[vox-tech] Debian Net Install Question

[Peter Jay Salzman]

On Thu 06 Jan 05,  9:15 AM, Robert G. Scofield <rscofield at afes.com> said:
> On Wednesday 05 January 2005 17:51, Rick Moen wrote:
> > Quoting Robert G. Scofield (rscofield at afes.com):
> > > So I downloaded this file:  sarge-i386-netinst.iso
> >
> > Also, download the matching md5sum.  Then run "md5sum" on the ISO, and
> > compare.  _Then_ burn it, and try to boot it.
> >
> 
> Before you read this, I want everyone to know I have been reading man md5sum 
> and info md5sum.  This stuff is confusing.  Here's what I've found:
> 
> 1)  If I type "md5sum sarge-i386-netinst.iso | mdsum -c", I get this nice 
> message:  "sarge-i386-netinst.iso: OK"
> 
> 2)  If I type "md5sum -c sarge-i386-netinst.iso", a bunch of stuff flys across 
> the screen ending with this not nice message:  "md5sum: WARNING: 906 of 906 
> listed files could not be read"
> 
> 3)  If I type "md5sum -c sarge-i386-netinst.iso | more", I appear to get the 
> same type of information flying across the screen as in (2); things like 
> files can't be opened or read or found, but at the end there is no final 
> message about 906 of 906 files not being read.
> 
> 4)  Peter's great command, mount -o loop sarge-i386-netinst.iso /mnt, leads to 
> this titillating file in sarge-i386-netinst.iso, "md5sum.txt."  The file 
> md5sum.txt lists hundreds and hundreds of files.  I guess what I'm supposed 
> to do is compare the file list in md5sum.txt to the output of some md5sum 
> command.  I suppose I could print md5sum.txt out and compare it to some list 
> of files that the command md5sum gives me.  But so far md5sum has not given 
> me a list like that found in md5sum.txt.  So far md5sum has given me the 
> three outputs listed in (1), (2), and (3) above.  A fourth output, from 
> "md5sum sarge-i386-netinst.iso" gives me a long code followed by the name of 
> the iso file.
> 
> Can anyone suggest what I should do next?
> 
> Thank you.
> 
> Bob

Hi Bob,

You need to do two things:


1. Get the md5sum for the entire ISO file by doing:

      $ md5sum sarge-i386-netinst.iso

   The output will look something like:

      6c190153f506bdf1474c728b8e5fcb90  sarge-i386-netinst.iso


2. Get what the md5sum *should* be for that file.  This may or may not be
   tricky unless you remember where you downloaded the ISO file from.  If
   you do remember, go back to where you downloaded the file from and
   look for a file called

      sarge-i386-netinst.041124.iso.md5.txt

   This file should contain what the md5sum of sarge-i386-netinst.iso
   should be.  If the two md5sums are the same, then you're good to go.
   If they're different, you have an incomplete/bad copy of the file.

   If you don't remember where you downloaded the file from, I say it
   "might" be tricky because there are many versions of
   sarge-i386-netinst.iso.  The last one appears to have been released
   15-Dec-2004 and should have an md5sum of

      6418fafe99f5e686283582833530e1c3

   There's also one that was released 25-Nov-2004 which should have an
   md5sum of

      cad07d235677fdf4ca931a1395206aba



What you may not realize is that md5sum prints a (semi) unique string for a
file.  Remember I mentioned that an ISO is a filesystem within a file?  That
means it makes sense to talk of the md5sum of an ISO file in two different
ways:

   1. You can talk about the md5sum of all the individual files inside
      the ISO file.  There will be many.

   2. You can talk about the md5sum of the ISO file itself.  There will
      be just one.

You don't want to compute md5sums for all the files inside the ISO; it's not
necessary.  It would only be necessary if you *KNEW* one of the files was
corrupted, and wanted to figure out which one.

If you just want to know if the entire ISO is okay or not, you just need to
compare the md5sum of the entire ISO file with what the md5sum of the ISO
file should be.

Pete

-- 
The mathematics of physics has become ever more abstract, rather than more
complicated.  The mind of God appears to be abstract but not complicated.
He also appears to like group theory.  --  Tony Zee's "Fearful Symmetry"

GPG Fingerprint: B9F1 6CF3 47C4 7CD8 D33E  70A9 A3B9 1945 67EA 951D