This is in the "Is this all there is/is this OK?" category.

I did (as root):

1) dd if=/dev/hda of=/some/file/not/onhda1/mbr.bin ibs=512 count=1
2) dd if=/dev/zero of=/dev/hda count=32767
3) dd if=/some/file/not/onhda1/mbr.bin of=/dev/hda
4) (ntfsclone command to restore filesystem to /dev/hda1)

And, as far as I can tell, everything is fine.  Machines boots into
Windows as before.  Is this all OK?  And, in particular, is there any
more "built-in" way to do it?  Perhaps a way to do a DOS "fdisk /mbr"
from Linux?

Note that this was all just testing; no actual data was harmed in the
performance of this test.

On 2007-04-05, Kenny McCormack <gazelle@xmission.xmission.com> wrote:

> This is in the "Is this all there is/is this OK?" category.
>
> I did (as root):
>
> 1) dd if=/dev/hda of=/some/file/not/onhda1/mbr.bin ibs=512 count=1
> 2) dd if=/dev/zero of=/dev/hda count=32767
> 3) dd if=/some/file/not/onhda1/mbr.bin of=/dev/hda

I just use "lilo -s [file]" to save a copy of the old boot record and 
"lilo -u [file]" to restore it.

-- 

John (john@os2.dhs.org)

In article <slrnf1b2r0.gi2.john@vector.os2.dhs.org>,
John Thompson  <JohnThompson@new.rr.com> wrote:
>On 2007-04-05, Kenny McCormack <gazelle@xmission.xmission.com> wrote:
>
>> This is in the "Is this all there is/is this OK?" category.
>>
>> I did (as root):
>>
>> 1) dd if=/dev/hda of=/some/file/not/onhda1/mbr.bin ibs=512 count=1
>> 2) dd if=/dev/zero of=/dev/hda count=32767
>> 3) dd if=/some/file/not/onhda1/mbr.bin of=/dev/hda
>
>I just use "lilo -s [file]" to save a copy of the old boot record and 
>"lilo -u [file]" to restore it.

That might work, too.  But is there anything wrong with my method?

In article <slrnf1b2r0.gi2.j[58]...@vector.os2.dhs.org>,
John Thompson  <JohnThomp[59]...@new.rr.com> wrote:

>On  2007-04-05,  Kenny  McCormack <gaze[60]...@xmission.xmission.com>
wrote:

> >> This is in the "Is this all there is/is this OK?" category.
> 
> >> I did (as root):
> 
> >> 1) dd if=/dev/hda of=/some/file/not/onhda1/mbr.bin ibs=512 count=1
> >> 2) dd if=/dev/zero of=/dev/hda count=32767
> >> 3) dd if=/some/file/not/onhda1/mbr.bin of=/dev/hda
> 
> >I just use "lilo -s [file]" to save a copy of the old boot record and
> >"lilo -u [file]" to restore it.
> That might work, too.  But is there anything wrong with my method?

I too prefer using dd, because it is more general.
If you also use mc, which does/shows 'everything', you can see the file
being written, its name, size, date ..permission; and even the contents
also in hex.

Save mbr ==>
dd if=/dev/hda1  of=./<fileIDnot used on current dir>  ibs=512 count=1
   will save 512 * 1 = the first 512 bytes of /dev/hda1

I'm not sure if /dev/hda  is good enough.
That's why it's best to have eg. mc to SEE what's happening.
The last 2 bytes of the 512 block must be hex 55 AA.
This is part of the check.
Here's a typical ending - the last 16 bytes.
0001f0: 79 20 74 6f 20 72 65 74 72 79 0d 0a 00 00 55 aa

Restore mbr ==>
dd if=./<the saved file> of=/dev/hda1

??! what's:
dd if=/dev/zero of=/dev/hda count=32767
all about ?
This would wipe the first 32767bytes, before restoring the mbr.

If I was at a linux box now I'd try the lilo way, and hope to
find where lilo had saved it; then use dd and compare the 2 files;
of course using mc.
NB. see about hda vs. hda1

== Chris Glur.