Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 19 Jun 1999 00:54:05 -0400 (EDT)
From:      "Crist J. Clark" <cjc@cc942873-a.ewndsr1.nj.home.com>
To:        dwhite@resnet.uoregon.edu (Doug White)
Cc:        holtor@yahoo.com, freebsd-questions@FreeBSD.ORG
Subject:   Re: Remote Format?
Message-ID:  <199906190454.AAA17230@cc942873-a.ewndsr1.nj.home.com>
In-Reply-To: <Pine.BSF.4.10.9906181744590.34137-100000@resnet.uoregon.edu> from Doug White at "Jun 18, 99 05:45:34 pm"

next in thread | previous in thread | raw e-mail | index | archive | help
Doug White wrote,
> On Wed, 16 Jun 1999, Holtor wrote:
> 
> > I've got several machines up at a colocation.
> > On one machine, I have 2 hd's. What I want to do.
> > Is backup my second HD on the first one, format
> > it, and slice it up into a 500 meg /tmp and the
> > rest back to the second HD, then restore..
> > Since my machine is a colo, I've got no console
> > access to it..would it be a wise idea to do this
> > remotly? Please advise..i'd really like to split up
> > this drive if possible into a /tmp.
> 
> Probably not since it involves a reboot,

Huh? It does? Not the way I read it.

> and you could flub the partition
> table setup and lock yourself out with 'No Operating System' sitting on
> the console. :-)

If he was mucking with the system partition or boot labels, I would
agree. However, I don't think he is. And in that case, I don't see how
local or remote would make much difference.

Here're the steps I see (assuming the second disk is slice '1' and
partition 'g' mounted at /usr/home),

1) Backup the second drive. I /assume/ you have enough space for the
   whole second drive in _one_ partition of the first? I'll assume
   that you are playing this risky and are tight for space,

   # dump -a0 -f - /usr/home | gzip > /usr/tmp/dump_da1.gz
   # umount /usr/home

2) OK, now, my preference is to edit a disklabel file, edit it, then
   read the edited version back in. (You might even make the new label
   as the first step in the whole process.) 

   # disklabel -r da1 > da1.dsklbl
   # your-favorite-editor da1.dsklbl

   In the editor, make the new halved partitions. Let's call the new
   one 'h'. Double-check the new label is correct. Triple-check. Now
   write the label,

   # disklabel -R -r da1 da1.dsklbl

3) OK, kewl, you have a new label... BUT this next step is the
   destructive one from which there is no turning back!

   # newfs /dev/da1s1g
   # newfs /dev/da1s1h

4) Whoo! Shiny new partitions on drive 1! Now, let's restore the one
   that was there before, 

   # mount /dev/da1s1g /usr/home
   # cd /usr/home
   # gunzip -c /usr/tmp/usr.home_da1.dmp | restore -rf -

5) OK, now the matter of your new /tmp. The current /tmp might have
   important files, and some disruption might be unavoidable if we
   move them. But a reboot is a big disruption too.

   # mv /tmp /oldtmp
   # mkdir tmp
   # mount /dev/da1s1h /tmp
   # cd /tmp
   # tar cf - -C /oldtmp . | tar xf -
   <check the move was OK, then...>
   # rm -rf /oldtmp

6) OK, almost done! Just add a line to the /etc/fstab to mount your
   new /tmp at startup. Oh, and we have that whole big backup of the
   old drive, 'rm /usr/tmp/usr.home_da1.dmp'.

OK... Now I did not see a need to reboot anywhere in there. But as I
mentioned, moving /tmp can break things that have open files there, a
reboot might not be too bad of an idea depending on just what this box
does and is running.

Did I miss anything? And was that what you were asking? :)
-- 
Crist J. Clark                           cjclark@home.com


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




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199906190454.AAA17230>