I have a small setup of Navision 4.0 (5 users), recently upgraded from ver 3.7 after using it for two years.
In my typical setup, I have a Linux based Domain server running DHCP and Mail Server. Navision is installed on a win2k machine which is a DHCP client to the Domain Server, so are all ther Navision clients.
I have recently started getting TCP/IP error: ECONNRESET on clients. Sometimes the client moves on after error report while at others it has to be restarted. What is DANGEROUS is that many a times data is lost and database gets corrupted.
I checked my network (approximately 21 hosts+servers) and it SEEMS to be ok. Other applications running on the network do not report any error, all stations PING correctly.
I am bewildered and not seeing any direction as to how to tackle this problem. Could I have some helpful tips please ? [:)]
I would bet that it is a faulty network. Most likely a bad network card, or maybe a Switch somewhere.
Basically Navision tests packet very thoughroughly, and if it finds a bad one, will generally clse the connection. Also asto the database corruption thing, I have a lot of experience with this, and I woudl say that more than half of all database corruptions are because of network problems, NOT because of hard disks. In fact I have only every once confirmed two Database corruptiona in Navision as due to a Hard Disk failure, one due to a bad sector, the other due to a SCSI cable terminator installed in the wrong position.
What happens is that if a network connection is bad, it can start retrying, and there is a very small chance that after a numebr of retries, corrupt data passes the check sum and gets onto the database.
You first thing to do is to get some packet monitoring hardware (not software) and track down if there are any corrupt packets transfering across the net. I am pretty sure there will be.
Forget PING, that would be like Michal Shumacher saying, “Yep I can see that the Engine is installed, so the car must be perfect”.
One important thing!!! Don’t get tied up into testing the machines that are disconnection, it could be any machine on the network sending bad pakets, you need to find out the problem. It could be anything as silly as an IP printer box causing this porblem. But Iwould put 50% chance that its a net work issue.
I tried searching for Packet/Network Monitoring Hardware as you had suggested but was not able to get any leads. Could you please tell me what company / brand can I search for this product ?
I tried searching for Packet/Network Monitoring Hardware as you had suggested but was not able to get any leads. Could you please tell me what company / brand can I search for this product ?
Many thanks for the inputs. The Navision Server is on Static IP. The clients, however, are still on DHCP.
I have added following lines to my Navision clients:
Hosts File: < ip of Navision server > < Name of Navision Server>
Sevices File: < Name of Navision Server > 2407/TCP
Things seem to be running but I cannot say whether this took care of the entire problem as it occurs occasionally. I’ll have to wait for some time before I can say that THIS was the right solution.
That could resolve the problem but I would still take notice of David’s suggestions.
My own workstation once had a faulty network card that created errors of the type you are seeing, never got to to find it until I replaced the PC.
I guess you are right. I am still having the problem after doing the settings I last described. It does not happen everyday, but when it happens, it happens multi-times in a given day. This is how it used to be earlier too.
Suspecting a network element responsible for this, how to pinpoint at one is what concerns me.
I do not get to see any collision indications and no other tell tale signs which point to the faulty element. Can someone suggest a measure, software based or hardware based which I could use to fish this bad network element out ?
I have a network of about 20 PC’s, 4 IP Cameras (LevelOne), wireless access point (Netgear), two switches (D-Link) and a DSL Router.