Re: How to burn Fedora DVDs to avoid readahead bug?

On Saturday 12 May 2007, D. Hugh Redelmeier wrote:
| From: Andre Robatino <andre@xxxxxxxxxxxxxxx>
|
| When I burned the Fedora ISOs to CDs, I used cdrecord with padding to
| avoid the readahead bug - see
|
| http://www.troubleshooters.com/linux/coasterless.htm

Your analysis matches my understanding.

| 3) Find some other tool besides cdrecord that can do the necessary padding
| for DVDs. Are there any?

I pad the iso file, burn the padded file, unpad the iso file, and then
check to see if the burn worked.

You can use the command isosize(8) to find out how much of an image
size is the actual iso9660 file system.

I wrote a utility that I use to pad (and unpad) an image file. I call
it "isopad". It pads by 128k bytes of zero which has been enough for
my systems.

To see if a burn worked, I just use cmp(1) to compare the contents of
the dvd with the image. I only check the filesystem part:
cmp --bytes `isosize whatever.ios` whatever.iso /dev/hdc
This uses the cmp(1) --bytes flag which is handy but not in the man
page.

================ isopad ================
#!/bin/sh

# isopad [+] [-] isofile...
#
# The Linux IDE CD driver in 2.6 tries to read ahead, even past the end of
the # CD or DVD. Even when the program issuing the original read request
was only # trying to read legitimate parts of the disc (albeit near the
end). # The result is spurious I/O errors and read failures.
# It does not seem that this driver bug is going to be fixed soon.
#
# This program is intended to facilitate a workaround. It can pad (or
unpad) # a .iso file so that, when it is burned, the resulting disc will
allow # reads past the end of the content to succeed.
#
# "+" means pad the following .iso files.
# "-" means remove all padding.
# neither means test file and iso sizes.
#
# To see how much readahead is enabled on a drive: hdparm -a /dev/hdc
#
# Why do the padding in place, rather than on a copy of the file?
# .iso files are usually quite large so copying takes a lot of time and
space. #
# Copyright 2005 D. Hugh Redelmeier
# License: GPL
# Version: Sat Jun 18 02:31:48 EDT 2005

# stop at the least sign of trouble
set -u -e

# op is "", "-", or "+": operation to be performed
op=""

for fn
do
case "$fn"
in
"-h"|"--help")
echo "Usage: $0 [-|+|] isofile..."
;;
"+"|"-")
op="$fn"
;;
*)
isosize -x "$fn"
isz=`isosize "$fn"`
fsz=`stat --format='%s' "$fn"`

# conventional block size for CDs
bs=2048

# my guess at a sufficient amount of padding (in blocks)
pb=64

if [ $fsz -lt $isz ]
then
echo "$fn is shorter ($fsz) than it should be ($isz)" >&2
exit 3
elif [ ` expr $fsz % $bs ` -ne 0 ]
then
echo "$fn file size ($fsz) is not a multiple of $bs" >&2
exit 4
elif [ ` expr $isz % $bs ` -ne 0 ]
then
echo "$fn isosize ($isz) is not a multiple of $bs" >&2
exit 5
else
case "$op" in
"")
if [ $fsz -eq $isz ]
then
echo "$fn: isosize == file size == $fsz"
else
echo "$fn: isosize $isz; file size $fsz"
fi
;;
"+")
echo "$fn: padding with $pb blocks of $bs zero bytes"
dd if=/dev/zero bs=$bs count=$pb >>"$fn"
;;
"-")
if [ $fsz -eq $isz ]
then
echo "$fn: already $fsz bytes"
else
echo "$fn: truncating from $fsz to $isz bytes"
dd if=/dev/null of="$fn" seek=$isz bs=1
fi
;;
esac
fi
;;
esac
done
================ end isopad ================

I'm totally bumfuzzled here folks, what with all this talk of padding etc.
AIUI, by definition, the iso IS an image of the disk, in the cases I've played
with, I have never ever come across an iso that had a mod%2048 of other than
zero, 2048 being the block size of the iso9660 file system.

And, in every case where a cd or dvd failed the k3b verify phase, I have been
able to grab a copy of kcalc, divide the size of the .iso as it exists on my
hard drive by 2048 to obtain a 'block count', at which point a

dd if=/dev/cdrom BS=2048 count=the_count_above >md5sum

and the result except for one occasion where my burner was actually doing an
exit stage left, a perfect md5sum for the disk, exactly the same as the hard
drive file.

And the disk of course Just Worked(TM)

So whats all this 'padding' for anyway? AIUI, any padding would break the
iso9660 file format except that it might be properly ignored due to being
completely beyond the the reach of the filesystems ken.

--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Call me bored, but don't call me boring.
-- Larry Wall in <199705101952.MAA00756@xxxxxxxx>

--
fedora-list mailing list
fedora-list@xxxxxxxxxx
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list

Relevant Pages

  • Re: How to burn Fedora DVDs to avoid readahead bug?
    ... I know about growisofs, but unfortunately it's incapable of putting padding on the DVD _after_ the ISO file. ... Fix the readahead bug. ...
    (Fedora)
  • Re: How to burn Fedora DVDs to avoid readahead bug?
    ... | When I burned the Fedora ISOs to CDs, I used cdrecord with padding to avoid ... I pad the iso file, burn the padded file, unpad the iso file, and then ... I wrote a utility that I use to pad an image file. ...
    (Fedora)
  • How to burn Fedora DVDs to avoid readahead bug?
    ... When I burned the Fedora ISOs to CDs, I used cdrecord with padding to avoid the readahead bug - see ... It turns out that the same precaution is necessary with DVDs, and until recently I was able to use cdrecord for this, like ... I know about growisofs, but unfortunately it's incapable of putting padding on the DVD _after_ the ISO file. ...
    (Fedora)
  • Re: "offline" media check of install DVD?
    ... From some experimenting burning the FC6 DVD ISO, I've found that the same bug affects the reading of DVDs, and that it's necessary to add the same type of padding in this case. ... The resulting discs avoid the read-ahead bug even though growisofs-burned ones don't. ...
    (Fedora)
  • Re: Re: Re: Jigdo help
    ... The problem is that jigdo create about a 4.2 Gb iso.tmp file and after exit. ... After I can start jigdo-lite again but don't change anything... ... If I rename the iso.tmp to .iso I can view with windows commander. ...
    (Debian-User)