Intel 82566DM Gigabit Lan can't get IP after update/reboot

Confused · Post by **Confused** » 2008/06/18 17:39:10

Everything was working fine before the last bunch of updates I did and rebooted the system.

After which, networking no longer works. Physically, the onboard NIC appears to be ok, there is light and the corresponding port light on the router blinks when I bring eth0 down and up.

Searching around, I found some info on commands that could be used to diagnost the problem and they all seem to indicate that the NIC is working fine mostly.

The following are manually keyed from another monitor since there's no network so pardon me if there's typo.

[code]
mii-tool -v
eth0: negotiated 100baseTx-FD flow-control, link ok
product info: vendor 00:aa:00, model 57 rev 0
basic mode: autonegotiation enabled
basic status: autonegotiation complete, link ok
capabilities: 100baseTX-FD 100baseTX-HD 10baseT-FD 10baseT-HD
advertising: 100baseTX-FD 100baseTX-HD 10baseT-FD 10baseT-HD flow-control
link partner: 100baseTX-FD 100baseTX-HD 10baseT-FD 10baseT-HD flow-control
[/code]

Seems to indicate to me that physically the link is fine, the NIC can see the router and they have happily negotiated the physical parameters of the link.

But
[code]
dmesg | grep eth
e1000: eth0: eth1000_probe: Intel(R) Pro/1000 Network Connection
ADDRCONF(NETDEV_UP): eth0: link is not ready
[/code]
Last line repeats a few times

Doing a "service network restart" will [FAILED] at "Determining IP information for eth0"

I've tried to set a static IP for eth0 but it's not working either. Pinging to or from the machine just doesn't work even static IP set. I've changed the network cables as well as swap a router but nothing works.

Would greatly appreciate any advice on what else can I do to fix this. Thanks!

p.s. running on x86_64 version, originally on a CentOS 5.0 install, patched until 2.6.18-53.1.21.el5xen

gerald_clark · Post by **gerald_clark** » 2008/06/18 17:42:59

What does uname -r show?
Try booting the previous kernel.
Now what does uname -r show?
Does the nic work?

Confused · Post by **Confused** » 2008/06/18 17:51:54

I've booted into a LiveCD and posting this from the machine so I guess that confirms that the problem is wholly software/configuration problem with my CentOS installation.

Going to reboot into CentOS again and get the uname info again

Confused · Post by **Confused** » 2008/06/18 18:04:45

uname -r
2.6.18-53.1.21.el5xen

However, it seems irrelevant now because networking has inexplicably started working again...

I would really like to know why on earth it didn't work after the first reboot but I guess it's one of those things that would forever stay a mystery. At least I certainly hope I don't have to deal with it again!

AlanBartlett · Post by **AlanBartlett** » 2008/06/18 19:24:50

Just one question for you, [b]Confused[/b]. Do you [i]need[/i] to run the [b]xen[/b] kernel?

If yes, then ignore everything else in this posting. If no, then please

[b]yum available kernel\*[/b]

examine the list and, if appropriate to your requirements,

[b]yum install kernel-2.6.18-53.1.21.el5[/b]

When going from the [i]xen[/i] to the [i]standard[/i] kernel, you should also edit your [b]/etc/sysconfig/kernel[/b] file such that it reads

[code]# UPDATEDEFAULT specifies if new-kernel-pkg should make
# new kernels the default
UPDATEDEFAULT=yes

# DEFAULTKERNEL specifies the default kernel package type
DEFAULTKERNEL=kernel[/code]

Confused · Post by **Confused** » 2008/06/19 17:09:23

[quote]
AlanJBartlett wrote:
Just one question for you, [b]Confused[/b]. Do you [i]need[/i] to run the [b]xen[/b] kernel?
[/quote]

I used to, was experimenting with running a copy of Windows on CentOS because of some legacy application that only works on Windows and with ActiveX. But the installation of Windows stopped working after updates so I'm giving up on that route and planning to go with VMWare next.

[quote]
If yes, then ignore everything else in this posting. If no, then please

[b]yum available kernel\*[/b]

examine the list and, if appropriate to your requirements,

[b]yum install kernel-2.6.18-53.1.21.el5[/b]

When going from the [i]xen[/i] to the [i]standard[/i] kernel, you should also edit your [b]/etc/sysconfig/kernel[/b] file such that it reads

[code]# UPDATEDEFAULT specifies if new-kernel-pkg should make
# new kernels the default
UPDATEDEFAULT=yes

# DEFAULTKERNEL specifies the default kernel package type
DEFAULTKERNEL=kernel[/code][/quote]

If I'm not wrong, this would install a non-Xen kernel, would it break other stuff that had been compiled locally or yum installed? I hate to have another OMG moment when the server starts telling me indirectly I've done something absolutely noobish that everybody else here would had assumed was to be done before changing kernel.

AlanBartlett · Post by **AlanBartlett** » 2008/06/19 17:35:59

[quote]If I'm not wrong, this would install a non-Xen kernel, would it break other stuff that had been compiled locally or yum installed?[/quote]
You're not wrong. It will [b]install[/b] the [i]standard[/i] kernel. However it will not [b]remove[/b] or [b]overwrite[/b] the existing [i]xen[/i] kernel.

At boot time, if you have a hidden menu, just press any key (the space bar will do) - then you will see the list of kernels available to boot. Select the [i]standard[/i] kernel and try it. If you have no problems with locally compiled software - and only you will know about that :-) - you can edit your [b]/etc/sysconfig/kernel[/b] file as I showed previously. Doing so will ensure that all further kernel updates will be of the [i]standard[/i] (i.e. [i]base[/i]) version.

Quite honestly, if you do not need to use a xen kernel, don't.

If you have any further queries - please post here again.

Post by **toracat** » 2008/06/19 18:29:59

[quote]
AlanJBartlett wrote:

Quite honestly, if you do not need to use a xen kernel, don't.
[/quote]
In fact, if you are planning to use vmware (as you hinted), you need to switch to the standard kernel.

dsysko · Post by **dsysko** » 2008/09/24 00:34:36

Hi,
I had this problem with all of my CentOS 5 boxen, but not my CentOS 4 systems. It plagued and vexed me until...

I found this in a ubuntu forum, and it totally solved my problem.
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/86437/comments/3

basically:

edit the config files in /etc/sysconfig/network-scripts/ifcfg-eth* thus:

iface eth0 inet static
address 192.168.3.13
netmask 255.255.255.0
gateway 192.168.3.103
pre-up modprobe -r e1000
pre-up modprobe tg3
pre-up modprobe e1000
pre-up sleep 5

Thanks Rajid!

Dustin

asgard45 · Post by **asgard45** » 2008/10/04 18:55:09

As far as I know you need an e1000e driver and not an e1000 driver for an 82566DC NIC.

CentOS