xfs
[Top] [All Lists]

xfs_fsr wierd bug?

To: linux-xfs@xxxxxxxxxxx
Subject: xfs_fsr wierd bug?
From: beevis@xxxxxxxxx
Date: Wed, 04 Jul 2001 19:07:25 +0200
Sender: owner-linux-xfs@xxxxxxxxxxx
Hello. I am running RH7.1 installed with the sgi boot disk.
I am currently using kernel 2.4.5, compiled with kgcc.
I have tried both patches for xfs 1.0 and 1.01, the same happens.
I am using the latest rpm : xfsdump-1.0.5-0, xfsprogs-1.2.7-0.

The problem:
after trying to defrag a partition with xfs_fsr, /dev/hdc4 ,
and unmounting it, I am unable to remount it without rebooting.
The same DOES NOT happen on other partitions, e.g. hdc2.
To be more precise, I use as root:
xfs_fsr v -t 200 /incoming
and run it until I see the partition is not reorganised anymore.

dmesg reports:

xfs_unmount: xfs_ibusy says error/16
XFS unmount got error 16
linvfs_put_super: vfsp/0xc7f83a60 left dangling!
VFS: Busy inodes after unmount. Self-destruct in 5 seconds.  Have a nice
day...
XFS: Filesystem has duplicate UUID - can't mount
XFS: Filesystem has duplicate UUID - can't mount

and so on, every attempt to remount the filesystem is useless.
xfs_repair creates lost+found and puts in some files.
One has to reboot, and afterwards everything is fine.
I have tried to remount in several manners, with the same result
(the usual: wrong filesystem, too many fs mounted et al.)

I have tried this on other partitions and nothing wierd
happens, I am able to remount it after reorganisation.

Here is the relevant part of /etc/fstab

/dev/hda2               /                       xfs    defaults        1
1
/dev/hdc1               swap                    swap    defaults
0 0
/dev/hda1               swap                    swap    defaults
0 0
/dev/hdc2       /usr/local      xfs      exec,dev,suid,rw 1 1
/dev/hdc3       /home   xfs      exec,dev,suid,rw 1 1
/dev/hdc4       /incoming       xfs      exec,dev,suid,rw 1 1

LILO stanza:

image=/boot/vmlinuz-2.4.5-xfs
        label=5
        initrd=/boot/initrd-2.4.5-xfs.img
        read-only
        root=/dev/hda2
        append="devfs=nomount,ramdisk_size=2500"


I can reproduce easily the problem. The hardware is fine, no bad
messages
about a broken disk and so on.

The hardware:
AMD K6-III 450 MHz, MB Asus P5A, 128M , 2 hd
hda: QUANTUM FIREBALL ST4.3A, ATA DISK drive
hdc: QUANTUM FIREBALL EX6.4A, ATA DISK drive

Thank you

Nicola Fabiano



<Prev in Thread] Current Thread [Next in Thread>
  • xfs_fsr wierd bug?, beevis <=