Cannot mount an older disk

[Konstantin Belousov]

On Wed, Oct 09, 2019 at 08:32:16AM +0100, Bob Eager wrote:
> On Tue, 8 Oct 2019 21:35:07 -0400
> "Mikhail T." <mi+t at aldan.algebra.com> wrote:
> 
> > Hello!
> > 
> > Going through older hard drives, I found one that still seems to work 
> > and was curious, what's on it. The OS -- 12.1-STABLE -- sees it find. 
> > The disklabel seems sane (except for the number of partitions):
> > 
> >     # /dev/ada1:
> >     8 partitions:
> >     #          size     offset    fstype   [fsize bsize bps/cpg]
> >        b:   12582912          0      swap
> >        c: 1465149168          0    unused        0     0 # "raw" part,
> >     don't edit
> >        d: 1452566256   12582912    4.2BSD     8192 65536 52352
> > 
> > and there are ada1, ada1b, and ada1d entries under /dev. So far so
> > good. Unfortunately, both mount and fsck tell me the same blatant
> > lie, that the device does not exist:
> > 
> >     # fsck -y /dev/ada1d
> >     Can't open /dev/ada1d: No such file or directory
> > 
> >     # mount /dev/ada1d /mnt
> >     mount: /dev/ada1d: No such file or directory
> > 
> > Any suggestions? Thank you! Yours,
> 
> Custom kernel? If so, try booting GENERIC. Might be that a support fs
> option is missing.

This is most likely a stray bsd label which appeared in the first (second ?)
block of the disk due to some pecularities of old partitioning tools.
Note the absence of the 'sN', i.e. MBR partition, between raw disk name
and bsd slice.  Some time at 9 or 10 lifetime priorities changed due to
switch to gpart.

I do not remember how this was worked around, most likely by zeroing second
block of the disk.  Of course, it is better to do the experiment on a copy
if the original is suspected to contain a useful information.