Subject: Re: NetBSD <-> Solaris 8 disklabel compatibility?
To: der Mouse <mouse@Rodents.Montreal.QC.CA>
From: Hauke Fath <hauke@Espresso.Rhein-Neckar.DE>
List: port-sparc
Date: 05/12/2002 19:34:05
At 12:43 Uhr -0400 12.5.2002, der Mouse wrote:
>>> Is Solaris completely ignoring the NetBSD-created label, or is it
>>> just failing to see the partitions that don't match the Sun
>>> constraints?
>
>> [hauke@pizza] ~ # fdisk /dev/rdsk/c1t2d0s2
>
>Solaris/sparc uses _fdisk_??  Ugh.  Are you sure you should be using
>fdisk for SCSI disks on Solaris?  I'd expect it to be only for IDE
>drives (Solaris/sparc does run on some machines with IDE drives).

Well, the SPARC Solaris 8 'fdisk' appears to work with SCSI disks, too.
Apparently an act of linuxation...

>Does Solaris have a "format" command?  (That's what SunOS and older
>Solaris used.)

Yes.

>Does it understand that disk?

format has an 'auto configure' mode. When I select that, it comes up with
reasonable drive data, but format:verify talks about a corrupt primary
label, and format:partition:print gives me a vanilla default layout.

>> Cool!  I have read of your sunlabel tool before but hadn't thought of
>> it in this context.
>
>Well, it may not be what you need, but it also may help.

If I know that Solaris cannot make any sense of a NetBSD label, I can
simply start the new installation by setting up Solaris first. That's the
point of planning and asking questions _before_ I go trashing my main
server.  ;)

>> Can the tools also re-write an existing NetBSD disklabel so that it
>> becomes acceptable to Solaris without losing the NetBSD
>> functionality?
>
>Only in some cases, at best - we don't know what it is about the NetBSD
>label that's upsetting Solaris.  If the NetBSD label conforms entirely
>to the Sun constraints, I think the only difference will be the
>embedded BSD label in the "spare space".  If that is what's making
>Solaris unhappy, then reading the label with sunlable and writing it
>back will fix it because it will clear that spare space.  If not, it
>may not, and I'm not sure what it's going to take.

Mhm. I used sunlabel to adjust the end of the 'c' partition so that it
equals that of the 'g' partition (both ending on cylinder boundary). If
this causes a re-write of the entire label, it has made no difference to
Solaris.

Mounting a Solaris UFS partition readonly in NetBSD, btw., only got me

/mnt: bad dir ino 2 at offset 0: mangled entry

'-- ick.'

	hauke


--
/~\  The ASCII Ribbon Campaign       "They that can give up essential liberty
\ /    No HTML/RTF in email          to obtain a little temporary safety
 X     No Word docs in email         deserve neither liberty nor safety."
/ \  Respect for open standards                    -- Benjamin Franklin, 1759