Network Card Error, ipconfig report error on eth0

[Marco A. Ramos]

Thank you very much Pete, I really appreciate your help.


As you said, I believe too, that the trouble could be the buffer, actually I
found other forum “when a card is in a fast network and it has few buffer
memory that card can drop packages”.

Well, the solution looks to increase the buffer memory, but I don’t know how
make it, and that failing is get in me in troubles. So now we will switch
the IP’s and cables, the eth2 winch is working fine we will put it on the
fast switch and the eth0 we will put it on the router with a slow connection
(128 kbps), may this is not the final solution but I can win time because
that erd. connection it is not operating now.

About your recommendations, the big problem is that I’m like 1,500 miles
away from the server and the server is opering a critical service (yes I
know, why don’t you test it fist?, but we tested it and work fine, just when
we put it to operate star to fail).

Change the port on the switch, I can’t do that at lest a probe that it is
failing, because it’s assigned by a government department.

The Nic is motherboard integrated, actually this is the strange because the
motherboard has two similar NIC integrated but just one fail!

Put other Nic on the server, the server is Rack Mount and it use small PCI
cards, I don’t have other now (bad bad bad), but it’s good idea to buy other
to have it as backup.


Every help or comment is really welcome!!!


Thank you all.

-----Original Message-----
From: redhat-list-bounces at redhat.com
[mailto:redhat-list-bounces at redhat.com]On Behalf Of Pete Nesbitt
Sent: Wednesday, 27 October, 2004 7:57 PM
To: General Red Hat Linux discussion list
Subject: Re: Network Card Error, ipconfig report error on eth0


On October 27, 2004 06:27 pm, Marco A. Ramos wrote:
> Hello every one,
>
> I have a RedHat server with 3 network card, two are working fine, but the
> other (eth0) is having problems, if I use ipconfig I can see many error
> report it.
>
> This problem card (eth0) is equal then the eth1, both has the same driver,
> and similar configuration.
>
> This networks are just use it for transmit and receives small plain text,
> so the problem card is working but block all files greater then 700b and
> some smaller pass but get empty.
>
> I'm looking for differences between the eth0 (bad card) and eth1 (good
> card) and I just found one (I don't see IRQ collision), when I check the
> file /proc/pci the eth0 (bad card) has after Master Capable. No bursts.
and
> the eth1 (good card) has Latency: 32, I believe that these is the problem,
> because I saw many other configurations and all have Latency: and number,
> but I don't found information about what it mean "No bursts" and how
change
> it to Latency:32.
>
>
>
> Any body know some thing about it?
>
> Thank you so much,
>

<...snip>

Looking at ifconfig for eth0:
RX packets:5829 errors:3863 dropped:0 overruns:0 frame:2169

The bad frames would appear to be a key. The nic may have a bad buffer of
something. Is there a switch at the other end, you could try a different
port
in case the switch port is sending bad frames.

Do you have a different machine to try the 'bad' nic in? (or another nic to
try in that box)?
--
Pete Nesbitt, rhce

--
redhat-list mailing list
unsubscribe mailto:redhat-list-request at redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/redhat-list