quick (and perhaps foolish) question: MD4sum of CD

[Elton Woo]

How does one check / verify the md5sum of a disk *after* the iso
is burned to the CD?

Elton
-- 
http://setiathome.ssl.berkeley.edu/stats/team/team_4504.html
"You only live once: let's make life BETTER for each other."
LINUX User #193975 [AMD ATHLON CPU] ICQ #149608718.

[Chris Kloiber]

On Fri, 2003-09-26 at 00:05, Elton Woo wrote:
>  How does one check / verify the md5sum of a disk *after* the iso
>  is burned to the CD?
>  
>  Elton 
Assuming the first disk boots properly, when you see the boot: prompt,
type:

boot: linux mediacheck

It will check the disk for you (an md5sum is implanted in each disk) and
it will tell you if any or all the disks are good.

-- 
Chris Kloiber
Red Hat, Inc.

[Vidyut Luther]

make an iso of the CD again and do a md5sum name of new iso ?.
usually tho, when you download the iso is where you get the problem, so
just check the md5sum of the iso you downloaded to the MD5SUM of the
files they gave you.


On Thu, 2003-09-25 at 23:05, Elton Woo wrote:
>  How does one check / verify the md5sum of a disk *after* the iso
>  is burned to the CD?
>  
>  Elton

[Elton Woo]

On September 26, 2003 12:12 am, Vidyut Luther Vidyut Luther 
<vid(a)linuxpowered.com&gt; wrote:
>  make an iso of the CD again and do a md5sum name of new iso ?.
>  usually tho, when you download the iso is where you get the problem, so
>  just check the md5sum of the iso you downloaded to the MD5SUM of the
>  files they gave you. 
I always run "md5sum *.iso" before burning the disk.
What I meant,
was "is there a command to run against the finished disk, before proceeding
to install (or boot the disk)?"

But I guess I can do as Chris Kloiber suggested and do it after booting.

Elton ;-)
-- 
http://setiathome.ssl.berkeley.edu/stats/team/team_4504.html
"You only live once: let's make life BETTER for each other."
LINUX User #193975 [AMD ATHLON CPU] ICQ #149608718.

[Tom Diehl]

On Fri, 26 Sep 2003, Elton Woo wrote:

>  On September 26, 2003 12:12 am, Vidyut Luther Vidyut Luther 
>  <vid(a)linuxpowered.com&gt; wrote:
>  > make an iso of the CD again and do a md5sum name of new iso ?.
>  > usually tho, when you download the iso is where you get the problem, so
>  > just check the md5sum of the iso you downloaded to the MD5SUM of the
>  > files they gave you.
>  I always run "md5sum *.iso" before burning the disk. What I meant,
>  was "is there a command to run against the finished disk, before proceeding
>  to install (or boot the disk)?"
>  
>  But I guess I can do as Chris Kloiber suggested and do it after booting. 
"md5sum /dev/cdrom" The problem with that is that it is possible for it
to fail and the cd still be good or so I am told. I always burn in DAO mode
and have never had problems with the above command failing.

HTH,

-- 
......Tom		Registered Linux User #14522	http://counter.li.org
tdiehl(a)rogueind.com	My current SpamTrap ------->	mtd123(a)rogueind.com

[Mikko Paananen]

On 2003-09-26 at 07:36, Tom Diehl wrote:

>  "md5sum /dev/cdrom" The problem with that is that it is
> possible for it
>  to fail and the cd still be good or so I am told. 
That won't usually work, md5sum will try to read past end of the disc,
generate an error and give you wrong checksum.

You have to read exactly right amount of data, isosize will give the
size of iso9660 filesystem on disc:

dd count=$[`isosize /dev/cdrom`/2048] bs=2048 if=/dev/cdrom | md5sum -

(2048 is the block size on cdrom.)

[Michael Schwendt]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Fri, 26 Sep 2003 11:56:56 +0300, Mikko Paananen wrote:

>  On 2003-09-26 at 07:36, Tom Diehl wrote:
>  
>  > "md5sum /dev/cdrom" The problem with that is that it is possible for it
>  > to fail and the cd still be good or so I am told.
>  
>  That won't usually work, md5sum will try to read past end of the disc,
>  generate an error and give you wrong checksum. 
No. It won't, because Tom mentioned DAO mode (Disk-At-Once) where you
don't get any run-out sectors at the end of the CD. I think in
cdrecord 2.0, it's called SAO mode (Session-At-Once).

>  You have to read exactly right amount of data, isosize will give the
>  size of iso9660 filesystem on disc:
>  
>  dd count=$[`isosize /dev/cdrom`/2048] bs=2048 if=/dev/cdrom | md5sum -
>  
>  (2048 is the block size on cdrom.) 
There's also the "readcd" utility to read an ISO image.


- -- 
Michael, who doesn't reply to top posts and complete quotes anymore.

> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.3 (GNU/Linux)
>
> iD8DBQE/dEyj0iMVcrivHFQRAt+vAJ9xHJNOu379Yx9oCQMfP2jw8gsxhwCeKBsg
> Nr156tcXwTq664uMnrdSpdM=
> =P3Qm
> -----END PGP SIGNATURE-----

[Leonard den Ottolander]

Hello Tom,

>  "md5sum /dev/cdrom" 
 Never realized it could be as easy as that :). I usually do a
dd if=/dev/cdrom | md5sum
 . Difference when using dd is that you could also specify a bs and count  
which will avoid errors about sectors not found at the end of the disk.

>  The problem with that is that it is possible for it to
>  fail and the cd still be good or so I am told. 
 Why would that be? Were you told that as well? The above mentioned errors 
should not cause md5sum to produce a wrong result.

Bye,
Leonard.

--
How clean is a war when you shoot around nukelar waste?
Stop the use of depleted uranium ammo!
End all weapons of mass destruction.

[Robert P. J. Day]

On Fri, 26 Sep 2003, Leonard den Ottolander wrote:

>  Hello Tom,
>  
>  > "md5sum /dev/cdrom"
>  
>  > The problem with that is that it is possible for it to
>  > fail and the cd still be good or so I am told.
>  
>   Why would that be? Were you told that as well? The above mentioned errors 
>  should not cause md5sum to produce a wrong result. 
i'm pretty sure i have had perfectly good CDs produce a bad MD5 sum
when running the above command.  note -- i didn't say their ISO
image files would produce a bad MD5 sum, just running "md5sum"
against the burned CD in the drive.

is it possible that there would be extraneous junk at the tail end
of the CD that would throw off the checksum?

rday

[Leonard den Ottolander]

Hello Robert,

>  > > "md5sum /dev/cdrom" 
>  i'm pretty sure i have had perfectly good CDs produce a bad MD5
> sum
>  when running the above command.
>  
>  is it possible that there would be extraneous junk at the tail end
>  of the CD that would throw off the checksum? 
 Maybe the error output also gets used in the checksum generation. As said, 
I use dd | md5sum for this. In that case I get correct checksums even if I 
do not specify a bs and count and see I/O errors. Probably because the 
error output does not get piped to md5sum.

 Ok, I did a test of md5sum /dev/cdrom[1] on both my cd-player and cd-
writer. I get a correct result with the writer, and no result due to an I/O 
error with the player (ie it chokes at the end of the disk, dd with no 
count works fine).

Bye,
Leonard.

--
How clean is a war when you shoot around nukelar waste?
Stop the use of depleted uranium ammo!
End all weapons of mass destruction.

[Tom Diehl]

On Fri, 26 Sep 2003, Leonard den Ottolander wrote:

>  Hello Tom,
>  
>  > "md5sum /dev/cdrom"
>  
>   Never realized it could be as easy as that :). I usually do a
>  dd if=/dev/cdrom | md5sum
>   . Difference when using dd is that you could also specify a bs and count  
>  which will avoid errors about sectors not found at the end of the disk.
>  
>  > The problem with that is that it is possible for it to
>  > fail and the cd still be good or so I am told.
>  
>   Why would that be? Were you told that as well? The above mentioned errors 
>  should not cause md5sum to produce a wrong result. 
The image is good but the cd has garbage on the back end. It is one of those
things that if the md5sum is good than all is well if it is bad then it might
still be good. :-)

HTH,

-- 
......Tom		Registered Linux User #14522	http://counter.li.org
tdiehl(a)rogueind.com	My current SpamTrap ------->	mtd123(a)rogueind.com

[Robert P. J. Day]

On Fri, 26 Sep 2003, Tom Diehl wrote:

>  The image is good but the cd has garbage on the back end. It is one
> of those
>  things that if the md5sum is good than all is well if it is bad then it might
>  still be good. :-) 
i think the trick is to checksum the ISO image as a file if 
you can, before you burn it to CD.

rday

[Elton Woo]

On September 26, 2003 12:40 pm, Robert P. J. Day "Robert P. J. Day" 
<rpjday(a)mindspring.com&gt; wrote:
>  i think the trick is to checksum the ISO image as a file if
>  you can, before you burn it to CD. 
I *always* run md5sum against the ISOs before burning.
If they fail, I *don't* waste my time burning the CD, but I do another
download. My question was HOWTO verify the CD *after* it is
burnt, but *before* booting it into install mode.

Elton ;-)
-- 
http://setiathome.ssl.berkeley.edu/stats/team/team_4504.html
"You only live once: let's make life BETTER for each other."
LINUX User #193975 [AMD ATHLON CPU] ICQ #149608718.

[Alexandre Oliva]

On Sep 26, 2003, Elton Woo <elwoo(a)videotron.ca&gt; wrote:

>  "is there a command to run against the finished disk, before
>  proceeding to install (or boot the disk)?" 
/usr/lib/anaconda-runtime/checkisomd5

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                 aoliva(a){redhat.com, gcc.gnu.org}
CS PhD student at IC-Unicamp        oliva(a){lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist                Professional serial bug killer

[Elton Woo]

On September 26, 2003 12:40 am, Alexandre Oliva Alexandre Oliva 
<aoliva(a)redhat.com&gt; wrote:
>  > "is there a command to run against the finished disk,
> before
>  > proceeding to install (or boot the disk)?"
>
>  /usr/lib/anaconda-runtime/checkisomd5 
However, I don't have anaconda-runtime
installed ...

Elton ;-)
-- 
http://setiathome.ssl.berkeley.edu/stats/team/team_4504.html
"You only live once: let's make life BETTER for each other."
LINUX User #193975 [AMD ATHLON CPU] ICQ #149608718.