Re: FILESYSTEM CORRUPTION
[Following up on debian-devel]
On Thu, 9 Apr 1998, John Goerzen wrote:
> I reported a similar bug 33 days ago against mount (#19039). It has been
> ignored by the maintainer of mount. I warned then, and I repeat today,
> that this bug CAN and DOES cause filesystem corruption!
>
> This bug relates to PCMCIA support. The PCMCIA refuses to unload if the
> network has not been switched off first, in some cases. However, the
> network refuses to switch off and the PCMCIA shutdown script hangs. I
> have found that if I remove the network card from the computer while it is
> in the PCMCIA shutdown script, it will go a bit further along. This
> happened to me last night, and I removed the card, and it got a but
> further and then hung. I finally shut off the computer, as this has
> happened before with little ill effects.
>
> However, when I turned it on today, my root filesystem was hozed so badly
> that the kernel gave a panic on boot. e2fsck would not fix it without
> removing files en masse. /usr was hozed seriously as well. I am looking
> at a full reinstall here.
>
> FORTUNATELY, /home was clean. (WHEW!) Had that one been messed up, I
> would be in a very serious situation.
>
> Anyway -- To the Mount maintainer -- LOOK AT THE BUG AND FIX IT. To the
> PCMCIA maintainer: This is a bug that you may not be aware of. Let me
> know if you need more details -- you probably do. But since I have to
> reinstall, it may be a few days.
(speaking a 'mount' maintainer)
I agree crash disks aren't fun at all, however from this email and from
your previous bug report, I fail to see where 'mount' is involved in this
infortunate process:
1/ the kernel still doesn't support forced umounts, so doesn't umount
consequently (& unfortunatly); although umount has preliminary '--force'
support (just try umount -f /something), it won't work until the
kernel-side is ready.
When you have run-away or zombies processes with open file descriptors,
it's the kernel that prevents the unmounting.
2/ when rebooting with an unclean filesystem, the '/' is mounted r/o
so e2fsck can be loaded to check all the filesystems BEFORE mount
attempts to mount them r/w.
The problem as you say involves PCMCIA (which fails to unload), you (for
turning off the machine) and the kernel (for panicing), but why would
mount be involved?
Cordialement,
--
- Vincent RENARDIAS vincent@{waw.com,pipo.com,debian.org} -
- Debian/GNU Linux: Pipo: WAW: -
- http://www.fr.debian.org http://www.pipo.com http://www.waw.com -
---------------------------------------------------------------------------
- "La fonctionnalite Son Visuel vous delivre des avertissements visuels." -
- [Message durant l'installation de Windows95]:wq
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: