Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 18 Mar 1998 09:16:18 -0600
From:      Richard Wackerbarth <rkw@dataplex.net>
To:        Mikael Karpberg <karpen@ocean.campus.luth.se>
Cc:        stable@FreeBSD.ORG
Subject:   Re: ATTENTION: Call for opinion re: root device naming change
Message-ID:  <l03130300b13591f2e951@[208.2.87.4]>
In-Reply-To: <199803181349.OAA04416@ocean.campus.luth.se>
References:  <350FC868.52809837@kew.com> from Drew Derbyshire - UUPC/extended software support at "Mar 18, 98 08:13:12 am"

next in thread | previous in thread | raw e-mail | index | archive | help
At 7:49 AM -0600 3/18/98, Mikael Karpberg wrote:
>According to Drew Derbyshire - UUPC/extended software support:
>> Richard Wackerbarth wrote:
>> > However, the sample is EXTREMELY BIASED toward individuals who closely
>> > track things. If this change gets into the distribution, I expect there
>> > to be a significant number of users who get "bitten" by the same
>>situation.
>>
>> I agree with this.  It was assumed by at least one person that most 2.2.6
>> RELEASE installs would be fresh installs -- it is not reasonable to assume
>> any such thing about about the _stable_ release.  You don't subscribe to the
>> Walnut Creek distribution just to smoke your hard drive every three
>>months.
>
>I would say it's very logical to assume anyone that is running -STABLE is
>reading the -stable mailing list. They should be more then aware about
>the priblem by now.
>Anyone that does a fresh install will not be bitten.
>Anyone that upgrades an existing system with the new release should not be
>bitten either, if sysinstall simply fixes yout fstab, mount, kernel and
>kernel source.
>Who's left?
>
>> Again, back it out.
>
>And get the same mess again? :-P
>Make sysinstall handle the problem instead.

Again, and someone please correct me if I am wrong, there are SOME HD
configurations
which work just fine with a Feb8 kernel and WILL NOT BOOT to single user mode
on the newest kernel. As far as I can tell, 1) the newest version of
"mount" is present,
2) all the possible entries are in /dev (sd0, sd0a, sd0s1a, sd0s2a, ... ,
sd0s5a)

It is my position that, unless there is a workable configuration for ALL
machines
presently running 2.2, these changes SHOULD NOT be a part of any 2.2 release.

I believe that systems such as Rodney's which are running the latest kernel
would
also continue to run on a newer kernel that has the slice changes backed out.
If this is correct, all we lose by backing the change out is the
"compatability"
with 3.0 that this introduces. "Compatability" with 3.0 is certainly desirable.
However, it is not required. IMHO, it is much more important to maintain a
working
system for ALL present users of the branch.

If the only required change were to "correct" the fstab entries, then that
could
(at least in theory) be done by a pre-installation script. At the very
least, we better
have a warning issued that there may be a problem. Otherwise, the customer
support
burden will be overpowering and the reputation of FreeBSD will suffer.

Richard Wackerbarth



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?l03130300b13591f2e951>