j. patrick campbell

I was sitting here happily listening to mp3s over the network and
chatting on IRC when BAM! everything just died. My friend on the same
switch was still on the net so i checked the switch and my box, all
cables in, all lights on, but the light on my pc's ethernet was solid
not flashing. I tried, could not ping my router internally. Not knowing
what else to do, i rebooted and everything is working. I would like to
know what to check to see why my network failed and maybe what steps i
could take to troubleshoot this next time without simply rebooting?

thanks,

~P
http://www.patrickcampbell.us
3 month old gentoo baby

Linux teh 2.6.6 #15 Mon Jul 19 17:18:45 EDT 2004 i686 AMD Athlon(tm) XP
1800+ AuthenticAMD GNU/Linux


--
gentoo-user@gentoo.org mailing list

john lowell

On Wed, 21 Jul 2004 22:42:19 -0400


Hi there, J. Patrick,.

If everything resolved with reboot, I'd be grateful and leave it at that. Your NIC may be showing the first evidences of trouble, I'd be particularly suspicious of that since you were unable to reach your router. If you have any type of dianostic program available you may want to run it. Rebooting is not intrinsicall evil, it's just overused. :-)

jlowell

--
gentoo-user@gentoo.org mailing list

j. patrick campbell

eep. this NIC is onboard my fairly new ASUS A7N8X-X


--
gentoo-user@gentoo.org mailing list

heiko wundram

Am Donnerstag, 22. Juli 2004 05:00 schrieb J. Patrick Campbell:


I have that same board, and my ethernet hasn't had any hiccups so far... It
_could_ be a driver issue (I once had this with an old VIA-Rhine based board
which would always dump my network after some time, so that I had to do
ifconfig eth0 down and ifconfig eth0 up), but I've not yet had this with the
NForce2 ethernet chipset (my box is up and running for about a month now).

What I do see is that I always get a kernel oops (sleeping in spinlock,
nothing really bad) the first time the card is initialized, but this only
appeared after I upgraded to 2.6.7-mm5 with my own patchset (mainly LIRC,
Reiser4, and no, I don't want to use Love sources, I patch my kernel
myself ), 2.6.4-ck2 with that very same patchset worked fine here, so I
guess that's certainly a driver thing.

Anyway, as I said before, what you might try to do is simply:

ifconfig eth0 down
(will reset the adapter to off state, but remember IP setting, etc.)

and after that:

ifconfig eth0 up
(will put the ethernet device back in action)

With almost all drivers doing ifconfig down will force a card reset, so you
should be fine here, in case everything was fine after reboot.

If you want to do it the "rc-script" way, try:

/etc/init.d/net.eth0 restart

This will remove all routes and all services that depend on net.eth0 (a lot),
and set them up again.

You have to call all commands as root.

Anyway, if those don't work, maybe try another kernel version, as I said, the
"reverse engineered" nforce ethernet driver for kernel 2.6.7 appears to be
borked (as I said, it works so far for me, just initialization goes somewhat
wrong), so maybe just using a different kernel will fix things for you... If
you're interested in my patchset, mail me offlist.

Heiko.

--
gentoo-user@gentoo.org mailing list

marco barreno

On Thu, Jul 22, 2004 at 06:00:47AM +0200, thus spake Heiko Wundram:

I sometimes get something similar to this, though with a vanilla
kernel.

I have a PCI wireless card on my laptop that uses the orinoco_pci
driver, and I occasionally experience a sudden stop in network
functionality. When this happens, I can't access the network at all,
and the ksoftirqd_CPU0 process starts using ~100% CPU (at nice 19, so
I don't notice anything slow down). To fix the problem I just take
down the network interface, remove then reload the orinoco_pci module,
and bring the network back up, and then everything works fine again.
I have a CPU monitor always running (gkrellm), so I know when it
happens because I see system (as opposed to user) CPU usage jump to
100%, in addition to not having network.

This problem happened much more frequently when I was using kernel
2.4.20 or thereabouts, and it happens very rarely now with vanilla
2.4.24 (I'm pretty sure that's the kernel version I have--I'm not on
that machine right now). It really only happens sometimes in the
middle of a very long network tranfer, near or over 100MB.

Good luck,
Marco

--
gentoo-user@gentoo.org mailing list

ian hastie

What is it you tried to ping? You LAN router or something in your local
computer? It might have been helpful to try pinging your IP address.

Well the first place to look is /var/log/messages at around the time you first
noticed the problem and a little before. This might give some hint as to the
cause. Some of the data you're lookin for should still be there. If you
could post any relevant info someone might be able to give a more informed
idea of what actually happened.

--
Ian.

--
gentoo-user@gentoo.org mailing list

benjamin allen

Hi

DOWN


I just upgraded to this board, and my onboard NIC (via-rhine) light is on all
the time, and blinks off when sending packets.

--
gentoo-user@gentoo.org mailing list