Mount root error / New device numbering?

[Fred Souza]

On Fri, May 14, 2010 at 00:25, Jeremy Chadwick <freebsd at jdc.parodius.com> wrote:
> Absolutely.  I've done it myself many times over the years, including
> remotely over serial console.  However, you said you did that then typed
> "exit" rather than "reboot", and the end result was a kernel panic.

Yes, I should have given it another clean try after changing
/etc/fstab. I guess my rustiness with the Good Stuff(tm) plus the
unexpected behavior made me panic myself.

> Honestly, I'm not surprised; the system was probably still confused
> about the root device.  I'm guessing some kernel innards (or maybe
> something picked up from boot2/loader) still referenced the "unknown
> root device" and caused the panic.

Could be. The system had just too many possible points of failure at
that point (its original kernel refused to boot for a few hours due to
the geometry mess, then all of a sudden started working normally, for
instance), so I take whatever I learned from that install as
experience. I'm glad I thought about the right thing to do, it may
have failed for a number of things in the way.

> Even on other operating systems, if I'm dropped (unintentionally or
> intentionally/by choice) into single-user mode, I reboot the system
> rather than exit out of single-user and hope that multi-user works from
> that point forward.  I've seen "exit" on Solaris fail and cause all
> sorts of mayhem (all sorts of system startup services (not rc/init!)
> failing, machine ending up in some sort of catatonic state).

Good to know, I never really paid much attention to those details (I
will from now on). Thank you a lot for the help, Jeremy. I will try
your suggestions in the morning and post back to tell what did I find
out.


Best regards,
Fred