network disconnection problem

[jheyrman]

We have a strange network problem on our LAN.

At certain times (randomly) our network clients seem to lose their network connection to a server, when this occurs they can't connect to one server but are still able to connect to other servers.
This results in many strange things and errors.

"Disk or network error" (MS-Access)
"Cannot save document you must select another file name" (MS-Word en MS-Excell)
"The file is in use by another user… " (MS-Excell)

We have the impression that the problem exists more when network traffic is low.
As a result of this we started pinging the servers every second from a client PC connected to the same switch as the servers. And we discovered that when this error occurs the server in question gives no replies on the ping command. (Sometimes for more than 15 seconds and this happens about 20 times a day)

We discovered now that when we open PcAnywhere sessions to the servers the problem disappears.
When the sessions are open the servers can respond to the ping command at all times and the clients don't lose their network connections to the servers anymore.

Does anyone know of this phenomena or knows another solution than keeping the PcAnywhere sessions open constantly.


Here are some specs of the things we have.

Servers:
Poweredge 1300 WinNT 4.0 SP6a (with Intel pro 10/100 LAN card)
Poweredge 4200 WinNT 4.0 SP6a (with 3com 3c980 server LAN card)
Poweredge 2500 WinNT 4.0 SP6a (with Intel 8255xx-based 10/100 LAN card)

Network:
Cisco Ethernet switch model 3524
All servers are connected to this switch and they negotiate 100Mbit Full Duplex.

Things we already tested but didn't help:
We tested with other switches (baystack, 3com) and hubs (Intel)
We changed the cabling.
Forced the auto-negotiation to 100/full 100/half 10/full 10/half
Other network cards (in clients and servers)
Checked out the temp directories on clients (MS KB: Q150943)


Best regards
Johan

 

[Required]

Johan,

Here are some questions and ideas.
When this problem happens, is your server (the console) locked, is it on the login screen, is there an open interactive session or some kind of screensaver ?
Did you check the energy saving setup ? PCAnywhere might be the one who doesn't let your server go to sleep...
You might try using the Performance Monitor to record the main activity of your server (CPU, memory, disks and network) and try to see if there is any change when the problem appears.
You could also monitor all the network traffic to and from this server, so you can see if there is no activity at all or if anything is going on when the problem appears.

HTH
Please keep us informed.

 

[jheyrman]

We have no screen saver's.

The problem happens in all situations, except in this situation.
If we open PcAnywhere to one of our servers and the server console is unlocked in an active session, then the problem doesn't occur.

Energy savings in NT4 server ??


Best regards
Johan

 

[Required]

Effectively, NT4 doesn't support Energy savings. I should have verified.

What can you tell us about the server and the network activity when the problem appears?
Is there a peak in the CPU or disk or network activity?
Does the server become completely silent on the network or does it receive and or send some kind of traffic?

 

[jheyrman]

There is no huge change in the cpu disk activity, we even measured a network disconnection when the sever was 100%idle.

During the error there is no network traffic is possibole to that specific server. And a this moment nobody on whole the network is able to access that specific server.

We also checked the error log on our cisco switch, but during the error there is no network traffic and the error counters don't count.


Best regards
Johan

 

[jheyrman]

Another thing that i just discovered when we let server X do a virus scan to server y then server x wil not give any errors during the virus scan just like when we open a PcAnywhere session to server x.

this makes it even more strange.

 

[TheLad]

What network speed and duplex have you got configured? Try to aviod full duplex if you have it configured. -----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------

 

[jheyrman]

Thanks for your response TheLad
the answer is already in my problem description

Network:
Cisco Ethernet switch model 3524
All servers are connected to this switch and they negotiate 100Mbit Full Duplex.

Things we already tested but didn't help:
Forced the auto-negotiation to 100/full 100/half 10/full 10/half
We did this on both sides as (SERVER/SWITCH)

 

[Required]

So, it is just as if the server disappeared from the network when it is not involved in frequent traffic.
When PCAnywhere or the Antivirus are running, they maintain a constant flow of network traffic and everything works fine. But when your server does not receive regular network packets, the problems appear.
Did you verify if the packets from the clients are routed to the server when the problem appears or if they are not routed correctly? I mean, does the server receive these packets and ignore them or doesn't it receive anything?

 

[emjaydee]

Daft question, but have you checked for duplicate IP addresses. We had a similar problem to this, and it was caused by a bay hub that had a mgt module in it, which had the same IP address as the server. It was difficult to track down, because the hub only seemed to advertise it's address at random intervals

 

[Required]

You might also have a look at this thread:

http://www.tek-tips.com/gviewthread.cfm/lev2/8/lev3/58/pid/558/qid/199692

 

[Required]

This page from Cisco might also be helpful:

http://www.cisco.com/warp/public/473/46.html

 

[jheyrman]

Many thanks for all those responses but we have the same problem with a baystack switch, and even with an ordinary hub.

We tested about 5 different NIC's in our servers (some 3com cards, intel, d-link, ......)

Best regards
Johan

 

[soundguy]

Hi,

I have a similar problem.
One of our NT servers 4.0 Sp6a keeps on stopping the NETLOGON service automatically. The error logs only mention that the network has forced a re-election. No reason why? and no explination why? dam strange. This is the PDC by the way, so sometimes it poses a bad problem.

I still haven't found a solution!

 

[Guest_imported]

this problem is related to net-bios. for some reason, windows 2000 and NT have trouble co-existing. iv'e been able to get them connected back together by running browser utilities (like browstat) that force the server/client to dump old/erroneous information and re-load new/accurate information. i still don't understand what the problem is but i am fairly certain it's due to the (porly conceived) way that microsoft implements network browsing through net-bios. for more information, check microsoft's description of the network browsing service at:

http://support.microsoft.com/default.aspx?scid=kb;EN-US;q188001

and good luck.

 

[Guest_imported]

Look into Johan suggestion carefully, you do not have to have screen saver on to have the energy saver on.
Check to see if the server was set to go to sleep or Turn off hard disk after so and so minutes "Its in the screensaver tab under power"

 

[jheyrman]

Up

 

[rjbj]

This may be silly but, I know that NT does not support power saving but some Bios programs go into a sleep mode if the machine is idle. They kill everything even the NIC card. Dell is notorious for this. I guess it's worh a shot. You have tried everything else.

 

[akarshs]

rjbj is right, just check powermanagement option in Bios.