Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 1 Mar 2005 18:02:08 -0500 (EST)
From:      Jerry McAllister <jerrymc@clunix.cl.msu.edu>
To:        zettel@acm.org (Leonard Zettel)
Cc:        questions@freebsd.org
Subject:   Re: restore
Message-ID:  <200503012302.j21N2O708047@clunix.cl.msu.edu>
In-Reply-To: <200503011705.48151.zettel@acm.org> from "Leonard Zettel" at Mar 01, 2005 05:05:47 PM

next in thread | previous in thread | raw e-mail | index | archive | help
> 
> On Tuesday 01 March 2005 04:36 pm, Jerry McAllister wrote:
> > >
> > > The production machine is working just fine.
> > > All the FreeBSD stuff is on ad0s1, a 40 GB hard drive.
> > > I have been using a 120 GB drive on ad0s1 for
> > > backup, and have (apparently successfully) done
> > > a dump of production / and /usr.
> > >
> > > So, I fire up the development machine with
> > > the 120 GB drive as the slave of controller 1,
> > > what I would like to be the main drive of the
> > > development system (a 40GB hard drive) as
> > > the master of controller 0, and disk 2 of the
> > > Free BSD CD-ROMs in the CD-ROM drive.
> > >
> > > Up comes beastie and I boot.
> > > I select "fixit" from the menu, followed by alt-F4.
> > >
> > > Then:
> > > mount /dev/ad0s1a /mnt (to make the new root accessible to
> > >                                          the system).
> > > mkdir backup                   (make a mount point for the 120GB drive)
> > > mount /dev/ad3s1a /backup (mount the 120 GB drive)
> > > newfs /dev/ad0s1a           (start the new root with a clean sheet
> > >                                           prior to doing a retore)
> > >
> > > BUT instead I get a diagnostic as follows:
> > > fstab: /etc/fstab:0: No such file or directory
> > > newfs: /dev/ad0s1a: failed to open disk for writing
> > >
> > > Could anybody tell me what I *should* be doing (bonus
> > > extra points for expalining why :-) )?"
> >
> > Well, I got a little confused as to which machine has which drive[s]
> > Especially when you say early on that everything on the production
> > machine is on ad0s1 - a 40 GB drive and then say you are doing
> > backups to ad0s1 - a 120 GB drive.   
> Typo -sorry about that; the 120G is ad1s1a when it is on
> the production machine.
> 

Figured that was probable.

> > That sounds like you have 
> > two ad0 drives on the machine at the same time.   I suspect something
> > is missing of twisted in the description.
> >
> > But, farther down seems to be your real problem.
> > The first question is did you look at the boot messages when you
> > came up in fixit and are sure that ad0 and ad3 are the devices you
> > need to be dealing with?  
> df verifies that ad0 is the 40G and ad3 is the 120G on the
> development machine. 
> 
> As I understand it, ad3 is the slave of the second IDE controller (1).  It 
> gets to be slave because the drive jumpers are set
> that way, so the 120 can be slave on the first controller on the
> production machine (ad1).  That way I can swap without redoing 
> the jumper.

Could be.  My only IDE machine I have never had more than one disk on.
SCSI is more orderly.   I would just check the messages as it boots or
look at dmesg(8) and verify the device names.
 
> > I kind of would have expected ad0 and ad1 
> > or maybe ad0 and ad2, but I am not used to mucking with IDE controllers.
> >
> > Next, why did you try and mount /dev/ad0s1a and then newfs /dev/ad0s1a
> > That should not work at all.   You don't newfs a mounted partition.
> 
> Gee, I didn't know that.  It wasn't clear when I read the handbook
> or man pages.  May give me something new to try--

That is probably your main problem.

> > Second, that would wipe what is on there - maybe you want that.
> 
> I do indeed want that.  
> 
> > I don't know why it complains about fstab at that point.  The fixit
> > does not create one, but I don't see where it is needed for what you
> > are trying.   
> Makes two of us.

I've been fooled before.   Most days.

////jerry

> 
> > Maybe, just doing the wrong thing with newfs got it 
> > to trying to check stuff.
> >
> > Anyway, There are some thoughts of things to work out.  Maybe
> > they will give you a clue of what to try next.
> >
> > ////jerry
> >
> Thanks.
> > >   -LenZ-
> 



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200503012302.j21N2O708047>