[gentoo-user] Re: WAS [.. min install isos] gentoo vm guest not booting

[Harry Putnam]

Mark Knecht <markknecht@×××××.com> writes:

> On Sun, Jun 26, 2011 at 7:19 AM, Harry Putnam <reader@×××××××.com> wrote:
> <SNIP>
>>
>> The boot process does get a little farther but comes to a halt at
>>
>>  Switching to clock source tsc
>>
>> And there it has sat for some time now.  Apparently is not going to continue.
>
> That EXACT symptom was EXACTLY what we saw on new hardware when we
> didn't have the /dev/null, /dev/console issue because they were
> missing from the recent tarballs. I actually had some photos that
> showed it but cannot find them. Probably deleted at this point. Here's
> a;link to the LKML thread and eventually Paul Hartman's pointer to
> what fixed it:
>
> https://lkml.org/lkml/2011/6/16/918
>
> Please consider reviewing the links I sent earlier about the 10 or so
> steps required to get them into the VM correctly.
>
> To verify, boot the VM with the install CD, mount the partitions and
> just

look at /dev. Is everything REALLY there?
>
> If /dev/null and /dev/console are there, and are special dev files and
> not just regular files, then make sure you are starting udev
> explicitly in rc-update.

[...]

First, thanks for going well beyond the call of duty and showing the
patience of Job.

Those files are there already and have been all along.

But anyway I rm'ed them and followed the steps resulting in:

(From chrooted shell)

root@sysresccd /dev % ls -l console null
crw-rw---- 1 root root 5, 1 Jun 26 17:33 console
crw-rw---- 1 root root 1, 3 Jun 26 17:33 null

rc-update |grep udev
     udev |                                        sysinit
udev-postmount | default

-------        ---------       ---=---       ---------      -------- 

You may not have noticed that 2.6.38-r3 where your config came from is
no longer available in portage... I used -r4 first time but now
redoing it all I'm using your config but on 2.6.39-r1.  Since I can't
get the exact one you used I may as well get the latest.

So to summarize... I've used a version I bastardized, of your config, on
a different kernel version... maybe not the smartest way to go...

I've recreated /dev/console and /dev/null using the steps you quoted.

I notice on a reboot of the install media and chrooting into gentoo,
that /dev/console now has 600 permissions even though I gave it 660.
And /dev/null now has 666 although I gave it 660.

  cd /dev

  ls -l console null
  crw------- 1 root root 5, 1 Jun 26 13:02 console
  crw-rw-rw- 1 root root 1, 3 Jun 26 13:02 null

Not sure if that is normal...but I think it is, anyway  building the
kernel now and we'll soon see how it works.

........

OK, well nothing better happens on this go around.

Boot still hangs up at `Switching to clocklsource tsc

What is so aggravating is that building a guest from debian works
first try.. no problems.  I guess I'll have to relearn apt-get and
other debian tricks if I want to get going on the reason I'm building
a linux guest.  I run gentoo on my main desktop but there, installs are
usually trouble free.

Seems like it should be just as easy with a vbox guest too since the
hardware is kept pretty simple,  but it never has been any
time I've tried it, which has been quite a few over the years.

I keep hoping it will work and try again some mnths later ... but it
never has yet.

If I want to get a tarsnap server in place (The reason I'm building a
guest) anytime soon I will have to drop gentoo as guest I'm afraid.

I may fool with it a bit more... but it's pretty discouraging.