eth0 NETDEV WATCHDOG problem

Hi!
I wrote last week, but i never get an aswer... so I reply here.

recently set up a small home server for my LAN. It run a SUSE 9.3 Pro
distro and offers the following services:
- web server apache+tomcat+php
- mysql server
- samba server to share some reiser fs directories from the server
- mldonkey file sharing server
- network antivirus

After installation I checked all and all seemed to work.
Since I have no place for another monitor in my PC room, I set up SUSE to
allow remote login via SSH, VNC, Web VNC, telnet, aiming to administer my
server from remote. All was working good.

The problem is that, after some hour of working (20 more or less), the
server stops to respond to ping or remote connection any more and it
suddenly become unreachable from hosts on LAN.

The first time it happened I though it was just freezed (yes, I suffer
the blue screen of death syndrome...); so I connected a monitor to the
server and verified that all the services was up and running. From
localhost I still could reach any service I set up, but not from remote
host on my LAN... From server localhost I can run KDE, play with console,
reach my "shared folders", connect to local web server and so on... From
remote hosts I simply can't...

ps -aux tell me all the services are running

ifconfig tell me that eth0 is well configured (IP / Subnetmask are ok),
but TXPackets 9747, errors 05, dropped 2596. Errors and dropped grow as
the time goes by...

I checked /var/log/messages.

Before the "freeze" it has just 2 lines:
syslog new configuration initialized;
<serverame> su: (to nobody) root on none

After the "freeze" it has more or less 1000 lines:

kernel: NETDEV WATCHDOG: eth0: transmit time out
kernel: transmit timed out. TXStatus 00 TXFrameID 00, resetting
kernel: 00 0bcca000 0bcca010 ... [cut 25 lines similar to this one]

Every 10-15 seconds, /var/log/messages add other 30 lines similar to the
above ones...

I think the guilty is my LAN card, but... what to do?
I bough it just for setting up my server, so it is 6 days old...
Any idea?
Is there some setting to change or is it broken?

The LAN card is:
D-Link DFE-550TX Fast Ethernet 10/100 Adapter

LAN configuration:
- 4 WinXP /2k clients
- static IP, class 192.168.1.x
- Linux server with no DHCP client/server enabled (192.168.1.2)
- ADSL router (192.168.1.1) with active DHCP server (this is needed
because my wife needs to connect her notebook to the lan, and the
notebook has DHCP client enable to fit her job LAN. So I have 1 dinamic
IP and all the others static, server too is static)


I thought it was a problem of ACPI / APM specifications, so two days ago
I rebooted the server adding acpi=off noapic parameters to the kernel at
Grub boot screen.

But the server's LAN card just freezed today...

Thanks in advance,
Ferdinando

ferdybassi@tin.it wrote:

> I think the guilty is my LAN card, but... what to do?

Bring it back, buy a real network card. (that is, don't buy a DLink again)
Buy a 3Com or an el cheapo Realtek.

--
Ruurd