Mark writes:
> So the problem only happens when overwriting the live system in either
> Etch or Squeeze. Lenny works exactly as expected...
They all work as expected: unpredictably.
> I could try upgrading the remote server from Etch to Lenny and then do
> it...
And get unpredictable results. W
On Mon, Apr 25, 2011, at 12:04:41 +0200, Klistvud wrote:
> Hmm, perhaps your live system is performing some disk writes *while*
> your dd is doing its thing, thus effectively overwriting
> (corrupting) some sectors already written by dd?
All of the OS versions I was trying in my test environment
On Mon, Apr 25, 2011 at 4:54 PM, Mark Kane wrote:
[...]
> So the problem only happens when overwriting the live system in either
> Etch or Squeeze. Lenny works exactly as expected so I guess I could try
> upgrading the remote server from Etch to Lenny and then do it, but it
> still seems odd that
Dne, 25. 04. 2011 10:54:52 je Mark Kane napisal(a):
On Mon, Apr 25, 2011, at 10:09:58 +0200, Klistvud wrote:
> It would seem we're finally getting somewhere. This should be fairly
> easy to test, namely: performing dd on an *unmounted* (not live)
> system and seeing if it finally works.
Yes, I t
On Mon, Apr 25, 2011, at 10:09:58 +0200, Klistvud wrote:
> It would seem we're finally getting somewhere. This should be fairly
> easy to test, namely: performing dd on an *unmounted* (not live)
> system and seeing if it finally works.
Yes, I tried this yesterday. Using dd to write the image to
On Mon, Apr 25, 2011, at 11:02:40 +0800, Huang, Tao wrote:
> to my understanding, writing with dd on a mounted device leads to
> unpredictable results, and never guarantees to work.
I understand what you mean but it does work 100% of the time on the
versions I mentioned including Lenny, just not E
Dne, 25. 04. 2011 05:02:40 je Huang, Tao napisal(a):
> Yes. Debian is installed and running on sda and I am trying to
> overwrite the current installation with this disk image.
to my understanding, writing with dd on a mounted device leads to
unpredictable results, and never guarantees to work.
On Mon, Apr 25, 2011 at 10:00 AM, Mark Kane wrote:
> On Mon, Apr 25, 2011, at 09:45:44 +0800, Huang, Tao wrote:
>> what do u mean by "live running file system"?
>> a mounted file system?
>
> Yes. Debian is installed and running on sda and I am trying to
> overwrite the current installation with th
On Mon, Apr 25, 2011, at 09:45:44 +0800, Huang, Tao wrote:
> what do u mean by "live running file system"?
> a mounted file system?
Yes. Debian is installed and running on sda and I am trying to
overwrite the current installation with this disk image.
Thanks,
-Mark
--
To UNSUBSCRIBE, email to
On Mon, Apr 25, 2011 at 7:48 AM, Mark Kane wrote:
> On Sun, Apr 24, 2011, at 22:54:48 +0800, Huang, Tao wrote:
[...]
> When using either of the above commands, there is no difference
> outputted from diff between the data written by Lenny (where it works)
> and Squeeze (where it doesn't).
>
> That
On Sun, Apr 24, 2011, at 22:54:48 +0800, Huang, Tao wrote:
> what about a binary diff on those two written images.
>
> i guess the difference should be at the beginning / end of the device.
> just diff their hex value instead of a real binary diff.
> try
>
> $ dd if=/dev/sda bs=1024 count=1 | hex
what about a binary diff on those two written images.
i guess the difference should be at the beginning / end of the device.
just diff their hex value instead of a real binary diff.
try
$ dd if=/dev/sda bs=1024 count=1 | hexdump -x > outfile
or
$ dd if=/dev/sda bs=1024 skip=30719 count=1 | hexd
On Sun, Apr 24, 2011 at 01:15:31AM +1200, Richard Hector wrote:
> On Sat, 2011-04-23 at 21:49 +0900, Osamu Aoki wrote:
> > You know 1M stands for 1*1000*1000*1000 not 1*1024*1024*1024.
>
> I'm half asleep, so I may not be thinking straight, but isn't that one
> too many multiplications, in both ca
On Sat, Apr 23, 2011, at 21:49:39 +0900, Osamu Aoki wrote:
> What was the size of image.img in exact bytes? What happened if you
> specified exact image size via "count"? What happened if bs=1024?
The size of image.img is 31457280 bytes. I have tried several ways
below and listed in parentheses w
On Sat, 2011-04-23 at 21:49 +0900, Osamu Aoki wrote:
> You know 1M stands for 1*1000*1000*1000 not 1*1024*1024*1024.
I'm half asleep, so I may not be thinking straight, but isn't that one
too many multiplications, in both cases?
Richard
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.de
Hi,
On Tue, Apr 19, 2011 at 07:46:06PM -0500, Mark Kane wrote:
> On Wed, Apr 13, 2011, at 07:13:13 -0500, Mark Kane wrote:
> I suppose I should have tested this before for thoroughness, but in
> Debian Lenny it works! So to summarize:
>
> # dd if=image.img of=/dev/sda bs=1M
You know 1M stands fo
On Wed, Apr 13, 2011, at 07:13:13 -0500, Mark Kane wrote:
> One thing to add is that the documentation I'm going off of was
> written back in 2008 and specifically mentions doing this
> successfully from Linux (though not sure which distro the author had
> used). I would think that if it worked bac
On Wed, Apr 13, 2011, at 21:19:43 +1000, Scott Ferguson wrote:
> Upon reflection I'd be willing to bet it's down to the difference
> between versions of coreutils - there's a couple of minor changes up
> until 8.9 (according to NEWS). While most of the changes I've read
> about improve speed and re
On 13/04/11 19:47, Mark Kane wrote:
> On Wed, Apr 13, 2011, at 14:33:34 +1000, Scott Ferguson wrote:
> I could be wrong, but from looking at the script a bit it looks like dd
> is used to create the image like this:
>
> dd of=tmp/disk.img if=/dev/zero count=30720 bs=1k
After a quick squiz I agre
On Wed, Apr 13, 2011, at 14:33:34 +1000, Scott Ferguson wrote:
> It's possible the script uses ddrescue instead of dd to create the
> image
> - though I still don't understand why that would make a difference.
I could be wrong, but from looking at the script a bit it looks like dd
is used to creat
On 13/04/11 13:10, Mark Kane wrote:
> On Wed, Apr 13, 2011, at 10:54:46 +1000, Scott Ferguson wrote:
>> What commands did you use to create the original image?
>>
>> If image.img was created by simply:-
>> dd if=/dev/sda of=/image.img
>> then bs is unnecessary.
>
> I did not create the image direc
On Wed, Apr 13, 2011, at 10:54:46 +1000, Scott Ferguson wrote:
> What commands did you use to create the original image?
>
> If image.img was created by simply:-
> dd if=/dev/sda of=/image.img
> then bs is unnecessary.
I did not create the image directly using dd but rather used a script
which ma
On 12/04/11 22:47, Mark Kane wrote:
> Hi everyone,
>
> I'm trying to write a bootable disk image to a hard drive using dd in
> Debian like so:
>
> dd if=image.img of=/dev/sda bs=1M
>
>
> -Mark
>
>
What commands did you use to create the original image?
If image.img was created by simply:-
On Tue, Apr 12, 2011, at 19:13:39 +0200, Klistvud wrote:
> Perhaps GRUB/LILO just doesn't find it? Additionally -- and I may be
> off target here -- shouldn't that be 'of=/dev/sdax' (a partition, not
> a device)?
Hi and thanks for the reply.
I should have mentioned this in my original message,
Dne, 12. 04. 2011 14:47:08 je Mark Kane napisal(a):
Hi everyone,
I'm trying to write a bootable disk image to a hard drive using dd in
Debian like so:
dd if=image.img of=/dev/sda bs=1M
Perhaps GRUB/LILO just doesn't find it? Additionally -- and I may be
off target here -- shouldn't that be
Hi everyone,
I'm trying to write a bootable disk image to a hard drive using dd in
Debian like so:
dd if=image.img of=/dev/sda bs=1M
dd completes without error and appears to have written this
successfully, however when trying to boot from sda the operating system
does not boot properly as if th
26 matches
Mail list logo
