*BSD News Article 22129


Return to BSD News archive

Newsgroups: comp.os.386bsd.bugs
Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!news.larc.nasa.gov!darwin.sura.net!europa.eng.gtefsd.com!uunet!ihz.compuserve.com!mfoley
From: mfoley@csi.compuserve.com (M Foley)
Subject: Re: YET ANOTHER disklabel AND IDE 2nd DRIVE PROBLEM - HELP!!
Message-ID: <CELsHv.3rH@csi.compuserve.com>
Organization: CompuServe Incorporated
Date: Sat, 9 Oct 1993 00:10:42 GMT
Lines: 32

I have received a significant number of messages from people responding
to my original trouble of not being able to create a disklabel on a
2nd IDE driver which was move straight from DOS (no pre-existing disklabel).
Let me respond to the three type of suggestions I received:

1.  Yes - I did try creating partitions, c and d, both separately and together.
    This did not help. Have any of you actually got the drive to disklabel
    simply by specifying these partitions? Or were these conceptual 
    suggestions?

2.  Yes - I do have a raw device properly configured for /dev/rwd1a. I just
    didn't list the raw partitions in my posting.

3.  For those who suggested writing trash or zeros to the raw device in
    order to kill the DOS partition, I tried that. I could not write to
    the raw device.

The only solution I found was to mount the drive as my first drive, boot
from Tiny BSD and run install. This creates a label (how?) which I can
then overwrite when I put things back. Two things:

1.  I'm really curious if those of you who made the above suggestions 
    actually have successfully disklabled, or if these were just ideas. If
    you were able to do this (on a non-pre-disklabeled drive, remember), and
    I could not, I would be willing to assume I did something wrong and
    resume experimentation.

2.  Since I did get a surprising number of **different** suggestions related 
    to this, I would like to suggest that offical proceedures for labeling
    virgin drives be posted, perhaps in the FAQ, even if the procedure is
    the work-around I found. If the only procedure **is** this workaround,
    perhaps this could be fixed in some future version of *bsd.