[ILUG] harddisk swap

[Brian Foster]

  | From: Timothy Murphy <tim at birdsnest.maths.tcd.ie>
  | Date: Thu, 29 Apr 2004 22:53:39 +0100
  | 
  | On Thursday 29 April 2004 19:54, Enda wrote:
  | > Where the drive class is the same, ie, SCSI / IDE etc, the device labels
  | > are the same, and with partition schemes been suggested identially out of
  | > the box, the probability of swapping MBR's and them working is very very
  | > high.
  | 
  | Have you ever done this?

 YES (years ago, when IDE was the “new‟ thing).
 multiple times, in fact.  (pre-IDE, my memory
 is this could be very tricky due(?) to the BIOS
 driver having to know about issues like landing
 zones and timing; details which the IDE hides.)

 in addition, with some restrictions, it is
 also possible to use the same MBR on drives
 with different geometries.

 offhand, I cannot see why SCSI/ATA/whatever
 would matter here, but have never tried it
 in that situation (i.e., between different
 data-bus technologies).  IDE, incidently,
 is not a data-bus technology per se, but the
 name for smarts on the drive --- Integrated
 Drive Electronics (should be Integrated Drive
 Firmware (IMHO))), albeit a driver for non-IDE
 drives using the same data-bus technology,
 AFAICR, won't work with an IDE drive (hence
 the confusion:  it seems similar to an ATA
 driver not working for SCSI or visa-versa).

 the main (only?) issue back then was the
 C/H/S data (Cylinder/Head/Sector; i.e.,
 the drive “geometry‟), both within the BIOS
 driver, and within the MBRs partitioning table.
 the MBRs boot program used standardized BIOS
 calls, and so was typically the same for most
 (all?) drives.  then, and I think now, you
 could often set the BIOS C/H/S data by hand,
 so _if_ you knew what you were doing, it was
 possible to use the “wrong‟ MBR, by lying to
 the BIOS.  typically, you had to doublely-lie,
 because often what was settable was the drive's
 “type‟ (nothing(?) more than pre-canned C/H/S),
 rather than raw C/H/S values --- so the trick
 was to find a sufficiently similar C/H/S “type‟.

 nowadays, since there are several different
 sets of BIOS calls (due to poor design; i.e.,
 silly limitations in older sets), I suppose
 an MBR program for one drive/BIOS might not
 work with a different drive and/or BIOS?
 but that is mere *speculation* on my part!

  | > Even in the case where you are the admin, and you manage to screw the
  | > "standards" up, as long as you do it consistently, you'll have a ball.
  | 
  | I don't know what this means.

 another term for what Enda is saying
 is “systematic error‟ --- does it make
 sense now?

cheers!
	-blf-

p.s.  it's been years (> 10?) since I was writing
     hard disc device drivers and boot-blocks, so
     I presume I have forgotten or misremembered
     some details.  apologies for any mistakes!

--
«How many surrealists does it take to    |  Brian Foster      Montpellier,
 change a lightbulb?  Three.  One calms  |  blf at utvinternet.ie      France
 the warthog, and two fill the bathtub   |    Stop E$$o (ExxonMobile)!
 with brightly-colored machine tools.»   |        http://www.stopesso.com