cdrecord

NOTE: see bottom for centos

   -r: rock ridge with useful file ownership modes and values (maybe)
        (all files owned by root, and have at least read permission)
        (this is NOT suitable for backup, use -R instead I think)
   -J: generate joliet names (up to 64 chars long),
       so it doesn't come out like crap on windows

% mkisofs -r -J -o test_image.isofs dirname/
# mount -t iso9660 -o ro,loop=/dev/loop0 test_image.isofs test_mountpoint/
       (actually the -t iso9660 is not necessary)
       (and ro, isn't necessary either, it ends up effectively read-only anyhow)
       the following works:
    # mount -o loop test_image.isofs test_mountpoint/

# cdrecord --scanbus  # to get the device number; the following assumes 0,0,0
# umount test_mountpoint/
# cdrecord -dummy -v speed=10000 dev=0,0,0 test_image.isofs
# cdrecord -v speed=10000 dev=0,0,0 test_image.isofs

NOTE: if cdrecord --scanbus says something like:
   cdrecord: No such file or directory. Cannot open '/dev/pg*'. Cannot open SCSI driver.
accompanied by a syslog message:
    Jul 13 01:27:07 ack modprobe: modprobe: Can't locate module char-major-97
It is because ide-scsi is not installed for some reason.
You need to run:
   # lsmod | grep ide
   # rmmod ide-cd
   # insmod ide-scsi
Sometimes there will be a process (the mount process?)
that keeps re-insmodding ide-cd every couple of seconds
(I observed this a long time ago on redhat 6.1;
it doesn't seem to be happening now on redhat 9, although maybe
that's just because I don't have a CD in right now).
To get around that, say (if there re 2 cd devices):
        rmmod ide-cd; insmod ide-cd ignore=hdc ignore=hdd
Note, all my /etc/grub lines say hdc=ide-scsi;
I tried adding hdd=ide-scsi but it didn't help.


NOTE: need to add info about verifying correctness-- use dd?


NOTE: in centos, need to use dev=ATAPI:1,0,0  and it works,
no need to fiddle with the kernel modules.  (but the error message
is mystifying if you don't know this).  Also need -v -v, to get
the nice progress messages.