> >>> This is RH7.3. I filed a bugsy wiht RH
> >>> but I was wondering if anyone else has seen this.
> >>> /dan
> >> I don't see a bug. It is the user's responsibility to
> >> configure/customize updfstab, is this right? It hardly matters what
> >> fstab calls a file system as long as it's mountable.
> >> Yuan Liu
> > No, it is a bug. updfstab probes the system for Zip drives, JAZZ,
> > cdroms, etc. It then enters the appropriate mount info into /etc/fstab.
> > There's no reason to modify updfstab configuration file for normal
> > things like ZIP drives etc.
> > It found the ZIP drive but this is what it put in fstab:
> > /dev/sdb4 /mnt/zip250.0 auto noauto,owner,kudzu 0 0
> > but /dev/sdb is the MO drive not the ZIP drive; the entry should be
> > /dev/sdc4 /mnt/zip250.0 auto noauto,owner,kudzu 0 0
> Yes, I understand that this is a nuisance - and just that. You can
> still access your MO drive under the mount point /mnt/zip250.0, right?
> Compare what dmesg says about your MO drive with entries in
> updfstab.conf.default (preshipped updfstab.conf contains only one
> include statement to refer to this default) and you'll see it's your MO
> drive that told kernel modules something that confused updfstab.
Nothing that would/should confuse the kernel. The MO is an HP C1113J
(extremely common). Vendor ID:HP Desc:C1113J Type:Optical. Also there's
nothing in updfstab.conf (RH 7.3) that vaguely resembles an MO or the
desc string of the MO. (I would assume that the "J" in C1113J isn't confused
with "jaz") According the man page upsftab only recognizes three classes:
hd, cdrom and floppy.
So, I guess I'll have to modify updfstab.conf. But, it's still a bug.
We sell turnkey system and I really hate to have to tweak things that
we didn't add or break.
> I would like my CD-RW mounted at /mnt/cd-rw because I also have a DVD
> drive (which I prefer to be called /mnt/dvdrom). Having coexisting
> /mnt/cdrom and /mnt/cdrom1 had been very frustrating at times. Do you
> consider this a bug, too? (Unlike the former Sun, HP, DEC, and IBM's own
> hardware spaces, IMO it's impractical to have one vendor to ship a
> default table that distinguishes all devices used in today's x86
> architecture, even though they created the monster. One thing RH can do
> is perhaps be more restrictive in default updfstab.conf so untested
> hardware won't appear and force users to configure themselves. Sorta
> defeated the purpose of kudzu but, hey, the*driver is there.)
> Yuan Liu
> > /dan