Red Hat Bugzilla – Bug 470530
Anaconda can not select the installation NIC from boot ksdevice option
Last modified: 2009-11-18 17:11:20 EST
Description of problem:
In anaconda 22.214.171.124 from rawhide the kernel cmdline option ksdevice does not work with ethx param.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Start the installation process from a pxe source on a multi homed system. (I start my test from a secondary harddisk.)
2. My grub entry is:
Title Fedora Rawhide
kernel /boot/vmlinuz.raw ksdevice=eth0 ks=http://<server>/ks.raw
3. boot this entry select keyboard and language and see what happens.
When come to the point where the network is configured in anaconda, it drops some text on the blue background and NetworkManager tries to setup a network, fails and gives a screen where I have to specify the kickstart device again although I have supplied it on the command line. When I select the device everything else until the "reboot" goes as expected.
After booting I expect anaconda to fetch the kickstart file and install the OS without any user interaction.
Working fine for me. I have two NICs in my system, eth0 is the one with an active connection. I boot with:
And it starts up just fine.
My only guess is eth0 is not the device you think it is. Can you specify it by MAC address and see what happens? As in ksdevice=MACADDR
If the device you specified via ksdevice is unusable, loader will drop you back to the network device selection screen.
At the moment I don't have the time to perform your test suggestion. I will do that later on this day.
Your assumption that my device is not the one I think it is, seems unlikely to me. When the NetworkManager fails, it pops up the network device selection screen. It gives me the choice between eth0 and eth1. I select eth0 (the same one I specified in the cmdline). With this choice NetworkManager starts the NIC without any problem and the kickstart process continues as expected.
You have an idee how I can verify your suggestion in the output of the other tty's?
Anyhow I will test again and report my findings.
Created attachment 322961 [details]
Tty4 after error
If I start the installation with ksdevice=$MACADDRESS I see the same behaviour, NetworkManager fails.
I enclose a picture of my tty4 screen after de network device selection screen has come up after the error.
Oddly you can see that DHCP has handed out a network lease and still NetworkManager fails.
loader's job at this point is to collect your network settings and pass them to NetworkManager. It does look like we're doing that correctly, but something is happening in NetworkManager when it tries to bring up the device. The last line in the tty4 output looks a little curious to me.
I'm going to reassign to NetworkManager because I think at this point, that's where the failure is occurring.
Can I get all your ifcfg-* files from /etc/sysconfig/network-scripts post-install? Looks like something is happening with the mac address locking.
Created attachment 323815 [details]
The requested ifcfg-*
In this attachment you'll find the requested ifcfg-* files.
I noticed, that after the installation according to ifcfg-eth1 the eth1 interface should not be up, but booting into the system shows the interface eth0 and eth1 both up.
This bug appears to have been reported against 'rawhide' during the Fedora 10 development cycle.
Changing version to '10'.
More information and reason for this action is here:
Any update on this issue? It is still ongoing with Anaconda v.126.96.36.199 on i386.
Dan, has there been any progress on this issue? It is keeping us from rolling out Fedora 10 in our environment. I am happy to provide any additional details or testing needed. Thanks.
Can somebody essentially 'cat /etc/sysconfig/network-scripts/ifcfg-eth0' and 'cat /etc/sysconfig/network-scripts/ifcfg-eth1' after the kickstart has failed, but *before* you re-pick it from the loader's network screen?
The screenshot above and the ifcfg files are from after the loader has re-queried for the device, and thus wont' be useful in figuring out the bug.
In order to cat the requested files I need to have a shell, but at the point where NetworkManager fails install.img is not loaded yet and so I have no shell.
Is there another way to get the information you asked for at that particular point?
This message is a reminder that Fedora 10 is nearing its end of life.
Approximately 30 (thirty) days from now Fedora will stop maintaining
and issuing updates for Fedora 10. It is Fedora's policy to close all
bug reports from releases that are no longer maintained. At that time
this bug will be closed as WONTFIX if it remains open with a Fedora
'version' of '10'.
Package Maintainer: If you wish for this bug to remain open because you
plan to fix it in a currently maintained version, simply change the 'version'
to a later Fedora version prior to Fedora 10's end of life.
Bug Reporter: Thank you for reporting this issue and we are sorry that
we may not be able to fix it before Fedora 10 is end of life. If you
would still like to see this bug fixed and are able to reproduce it
against a later version of Fedora please change the 'version' of this
bug to the applicable version. If you are unable to change the version,
please add a comment here and someone will do it for you.
Although we aim to fix as many bugs as possible during every release's
lifetime, sometimes those efforts are overtaken by events. Often a
more recent Fedora release includes newer upstream software that fixes
bugs or makes them obsolete.
The process we are following is described here:
I believe this bug was also present in Fedora 11 but I am not sure about 12. Would I be able to grab the requested information using a serial terminal?
You should be able to, you can hit Ctl+Z and that will drop you to a busybox shell. Then you can grab /var/log/messages or I think in the loader it's /tmp/install.log or something like that. You should also be able to get to the /etc/sysconfig/network-scripts/ifcfg-* from there.
Dan, I reran the Fedora 11 installer on a few of our motherboards and the issue does not seem to be present. Sorry for any confusion, it seems it would be safe to close this bug now.
Ok, thanks for the report.