Mailinglist Archive: opensuse-bugs (8114 mails)

< Previous Next >
[Bug 246959] New: ext3 self-destruct on openSUSE 10.2 (kernel-default-2.6.18.2-34)
  • From: bugzilla_noreply@xxxxxxxxxx
  • Date: Tue, 20 Feb 2007 08:23:16 -0700 (MST)
  • Message-id: <bug-246959-21960@xxxxxxxxxxxxxxxxxxxxxxxxx/>
https://bugzilla.novell.com/show_bug.cgi?id=246959

Summary: ext3 self-destruct on openSUSE 10.2 (kernel-default-
2.6.18.2-34)
Product: openSUSE 10.2
Version: Final
Platform: i686
OS/Version: Linux
Status: NEW
Severity: Normal
Priority: P5 - None
Component: Kernel
AssignedTo: kernel-maintainers@xxxxxxxxxxxxxxxxxxxxxx
ReportedBy: matthias.andree@xxxxxx
QAContact: qa@xxxxxxx


I've just had an i686 uniprocessor system (Athlon XP 1800+) effectively wipe
itself out last Sunday.

Software: kernel-default-2.6.18.2-34. e2fsprogs-1.39-21.

Disk layout:

/dev/hda1 vfat (unmounted and empty)
/dev/hda3 ufs (unmounted, FreeBSD 6.2-PRERELEASE)
/dev/hda5 ext3 as /
/dev/hda6 ext3 as /home
/dev/hda7 xfs as /musik

There was a preceding issue where the system lost a few files three weeks ago:
the hard disk drive (Seagate Barracuda ATA IV) remapped 5 sectors from
/dev/hda5 and lost a few files. This was quickly repaired with e2fsck -fy
/dev/hda5, rpm -Va and reinstalling the damaged packages.

The current issue at hand however was violent however. The system remounted the
file system R/O after finding a bitmap mismatch (logging below); several
commands could not be found any more on the system after that event and the
system was unbootable, so I rebooted into the DVD-based (original openSUSE 10.2
box DVD) rescue system and ran e2fsck there. This came up with lots of
inconsistencies in e2fsck -pf /dev/hda, so I ran e2fsck -fy /dev/hda5
(e2fsprogs-1.39-21) which then relocated nearly 107,000 (one hundred and seven
thousand!) files to /lost+found on /dev/hda5. /boot/grub and several /etc files
and executables were missing, rendering the system unbootable, so I moved the
rest into OLD/ and installed Ubuntu 6.10, since I don't trust the openSUSE 10.2
kernel for the nonce, until it's clear what caused this.

I don't think it was e2fsck though, since before the reboot, some commands and
/boot/grub had already gone missing, so I suspect kernel bugs here that
systematically trashed the system.

/dev/hda6 (according to e2fsck -pf) and /dev/hda7 (according to xfs_check) are
undamaged.

These are, in a sense, the final words the system uttered over the network to
the loghost, I haven't found any other suspicious messages after the bootup at
21:14 that day.

Feb 18 21:29:59 rho su: (to beagleindex) root on none
Feb 18 21:30:51 rho su: (to beagleindex) root on none
Feb 18 21:33:49 rho syslogd: /var/log/messages: Read-only file system
Feb 18 21:33:49 rho syslogd: /var/log/warn: Read-only file system
Feb 18 21:33:49 rho syslogd: /var/log/warn: Read-only file system
Feb 18 21:33:49 rho kernel: EXT3-fs error (device hda5): ext3_free_inode: bit
already cleared for inode 1522610
Feb 18 21:33:49 rho kernel: Aborting journal on device hda5.
Feb 18 21:33:49 rho kernel: ext3_abort called.
Feb 18 21:33:49 rho kernel: EXT3-fs error (device hda5): ext3_journal_start_sb:
Detected aborted journal
Feb 18 21:33:49 rho kernel: Remounting filesystem read-only
Feb 18 21:33:49 rho kernel: EXT3-fs error (device hda5) in ext3_delete_inode:
IO failure
Feb 18 21:33:49 rho kernel: __journal_remove_journal_head: freeing
b_committed_data
Feb 18 21:33:49 rho kernel: __journal_remove_journal_head: freeing
b_committed_data


--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

< Previous Next >