On Mon, 2007-12-31 at 12:16 -0500, Robert Moskowitz wrote:
>
> William L. Maltby wrote:
> ><snip>
> >
> Over with DSL, I learned that they DON'T use the form: dev=/dev/scd0,
> but rather something like dev=0,0,1 where the 0,0,1 was learned via
> cdrecord -scanbus. So I have been doing the -scanbus and using things
> like dev=5,0,0 and I know that I am hitting the CDRW.
*Just* as an FYI, on my 5.1 the -scanbus showed what seemed invalid
results (-2, -2, -2), but other commands (like cdrecord -prcap ...)
showed bus 0,0,0. But I couldn't access the IDE drive with it. In days
of yore (not your or you're! ;-) I always had used the -scanbus results
successfully. I do know that things changed over time, with some
improvements in the kernel, and there was a time where I could use both
that and the /dev/hdc type of references.
Currently, on my 5.1 IDE, what works is dev=/dev/cdrom or /dev/dvd, etc.
Also, an insmod of sr_mod shows a use count of 0 (->> the same as
yours), indicating that this is apparently not needed anymore.
So, I suggest trying
cdrecord -prcap dev=/dev/<yours> # <yours> = sd? or cdrom or dvd
Because I have not kept up with all the changes over the years, I can't
say what promise this holds, but it *does* work on my 5.1 IDE DVD-+R/W.
As to the GUIs, I *believe* they normally use the symlinks so that they
are not subjected to confusion based on local hardware
configurations/device names.
> Interestingly,
> the first number of the triplet seems to increment every time I unplug
> and plug back in the CDRW.
A disconnect does not immediately unload the modules and their memory
IIRC. Then a connect would/could engender a new bus detection? Just
guessing.
> > I went through each post in the archives in the thread and, using
> > Firefox's search feature, didn't see a post where you showed a "dev="
> > being used with cdrecord. Maybe it's worth a try. Putting a written disc
> > in the ide one should provoke an automount. Then a mount command would
> > show what device is mounted. Then an ls would show which symlinks are
> > pointing to that device. Then you could try cdrecord with a different
> > one to make sure the correct one is accessed.
> >
> definitely using dev= And like I pointed out, with the Storix adding
> -force speed=4 and it does burn the CD. But the HP drive not. I am
> thinking next I try a speed=2 with the HP, as I recall it is a 6x writer.
The -prcap will show a table of allowed write speeds, if the device
(p)rom contains them. FI mine has 31, 15 for cdrom and 4x, 2x for DVD.
It also shows a table of permitted read speeds. I suspect that speeds
not matching those tables would default to one of the allowed values.
> > Presuming that you tried all the things people mentioned in the other
> > thread, the only course I see now is to make sure that the
> > "configuration" on this box matches what is on the others as closely as
> > possible. What modules loaded, /etc/cdrecord.conf, symlinks (allowing
> > for multiple devices or not multiples), GUI config set up correctly
> > (which device), ...
> >
> > The scsi stack needed hasn't changed much in the last few years, as has
> > not the devices used, config files used, ... The major variables I
> > (guess I?) see are multiple devices (opportunity for confusion), USB 1.2
> > (?), configuration files and related, hardware status (good vs.
> > failed)...
> >
> > Have you, using the same cable and usb port, tried another device?
> >
> I have LOTS of cables and use the indiscriminately. I am not careful
> which USB port I use, and I have done this on a few different computers
> loaded with Centos 5. I only have the 2 devices, so I am limited
> there. The Storix is a 'newer' device than the HP, as it is a DVD/CDRW
> and the HP is only a CDRW.
> > A reading of Schilly's docs and man page might suggest something we have
> > overlooked.
> Schilly's doc?
Joerg(?) Schilling, the author of cdrtools. I always forget that
"shily*" references to domains, libraries, etc. is not really part of
his name.
See /usr/share/doc/cdrecord-2.01 for a lot of good info, as well as his
usual rants about the sorry state of scsi in Linux.
>
> I got tired of trying to read the man page, so found a version of it
> online: http://man-wiki.net/index.php/1:cdrecord and so far have not
> found anything else to try.
> <snip sig stuff>
--
Bill
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos
