Re: [gentoo-user] strange error during boot

[Joost Roeleveld]

On Thu, February 12, 2009 2:26 pm, John covici wrote:
> on Thursday 02/12/2009 Joost Roeleveld(joost@××××××××.org) wrote
>  > On Thu, February 12, 2009 2:05 pm, John covici wrote:
>  > > on Thursday 02/12/2009 Joost Roeleveld(joost@××××××××.org) wrote
>  > >  > On Thu, February 12, 2009 10:52 am, John covici wrote:
>  > >  > > Hi.  I just upgraded a gentoo system from about August 2008 to
>  > > current
>  > >  > > -- including updating baselayout and openrt and now  when I boot
> I
>  > > get
>  > >  > > a series of messages quite early in the boot modprobe: fatal
> /sys is
>  > >  > > not mounted.  Eventually it does boot and all seems to work with
> the
>  > >  > > exception of the script for my hsfmodem, but I am curious as to
> what
>  > >  > > those message mean and if there is a way to fix them.
>  > >  > >
>  > >  > > Any assistance would be appreciated.
>  > >  >
>  > >  > Did you include sysfs support to your kernel and do you have a
>  > > directory
>  > >  > '/sys'? (SYSFS)
>  > >  > This can be found in: File systems / Pseudo filesystems in the
> kernel
>  > >  > configuration.
>  > >  >
>  > >  > The '/sys' filesystem is as important as '/proc' these days.
>  > >
>  > > The plot thickens -- by the time I log in after booting, /sys is
>  > > mounted with the correct file system.  Still very strange.
>  >
>  > Hmm... so, something does solve the problem you are seeing at the
>  > beginning later on.
>  > Did you update all the configuration files (including the ones in
>  > /etc/init.d/.. )?
>  > It could be that something there is not set correctly.
>  >
>  > For now, I am assuming the issue is in the boot-sequence/runlevel.
>  >
>  > Can you check which services are in your boot-runlevel?
>  > I have:
>  > bootmisc, checkfs, checkroot, clock, consolefone, hostname, keymaps,
>  > localmount, modules, net.lo rmnologin and urandom.
>  > Think these are the default ones.
>  >
>  > Do you use an initrd? If yes, did you update this as well?
>
> I regenerated the initrd, but I am still using 2.6.20 kernel which I
> will update soon, but I wonder if this is the problem -- something
> wrong with the initrd, but regenerating did not fix it.  In my boot
> level I have
> bootmisc@
> consolefont@
> device-mapper@
> fsck@
> hibernate-cleanup@
> hostname@
> hwclock@
> keymaps@
> localmount@
> modules@
> mtab@
> net.lo@
> procfs@
> root@
> swap@
> sysctl@
> termencoding@
> urandom@
> in my sysinit I have
> devfs@
> dmesg@
> udev@

Do you have "device-mapper" in your boot-level?
In that case, you might want to check which init-script mounts the '/sys'
filesystem as this script requires the /sys filesystem to be mounted.

May I ask why you have this added as I don't use it with my LVM drives.

--
Joost