Re: Can Grub start Windows XP from "other" partition

[Michael D. Setzer II]

On 22 Nov 2012 at 0:11, Felix Miata wrote:

Date sent:              Thu, 22 Nov 2012 00:11:57 -0500
From:                   Felix Miata <address@hidden>
Organization:           less than infinite
To:                     address@hidden
Subject:                Re: Can Grub start Windows XP from "other" 
partition

> On 2012-11-22 03:36 (GMT+0100) Ulf Zibis composed:
> 
> >> Grub Legacy can do the very same partition type modifications at
> >> boot time. I doubt this capability has been dropped from Grub2. I
> >> don't use Grub2, and install no bootloader during installation of
> >> non-*buntu distros that do not offer to install Grub Legacy.
> 
> > Can you point me to the paragraph in Grub Legacy docs, where this is
> > described? This would ease me to possibly find that feature in Grub
> > 2.
> 
> http://orgs.man.ac.uk/documentation/grub/grub_4.html#SEC22
> 
> >>> I remember that I had read a tutorial about "duplicating Windows"
> >>> to 2 partitions for the purpose to have a working installation and
> >>> a test installation to try out dangerous things without corrupting
> >>> the first. I'm pretty sure, that the 2nd one was created by just
> >>> copying the 1st, and if booted into the 2nd, partition 1 appeared
> >>> as D:. Unfortunately I do not find this tutorial again by Google.
> 
> > Do you have an idea, how this could have been possible, or do you
> > know about a doc which may help me?
> 
> Maybe it could be done with XXCOPY while booted to a maintenance
> system, but not by a partition clone. Registry modifications would be
> required as well as adjusting boot.ini.
> 
> One reason cloning cannot be expected to work can be seen by a study
> of http://fm.no-ip.com/Tmp/Dfsee/dfsw5000.txt
> 
> Notice that the three primary partitions seem to be equal in size,
> except that they are not. The first is smaller by the number of
> sectors in a track minus one. To try to clone between a first and an
> other would cause sector counts to be out of sync with "cylinder"
> alignment, something Linux can cope with, but WinXP can't.
> 
> >>>> Windows needs a primary to be C:, but it needn't be "installed"
> >>>> to C:.
> 
> >>> In other words, Windows needs the NTloader, boot.ini etc. in the
> >>> first Windows-visible-native-typed partition which is always named
> >>> as C:. If Windows itself is "installed" in any other other
> >>> partition, it would be named different, e.g. D:, E: ... right?
> 
> >> Yes.
> 
> > But how does it work, if there are 2 Windows installations with 2
> > NTloaders?
> 
> Proper configuration of menu.lst, including hiding and unhiding of
> primaries.
> 
> > If Windows can only be booted by the NTloader from the 1st
> > partition,
> 
> Where did you get this idea? WinXP can only be booted from an only
> visible primary.
> 
> > how can Grub tell it to boot the Windows installation from the 2nd
> > partition?
> 
> Windows can be booted by any primary partition that contains NTLDR, if
> it is an unhidden partition, and there no other Windows native
> unhidden primary partitions.
> 
> > But if there are 2 Windows installations on a system, both listed in
> > Grub's start menu, each has it's own NTLDR + boot.ini with only 1
> > default entry, how can Grub manage to chainload the right NTLDR, or
> > instruct the NTLDR in the 1st partition to boot the Windows
> > installation in the 2nd partition?
> 
> Can't do that. Once chainloaded, you have to use the Windows boot menu
> from a properly constructed boot.ini. This will be necessary if you
> make only one primary to be used as C:. You only put two Windows
> stanzas in Grub's menu if you have two different primaries that could
> be C:.
> 
> > If I have understood right, Grub can boot Linux from any arbitrary
> > partition, but in case of Windows it should fallback to the BIOS
> > which boots the Windows installation from the "active" partition,
> > which can only be one of the two.
> 
> You don't have it right.
> 
> > Or ist it like this? :
> > Grub can start *any* NTLDR regardless of the "activeness" of it's
> > enclosing partition.
> 
> "Active" is only relevant with regard to MBR code. It indicates the
> jumpto partition. With Grub on a primary and standard code in the MBR,
> active must be and remain the primary that Grub is on, unless you wish
> Grub to be a secondary boot loader chainloaded from elsewhere, such as
> via boot.ini on a C:.
> 
> > In case of the NTLDR of the 2nd Windows-visible-native-typed
> > partition it will boot the 2nd Windows partition, but "see" the 1st
> > Windows partition as C: and the 2nd as D:, which is the big problem
> > here if it was originally installed on C:
> 
> Windows cannot boot a system with more than one visible native type
> primary per HD. In a single HD system, D: will always be a logical,
> and C: will be the one and only visible primary. -- "The wise are
> known for their understanding, and pleasant words are persuasive."
> Proverbs 16:21 (New Living Translation)
> 

In regards to making a backup copy of the partition. I'm the 
maintainer of the opensource g4l project that can make an 
ntfsclone image of a partition that can be stored on another 
partition or disk and then restored. In my classroom, I keep an 
image of the XP on partition /dev/sda6 and have a boot option on 
the grub menu that can restore the image in about 12 minutes, so 
students can try things with the worst being the need to restore 
the image. The ntfsclone option is fast, and only backs up used 
space as contrasted to the raw image options. I would also 
recommend backing up the mbr, just in case something also 
modifies it. 

Project is on sourceforge and is free.


>   Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
> 
> Felix Miata  ***  http://fm.no-ip.com/
> 
> _______________________________________________
> Help-grub mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/help-grub


+----------------------------------------------------------+
  Michael D. Setzer II -  Computer Science Instructor      
  Guam Community College  Computer Center                  
  mailto:address@hidden                            
  mailto:address@hidden
  http://www.guam.net/home/mikes
  Guam - Where America's Day Begins                        
  G4L Disk Imaging Project maintainer 
  http://sourceforge.net/projects/g4l/
+----------------------------------------------------------+

http://setiathome.berkeley.edu (Original)
Number of Seti Units Returned:  19,471
Processing time:  32 years, 290 days, 12 hours, 58 minutes
(Total Hours: 287,489)

address@hidden CREDITS
SETI        13309738.310486   |   EINSTEIN     9114415.909852
ROSETTA      5395483.331241   |   ABC         15478466.869275