "Spawn Tasks from Application server".  Note that all the users did NOT have this option set.  I went through all of them and enabled them for all users.  There are easily more than 15 users hitting it at one time.

Now this is running on a high end speced Hyper-V host running windows 2016 and the VM it is on allocates a lot of resources (64GB memory, 4 VCpus).  Now we did finish the last 4 hours of the day with no issues and it does seem to be working ok but I am not convinced the error will not come back.  Time will tell tomorrow.

But are there limitations that would cause it to give these errors when a certain number of users/processes are hit?  It never happened on the Sage 2018 server running on a much lower speced 2008 R2 VM (again on the same hyper-V host).  Should we be running an application server as opposed to a service?  I read somewhere where if over 15 users we should run it as an application server but I think that is old advice and we had it as a service on the old system.  

I am curious if the spawn task on server will help or just mask the issue

Thanks for any feedback.  We will see how things shake out tomorrow.

Thanks for any feedback

Ray