I'm using Redhat Enterprise 4 on a Dell Poweredge 2550 server. There are two built in network ports on this system. Eth0 is gigE port and eth1 is 100bp. The problem that I'm having is that that after moving a large amount of data through my eth0 adapter (like 2 TB) it stops being able to see any network. I've tried restarting the network and that doesn't work. The only thing that does work is rebooting it completely.

Any ideas as to why this would be happening? It's a new server install to version 4. When I was using version 3 on this same server I never had this problem.

Thansk!

# ifconfig eth0 down
# ifconfig eth0 up

No luck? You shouldn't ever have to reboot your computer (sans kernel compile)

Poetics,

That's exactly what I think. I've tried making it go up and down and the only thing that helps is rebooting it. Quite a pain in the rear...

What're the results of an `ifconfig -a` after it goes down? Does it lose it's IP?

mii-tool -r -R
&
service network restart.

Thats it.

This is something interesting I see in my log file and I'm not sure if this has anything to do with my problem. Other then this I don't see anything out of the ordinary.


Fsmportmapper on host 127.0.0.1 no responding. retrying...


What is the Fsmportmapper?

Using the toold mii-tool as Amisharma suggeted I noticed that it reports both NIC cards running at 100BbaseTx-FD. Is that accurate?


eth0: negotiated 100BaseTx-FD, link ok
eth1: negotiated 100BaseTx-FD, link ok


If that's correct then the switch isn't allowing the GigE NIC to run full speed.

I just wanted to add to this that the 'ethtool' is replacement for 'rii-tool' for those that care. Using ethtool shows that correct reading of that NIC speed.

I have a similar problem .. but with a switch in my network.

After a few hundered GB thu it .. the darn thing stops processing packets.
I believe there is there a bug in the firmware that it ends up running out of buffers. You might be facing something similar, there is no way to make sure that's what's going on...

ifconfig can't help .. this is a issue one level lower .. try unloading/loading mod to see if it rectifies.
-

Thanks for all your responses... What it turned out to be is that Broadcomm doesn't work well with Redhat. After trying several of their fixes I decided to just disable the port on the server and install and Intel GigE card instead.

Try upgrading your broadcom driver from the site. Coz i had the same problem with broadcom driver with RHEL3 on my laptop as i had dual booting with XP too.
Upgradation did solved all the issues for me.

Google it.

Hi guys. This is my first post on this forum :-) So here is my problem. I have just got hold of this dell box and want to put my slackware 12 on it.

I have successfully booted knoppix on this box and checked my emails - so hardware is not faulty.

I did a fresh install of slackware 12 in the box and although the nic is seen and the irq is noted when the boot is taking place, when everything is finished, the nic is not allocated any irq.

I thought maybe the DVD I got was faulty, so I put a hard drive in another box and did an install, no problem, can check my email.

I thought maybe the on board nic in the dell was not fully supported so I put a realtek into a pci slot, turned off the onboard nic, re-tested with knoppix - still good, and went back to the slackware. Still not going. I only get the default local host 127.0.0.1 when I do an ifconfig. When I do any other searches for eth0 it says the card doesn't exist. As mentioned above, the nic is detailed in var/adm/messages with its irq etc correct.

Any suggestions?
EastWind (too)

I ran into a similar problem after installing Slackware 12.1 on a machine. Get to a command prompt and enter to see if udev has renamed your NIC. On my machine udev renamed eth0 to eth3. If this is the case on your box just open rc.inet1.conf and configure eth3 (or whatever comes up in dmesg) appropriately (don't forget to remove the configure info from eth0, assuming it's been filled in).