[3908] in Release_7.7_team
Re: Different Solaris update miniroot
daemon@ATHENA.MIT.EDU (miki@MIT.EDU)
Sun Jul 6 12:31:02 2003
From: miki@MIT.EDU
Message-ID: <1057509058.3f084ec2d4ff1@webmail.mit.edu>
Date: Sun, 6 Jul 2003 12:30:58 -0400
To: Greg Hudson <ghudson@mit.edu>
Cc: release-team@mit.edu
In-Reply-To: <200307040254.h642sfLp026286@error-messages.mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
I switched in the dev cell from a Solaris9 miniroot to a Solaris8 miniroot.
I re-tested (although the Solaris8 miniroot was once tested for all platforms)
the Sun-Blade and the Ultra5/10. I tested as well that going from
Athena8.3 to Athena9.2 is OK om Ultra5/10.
Yes, I think that we should have two miniroots one based on Solaris9 and one
based on Solaris8 and for the time being use the one based on Solaris8.
Miki
> Short form: the update miniroot for Solaris has changed in the dev
> cell, in order to work around the problem with hardware installed at
> 8.3. Before the public release, we need people to test dev cell
> updates from 9.1 to 9.2 on various different kinds of Sun hardware to
> make sure that the new miniroot volume works. Please help out if you
> are able to.
>
> Long form: miki thinks our update problem is not related to the dad
> 134 device (in particular, her test machine demonstrated the failure
> even though the dad device was set to 136) and that it is related to
> having devices in /etc/name_to_major which aren't appropriate for the
> hardware. Or something. At any rate, the current stopgap solution is
> to switch from a Solaris 9 update miniroot to a Solaris 8 update
> miniroot.
>
> In the long run, it actually does make sense to use an update miniroot
> matching the version you're updating from, rather than the version
> you're updating to. That way you can just copy in the
> /etc/name_to_major from the root. It does mean we need two different
> update miniroots, though (one if you're updating from 9.1 to 9.2,
> another if you're doing a 9.2 patch release--assuming we ever do a
> miniroot patch release, of course).
>