Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #187237 > unrolled thread

netinst.iso on USB flash drive -- unexpected results

Back to article view | Back to linux.debian.user

Contents

  netinst.iso on USB flash drive -- unexpected results Richard Owlett <rowlett@cloud85.net> - 2017-09-26 13:40 +0200
    Re: netinst.iso on USB flash drive -- unexpected results "Thomas Schmitt" <scdbackup@gmx.net> - 2017-09-26 14:10 +0200
      Re: netinst.iso on USB flash drive -- unexpected results Richard Owlett <rowlett@cloud85.net> - 2017-09-26 14:40 +0200

#187237 — netinst.iso on USB flash drive -- unexpected results

<BACKGROUND>
Yesterday I had attempted to use netinst {on USB flash} to do an 
atypical install to another USB device. It failed due to at least an 
invalid /boot/grub/grub.cfg on the target device. I suspect I know what 
is happening, but need to retrace my steps to document how to reproduce 
the problem.

<Current>
I downloaded a fresh copy of debian-9.1.0-i386-netinst.iso .
As root I then did:
   dd if=/dev/zero bs=2048 count=1 seek=16 of=/dev/sdb
   Created a msdos partition table using Gparted
   dd if=/home/richard/Downloads/debian-9.1.0-i386-netinst.iso 
of=/dev/sdb bs=4M
    parted /dev/sdb print
RESULTING IN FOLLOWING
   Warning: The driver descriptor says the physical block size is 2048
   bytes, but Linux says it is 512 bytes.
   Ignore/Cancel? i
   Model: Verbatim STORE N GO (scsi)
   Disk /dev/sdb: 31.0GB
   Sector size (logical/physical): 2048B/512B
   Partition Table: mac
   Disk Flags:

   Number  Start   End     Size   File system  Name   Flags
    1      2048B   6143B   4096B               Apple
    2      1942kB  2253kB  311kB               EFI

<Question>
Should have I expected this result?

TIA

[toc] | [next] | [standalone]

#187240

Hi,

Richard Owlett wrote:
> Should have I expected this result?

As far as the parted report is concerned: Yes.


> parted /dev/sdb print
> RESULTING IN FOLLOWING
>  Warning: The driver descriptor says the physical block size is 2048
>  bytes, but Linux says it is 512 bytes.
>   Partition Table: mac

It probably complains that the Apple Partition Map blocksize 2048 is
not the same as the device block size. That's ok. Device block size 512
is expected for the use of the partition tables inside the bootable
debian-cd ISOs.

(It is rather questionable whether the Apple Partition Map is used
 by any boot firmware. Especially since there is no HFS+ partition.)


There is also an MBR (aka "DOS") partition table. Inquire by

  $ /sbin/fdisk -l debian-9.0.0-i386-netinst.iso
  ...
  Device                         Boot Start    End Sectors  Size Id Type
  debian-9.0.0-i386-netinst.iso1 *        0 767999  768000  375M  0 Empty
  debian-9.0.0-i386-netinst.iso2       3868   4475     608  304K ef EFI (FAT-12/16

The type "Empty" of partition 1 is an essential feature, not a bug.
With any other type, EFI would not accept partition 2.


With the 9.0.0 ISO i get from parted

  Number  Start   End     Size   File system  Name   Flags
   1      2048B   6143B   4096B               Apple
   2      1980kB  2292kB  311kB               EFI

The first partition is the Apple Partition Map itself.
The second partition marks the same block range as the EFI partition in the
MBR partition table. But parted's kB are merchant's KB (1000, not 1024),
whereas fdisk reports blocks of 512 bytes.
  3868 * 512 = 1,980,416


Have a nice day :)

Thomas 

[toc] | [prev] | [next] | [standalone]

#187243

On 09/26/2017 06:58 AM, Thomas Schmitt wrote:
> Hi,
>
> Richard Owlett wrote:
>> Should have I expected this result?
>
> As far as the parted report is concerned: Yes.
>

Thank you. I got the same results as you did when running fdisk.
Now I can run my test installs to identify the source of my problems {I 
suspect I know the underlying problem}.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.user

csiph-web