[Ovirt-devel] node boot error

[Alan Pevec]

On Wed, Jun 24, 2009 at 2:07 AM, Ian Main <imain at redhat.com> wrote:

> On Tue, 23 Jun 2009 17:15:33 +0400
> dima vasiletc <pronix.service at gmail.com> wrote:
>
> > Hello
> > i have fedora 11 and build from next branch ovirt
> > during boot by network (pxe) second node appear error.
> > host stay anavailable and not sent to manage node identify data.
> > ping return "unreachable host"
> >
> > Where may be error ?
>
> To see the real error you have to look at the console on the booting
> node.  We're having issues with fedora 11 nodes, Alan and I are working
> on resolving it right now.  Hopefully it won't be long!  I'm getting
> frustrated with this one myself :).


Two issue were identified with livecd initrd in F11:
- PXE boot failing with "Could not find any loop device." - loop is built-in
since F10 so actually /dev/loop* was missing b/c udev didn't create it in
time.
  Fixed by adding back udevadm settle (removed for F11 in
http://git.fedorahosted.org/git/?p=mkinitrd;a=commitdiff;h=0016a5fb236839db31344fef1c87ad68680e301b)

- udev sometimes hangs on startup, root cause unclear:
Running plymouthd
udev: starting version 141

Note that I didn't get any injected echo output after nash daemonize call in
/init !?
Workaround for this hang is to remove plymouth completely from initrd
scripts, but that's clearly suboptimal.

Attached is the patch for livecd-iso-to-pxeboot.sh which adds above
workarounds into initrd, proper fixes need to be discussed with mkinitrd
guys,
but with this we can at least get reliable PXE boot.

Alan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/ovirt-devel/attachments/20090624/a4c6195f/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: livecd-iso-to-pxeboot.diff
Type: application/octet-stream
Size: 1128 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/ovirt-devel/attachments/20090624/a4c6195f/attachment.obj>