Is anything being done to un-break partition names?

[Pawel Jakub Dawidek]

On Thu, Jun 04, 2009 at 04:18:09PM -0700, Juli Mallett wrote:
> Hey folks,
> 
> If I install 7.2 (or old 8-CURRENT) and partition a drive "dangerously
> dedicated" and answer No when asked if I want to create a true
> partition entry, and then install as normal, my system is set up with
> partitions named like da0s1a.  Newer 8-CURRENT instead names the
> devices da0a, which means root mount fails, etc., until one updates
> /etc/fstab.  This also seems to confuse sysinstall, which appears to
> expect labeling da0s1 to work even if you're in dangerously-dedicated
> mode — though I might be misunderstanding the interactions there;
> randi@ suggests it's just a problem with sanitizing disk names in
> libdisk, although when I built sysinstall with a patched libdisk and
> tried to use it when booting from an 8-CURRENT (snapshot as of a few
> weeks back) livefs disk, it seemed to have other problems with the
> device names.
> 
> This seems like a huge POLA violation and has eaten several hours of
> my life in terms of fixing servers that were tracking 8-CURRENT and
> failed to boot up because of the need to change /etc/fstab that wasn't
> documented in UPDATING.

I was bitten by the exactly the same thing. Unfortunately in my case I
was upgrading from 7.0 or something and kernel.old didn't work for me
with new userland. So I had to compile GEOM_PART_MBR out and compiled
GEOM_MBR in, everything on another machine and then transfer new kernel
using nc(1) and tar(1), because sshd(8) didn't work properly. And I was
in a hury. All in all, a huge disappointment.

BTW. I wasn't able to boot my system using ufs:/dev/ad0a on mountroot
prompt.

> Is anything being done to add compatibility slice names, or to teach
> mergemaster about the change?  I don't know enough about what all is
> going on on disk to know whether this is something that just affects
> dangerously-dedicated disks, but it seems to be consistently biting
> me, and I can only imagine how much trouble it's going to cause
> others.  Was this change even intentional?

I don't think it was. For me it's a bug in GEOM_PART_MBR, which has
problems detecting MBRs properly.

Shame on you, Marcel!:)

-- 
Pawel Jakub Dawidek                       http://www.wheel.pl
pjd at FreeBSD.org                           http://www.FreeBSD.org
FreeBSD committer                         Am I Evil? Yes, I Am!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-geom/attachments/20090605/d5ccdbc8/attachment.pgp