Randomly kicked out of active session

This is the problem.  Client with 20 users is complaining that some of their users are being kicked out of sessions.  Users will be kicked out of Sage while entering a document such as a PO.  Or kicked out while browsing for almost anything.  When the local network group reviews user logs, it 'strongly' suggests that they were disconnected due to an application crash.  It seems that this happens very frequently.  This is not happening to all users and seems to be confined to those using the operations modules of IC, OE and PO.

The client is operating in Azure environment with a SQL server and 2 vmhosts.  The full Sage300 programs loaded on the hosts.  They are using Sage300 v2025.  SQL server is Win2022 datacenter and SQL 2022.  Server has 32 gigs of ram.  Vmhosts are Win11 Ent multisession with 32 GB.  I am not sure what each user is allocated when they log in.  The client does not access Sage locally.  They log into the vmhost and then launch Sage from there.

I know this is quite vague at the moment but the client did not experience this type of problem when using an onprem server.  It has been suggested by the client that google searches show that this is a common problem.  It is not common to me and I'm looking to see why this is a Sage problem.

Is this a common issue??  Have others experienced this??  Is this a resourse issue on the workstations?? or Server??

Any help would be appreciated.  Thanks.

Terry

  • 0

    Are the users seeing any message at all when this happens?

    We had one customer with a weird intermittent error that I eventually solved but it took daylight savings to figure out. The user would enter a transaction and get a SQL disconnect message. This user was often the only one impacted. What we discovered was that it happened at roughly the same time each day when she came back from lunch. No windows logs recorded any issues. When we changed the clocks in the spring, the time frame of the error moved ahead one hour as well.

    So I ran a ping in a batch file that recorded the ping to the database server results every ten seconds.

    On Wednesday, at about 12:30pm, or 1:30pm after "leap forward", the database connection would be severed on VM #5 (they had six VMs for load balancing).

    Sage won't tolerate a break in the database connection for more than about 30 seconds. If you have left a window open and the database connection is cut from 'underneath it' then you'll see an error when you make the screen go to the database. If you just close the screen and open it again you'll be a refreshed DB connection.

    What was happening was that the mid-day image backup of VM#5 was failing and that caused the timeout. Because this was happening "outside" of the server, nothing showed up in any of the logs we had access to. We finally got the attention of IT with specific proof that this VM, during this time frame every week, would be unable to ping the database server for about five minutes. They saw their VM logs shows problems at that time. (This had been going on for months and months - IT company was fired since they couldn't be bothered to look since it was a 'software issue'.)

    So - my point is, do the users see an error message when the PO/OE screen closed up? Is so, does it indicate that there was a SQL client communication error? It's very rare for windows to close without any dying last message. And when that happens - I like to make sure that the video card drivers are up to date. I've seen updating those fix weird crashing issues.

    Good luck!

  • 0 in reply to Django

    Thanks Django. 

    The users are not getting any error.  The screen just closes on them.  I have asked them to document the problem with more detail so we can see if there is anything common.