Navision users are being dropped after 4 minutes of inactivity with an ECONNRESET error.

We are running a single server instance of Navision 2.5 on Windows SBS 2003 with SP2. We are using database.fdb.

This has been running fine for almost a year. Now we get some kind of 4 minute time out happening.

We’ve been running cports.exe to see what’s happening.

User are connecting on port 2407 and as long as they are actively using Navision, they stay connected. If they take a break, the port gets dropped after about 4 minutes of inactivity. When the user tries to do something again they get the ECONNRESET error and have to shutdown and restart the Navision client.

The server is a Dell SC440 w 4GIG Ram. The LAN card is an Intel(R) PRO/1000 Server Adatper with a static ip address.

I’m looking far any ideas on how to fix this as its a consistent, repeatable error that’s getting quite annoying.

Native Server timeout has defined by value KeepAliveTime in TCP parameters located registry.

Check you current value in HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters

Do you have any suggestions on what the KeepAliveTime value needs to be set at?

Also do any of:




make a difference?

Thanks - Al

I guess the other things we need to know are:

Are the registry changes done only on the server machine or are they needed on the workstations also?

What about workstations not running the Navision client, do they have to be changed as well?


You set it on the server. In you aprticular case you probably want to set it to zero.

So far we have set the server KeepAliveTime to 2 hours 2 minutes and zero. Restarted Navision, rebooted the server, and still no difference.

The users are still being dropped after 4 minutes of idle time. Its like the KeepAliveTime value is being completely ignored.

Would appreciate any other ideas even ones that include waving rubber chickens over the server ( or maybe not :).

That’s a bit silly isn’t it? Any sacrifical rituals involve blood, so I think plastic ones wont work. (Maybe you could use ketchup and the servers wont know the difference).

My next guess though would be that someone has installed or reconfigured a firewall somewhere. Do a trace and see if packets are being routed somewhere along the way to the server.

Resolved YAY!!!

The same Server we are running Navsion on is also running Wingate Proxy Server. It turns out that Wingate 6.5.2 had known problems with timeouts.

Upgraded to newly realsed 6.6.2 and suddenly our Navsion timeout probems vanished.

We couldn’t understand why none of the registry value changes we did made any difference at all.

We managed to cancel the order for rubber chickens and ketchup.