* NFS stale mount after chroot...
@ 2001-12-10 1:57 Daniel Freedman
2001-12-10 2:18 ` Daniel Freedman
2001-12-10 12:06 ` Anders Peter Fugmann
0 siblings, 2 replies; 4+ messages in thread
From: Daniel Freedman @ 2001-12-10 1:57 UTC (permalink / raw)
To: linux-kernel; +Cc: Daniel Freedman
Hi,
It seems like I can generate reproducible stale NFS mounts by mounting
a partition, chroot'ing into that mount, immediately exiting the
chroot, and then finding myself unable to unmount the NFS partition.
I'm pretty sure I've confirmed that nothing is using the partition
(both with fuser and lsof) and even tried to force umount the
partition (which seems like it should definitely umount it, rather
than returning with the same "device is busy" errors), to no avail.
The only method which I've used that seems to be able to get rid of
this NFS mount, is to reboot the NFS client, and clearly that's not a
good one at all. If I'm missing something obvious here, my apologies
in advance. Also, if there's any further information I can provide,
I'd be happy to help. The dump of my procedure follows this message.
Thanks again and take care,
Daniel
--------
On NFS client:
--------
freedman@feynman:/var/space/freedman$ ls -l
total 4
drwxr-xr-x 2 freedman arias 4096 Dec 9 14:46 node1
freedman@feynman:/var/space/freedman$ su
Password:
feynman:/var/space/freedman# mount -t nfs newton:/var/tftpboot-NFS/ ./node1/
feynman:/var/space/freedman# mount
/dev/hda6 on / type ext2 (rw,errors=remount-ro,errors=remount-ro)
proc on /proc type proc (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
/dev/hda5 on /boot type ext2 (rw)
/dev/hda7 on /usr type ext2 (rw)
/dev/hda8 on /var type ext2 (rw)
/dev/hda12 on /var/space type ext2 (rw)
newton:/home on /home type nfs (rw,hard,bg,intr,rsize=8192,wsize=8192,addr=192.168.0.2)
newton:/var/tftpboot-NFS/ on /var/space/freedman/node1 type nfs (rw,addr=192.168.0.2)
feynman:/var/space/freedman# cd node1/
feynman:/var/space/freedman/node1# ls
bin boot cdrom dev etc floppy home initrd lib mnt opt proc root sbin tmp usr var vmlinuz
feynman:/var/space/freedman/node1# cd ..
feynman:/var/space/freedman# chroot node1/
feynman:/# ls -l
total 84
drwxr-xr-x 2 45 45 4096 Dec 8 03:06 bin
drwxr-xr-x 2 45 45 4096 Dec 10 01:05 boot
drwxr-xr-x 2 45 45 4096 Dec 8 03:06 cdrom
drwxr-xr-x 5 45 45 20480 Dec 8 03:07 dev
drwxr-xr-x 38 45 45 4096 Dec 10 01:19 etc
drwxr-xr-x 2 45 45 4096 Dec 8 03:06 floppy
drwxrwsr-x 2 45 45 4096 Nov 28 10:25 home
drwxr-xr-x 2 45 45 4096 Dec 8 03:06 initrd
drwxr-xr-x 4 45 45 4096 Dec 8 03:06 lib
drwxr-xr-x 2 45 45 4096 Nov 28 10:25 mnt
drwxr-xr-x 2 45 45 4096 Dec 8 03:06 opt
drwxr-xr-x 2 45 45 4096 Nov 28 10:25 proc
drwxr-xr-x 2 45 45 4096 Dec 8 03:44 root
drwxr-xr-x 2 45 45 4096 Dec 10 01:05 sbin
drwxrwxrwt 2 45 45 4096 Dec 10 01:07 tmp
drwxr-xr-x 11 45 45 4096 Dec 8 18:34 usr
drwxr-xr-x 14 45 45 4096 Dec 9 22:44 var
lrwxrwxrwx 1 45 45 19 Dec 8 20:26 vmlinuz -> boot/vmlinuz-2.4.16
feynman:/# id
uid=0(root) gid=0(root) groups=0(root)
feynman:/# exit
exit
feynman:/var/space/freedman# mount
/dev/hda6 on / type ext2 (rw,errors=remount-ro,errors=remount-ro)
proc on /proc type proc (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
/dev/hda5 on /boot type ext2 (rw)
/dev/hda7 on /usr type ext2 (rw)
/dev/hda8 on /var type ext2 (rw)
/dev/hda12 on /var/space type ext2 (rw)
newton:/home on /home type nfs (rw,hard,bg,intr,rsize=8192,wsize=8192,addr=192.168.0.2)
newton:/var/tftpboot-NFS/ on /var/space/freedman/node1 type nfs (rw,addr=192.168.0.2)
feynman:/var/space/freedman# cat /proc/mounts
/dev/root.old /initrd cramfs rw 0 0
/dev/root / ext2 rw 0 0
proc /proc proc rw 0 0
devpts /dev/pts devpts rw 0 0
/dev/hda5 /boot ext2 rw 0 0
/dev/hda7 /usr ext2 rw 0 0
/dev/hda8 /var ext2 rw 0 0
/dev/hda12 /var/space ext2 rw 0 0
newton:/home /home nfs rw,v3,rsize=8192,wsize=8192,hard,intr,udp,lock,addr=newton 0 0
newton:/var/tftpboot-NFS/ /var/space/freedman/node1 nfs rw,v3,rsize=8192,wsize=8192,hard,udp,lock,addr=newton 0 0
feynman:/var/space/freedman# umount /var/space/freedman/node1/
umount: /var/space/freedman/node1: device is busy
feynman:/var/space/freedman# fuser ./node1/
feynman:/var/space/freedman# lsof|grep node1
feynman:/var/space/freedman# umount /var/space/freedman/node1/
umount: /var/space/freedman/node1: device is busy
feynman:/var/space/freedman# umount -f /var/space/freedman/node1/
umount2: Device or resource busy
umount: /var/space/freedman/node1: Illegal seek
feynman:/var/space/freedman#
--
Daniel A. Freedman
Laboratory for Atomic and Solid State Physics
Department of Physics
Cornell University
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: NFS stale mount after chroot...
2001-12-10 1:57 NFS stale mount after chroot Daniel Freedman
@ 2001-12-10 2:18 ` Daniel Freedman
2001-12-10 12:06 ` Anders Peter Fugmann
1 sibling, 0 replies; 4+ messages in thread
From: Daniel Freedman @ 2001-12-10 2:18 UTC (permalink / raw)
To: linux-kernel
On Sun, Dec 09, 2001, Daniel Freedman wrote:
>
> Hi,
>
> It seems like I can generate reproducible stale NFS mounts by mounting
> a partition, chroot'ing into that mount, immediately exiting the
> chroot, and then finding myself unable to unmount the NFS partition.
> I'm pretty sure I've confirmed that nothing is using the partition
> (both with fuser and lsof) and even tried to force umount the
> partition (which seems like it should definitely umount it, rather
> than returning with the same "device is busy" errors), to no avail.
> The only method which I've used that seems to be able to get rid of
> this NFS mount, is to reboot the NFS client, and clearly that's not a
> good one at all. If I'm missing something obvious here, my apologies
> in advance. Also, if there's any further information I can provide,
> I'd be happy to help. The dump of my procedure follows this message.
>
> Thanks again and take care,
> Daniel
Arggg... (Continuing above, appending to previous email's client
output, and sorry for leaving off some info...)
Here's my general system info, as well as a counterpoint successful
attempt (to hopefully narrow the search area) at unmounting the same
NFS fs, the only difference is that I didn't do a chroot into that
mount in the interim. Hope this helps as well.
Take care,
Daniel
--------
NFS Client:
________
feynman:/var/space/freedman# uname -a
Linux feynman 2.4.16 #1 SMP Wed Nov 28 12:48:20 EST 2001 i686 unknown
feynman:/var/space/freedman# cat /boot/config-2.4.16 |grep NFS
CONFIG_NFS_FS=m
CONFIG_NFS_V3=y
# CONFIG_ROOT_NFS is not set
CONFIG_NFSD=m
CONFIG_NFSD_V3=y
CONFIG_NCPFS_NFS_NS=y
feynman:/var/space/freedman# cat /proc/sys/kernel/tainted
0
feynman:/var/space/freedman# lsmod
Module Size Used by Not tainted
r128 91608 1
emu10k1 58880 0 (unused)
sound 56940 0 [emu10k1]
soundcore 4036 7 [emu10k1 sound]
ac97_codec 9760 0 [emu10k1]
nfs 73692 1
lockd 48704 1 [nfs]
sunrpc 65108 1 [nfs lockd]
eepro100 17584 1
rtc 6296 0 (autoclean)
unix 16004 19 (autoclean)
ide-disk 6880 8 (autoclean)
ide-probe-mod 8096 0 (autoclean)
ide-mod 132876 8 (autoclean) [ide-disk ide-probe-mod]
ext2 32960 5 (autoclean)
feynman:/var/space/freedman# mount -t nfs newton:/var/tftpboot-NFS/ ./node1/
feynman:/var/space/freedman# cat /proc/mounts
/dev/root.old /initrd cramfs rw 0 0
/dev/root / ext2 rw 0 0
proc /proc proc rw 0 0
devpts /dev/pts devpts rw 0 0
/dev/hda5 /boot ext2 rw 0 0
/dev/hda7 /usr ext2 rw 0 0
/dev/hda8 /var ext2 rw 0 0
/dev/hda12 /var/space ext2 rw 0 0
newton:/home /home nfs rw,v3,rsize=8192,wsize=8192,hard,intr,udp,lock,addr=newton 0 0
newton:/var/tftpboot-NFS/ /var/space/freedman/node1 nfs rw,v3,rsize=8192,wsize=8192,hard,udp,lock,addr=newton 0 0
feynman:/var/space/freedman# cd node1/
feynman:/var/space/freedman/node1# ls
bin boot cdrom dev etc floppy home initrd lib mnt opt proc root sbin tmp usr var vmlinuz
feynman:/var/space/freedman/node1# id
uid=0(root) gid=0(root) groups=0(root)
feynman:/var/space/freedman/node1# cd ..
feynman:/var/space/freedman# umount /var/space/freedman/node1/
feynman:/var/space/freedman# cat /proc/mounts
/dev/root.old /initrd cramfs rw 0 0
/dev/root / ext2 rw 0 0
proc /proc proc rw 0 0
devpts /dev/pts devpts rw 0 0
/dev/hda5 /boot ext2 rw 0 0
/dev/hda7 /usr ext2 rw 0 0
/dev/hda8 /var ext2 rw 0 0
/dev/hda12 /var/space ext2 rw 0 0
newton:/home /home nfs rw,v3,rsize=8192,wsize=8192,hard,intr,udp,lock,addr=newton 0 0
feynman:/var/space/freedman#
--
Daniel A. Freedman
Laboratory for Atomic and Solid State Physics
Department of Physics
Cornell University
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: NFS stale mount after chroot...
@ 2001-12-10 3:42 Tyler BIRD
0 siblings, 0 replies; 4+ messages in thread
From: Tyler BIRD @ 2001-12-10 3:42 UTC (permalink / raw)
To: linux-kernel; +Cc: freedman
I wonder if bash, csh etc believe in this case that your NFS partition is actually
still the root. Try to make sure that the NFS partition is not the root possibly be exiting bash
and loging in again.
I'm probably wrong on this.
Tyler
>>> Daniel Freedman <freedman@ccmr.cornell.edu> 12/09/01 06:57PM >>>
Hi,
It seems like I can generate reproducible stale NFS mounts by mounting
a partition, chroot'ing into that mount, immediately exiting the
chroot, and then finding myself unable to unmount the NFS partition.
I'm pretty sure I've confirmed that nothing is using the partition
(both with fuser and lsof) and even tried to force umount the
partition (which seems like it should definitely umount it, rather
than returning with the same "device is busy" errors), to no avail.
The only method which I've used that seems to be able to get rid of
this NFS mount, is to reboot the NFS client, and clearly that's not a
good one at all. If I'm missing something obvious here, my apologies
in advance. Also, if there's any further information I can provide,
I'd be happy to help. The dump of my procedure follows this message.
Thanks again and take care,
Daniel
--------
On NFS client:
--------
freedman@feynman:/var/space/freedman$ ls -l
total 4
drwxr-xr-x 2 freedman arias 4096 Dec 9 14:46 node1
freedman@feynman:/var/space/freedman$ su
Password:
feynman:/var/space/freedman# mount -t nfs newton:/var/tftpboot-NFS/ ./node1/
feynman:/var/space/freedman# mount
/dev/hda6 on / type ext2 (rw,errors=remount-ro,errors=remount-ro)
proc on /proc type proc (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
/dev/hda5 on /boot type ext2 (rw)
/dev/hda7 on /usr type ext2 (rw)
/dev/hda8 on /var type ext2 (rw)
/dev/hda12 on /var/space type ext2 (rw)
newton:/home on /home type nfs (rw,hard,bg,intr,rsize=8192,wsize=8192,addr=192.168.0.2)
newton:/var/tftpboot-NFS/ on /var/space/freedman/node1 type nfs (rw,addr=192.168.0.2)
feynman:/var/space/freedman# cd node1/
feynman:/var/space/freedman/node1# ls
bin boot cdrom dev etc floppy home initrd lib mnt opt proc root sbin tmp usr var vmlinuz
feynman:/var/space/freedman/node1# cd ..
feynman:/var/space/freedman# chroot node1/
feynman:/# ls -l
total 84
drwxr-xr-x 2 45 45 4096 Dec 8 03:06 bin
drwxr-xr-x 2 45 45 4096 Dec 10 01:05 boot
drwxr-xr-x 2 45 45 4096 Dec 8 03:06 cdrom
drwxr-xr-x 5 45 45 20480 Dec 8 03:07 dev
drwxr-xr-x 38 45 45 4096 Dec 10 01:19 etc
drwxr-xr-x 2 45 45 4096 Dec 8 03:06 floppy
drwxrwsr-x 2 45 45 4096 Nov 28 10:25 home
drwxr-xr-x 2 45 45 4096 Dec 8 03:06 initrd
drwxr-xr-x 4 45 45 4096 Dec 8 03:06 lib
drwxr-xr-x 2 45 45 4096 Nov 28 10:25 mnt
drwxr-xr-x 2 45 45 4096 Dec 8 03:06 opt
drwxr-xr-x 2 45 45 4096 Nov 28 10:25 proc
drwxr-xr-x 2 45 45 4096 Dec 8 03:44 root
drwxr-xr-x 2 45 45 4096 Dec 10 01:05 sbin
drwxrwxrwt 2 45 45 4096 Dec 10 01:07 tmp
drwxr-xr-x 11 45 45 4096 Dec 8 18:34 usr
drwxr-xr-x 14 45 45 4096 Dec 9 22:44 var
lrwxrwxrwx 1 45 45 19 Dec 8 20:26 vmlinuz -> boot/vmlinuz-2.4.16
feynman:/# id
uid=0(root) gid=0(root) groups=0(root)
feynman:/# exit
exit
feynman:/var/space/freedman# mount
/dev/hda6 on / type ext2 (rw,errors=remount-ro,errors=remount-ro)
proc on /proc type proc (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
/dev/hda5 on /boot type ext2 (rw)
/dev/hda7 on /usr type ext2 (rw)
/dev/hda8 on /var type ext2 (rw)
/dev/hda12 on /var/space type ext2 (rw)
newton:/home on /home type nfs (rw,hard,bg,intr,rsize=8192,wsize=8192,addr=192.168.0.2)
newton:/var/tftpboot-NFS/ on /var/space/freedman/node1 type nfs (rw,addr=192.168.0.2)
feynman:/var/space/freedman# cat /proc/mounts
/dev/root.old /initrd cramfs rw 0 0
/dev/root / ext2 rw 0 0
proc /proc proc rw 0 0
devpts /dev/pts devpts rw 0 0
/dev/hda5 /boot ext2 rw 0 0
/dev/hda7 /usr ext2 rw 0 0
/dev/hda8 /var ext2 rw 0 0
/dev/hda12 /var/space ext2 rw 0 0
newton:/home /home nfs rw,v3,rsize=8192,wsize=8192,hard,intr,udp,lock,addr=newton 0 0
newton:/var/tftpboot-NFS/ /var/space/freedman/node1 nfs rw,v3,rsize=8192,wsize=8192,hard,udp,lock,addr=newton 0 0
feynman:/var/space/freedman# umount /var/space/freedman/node1/
umount: /var/space/freedman/node1: device is busy
feynman:/var/space/freedman# fuser ./node1/
feynman:/var/space/freedman# lsof|grep node1
feynman:/var/space/freedman# umount /var/space/freedman/node1/
umount: /var/space/freedman/node1: device is busy
feynman:/var/space/freedman# umount -f /var/space/freedman/node1/
umount2: Device or resource busy
umount: /var/space/freedman/node1: Illegal seek
feynman:/var/space/freedman#
--
Daniel A. Freedman
Laboratory for Atomic and Solid State Physics
Department of Physics
Cornell University
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: NFS stale mount after chroot...
2001-12-10 1:57 NFS stale mount after chroot Daniel Freedman
2001-12-10 2:18 ` Daniel Freedman
@ 2001-12-10 12:06 ` Anders Peter Fugmann
1 sibling, 0 replies; 4+ messages in thread
From: Anders Peter Fugmann @ 2001-12-10 12:06 UTC (permalink / raw)
To: Daniel Freedman; +Cc: linux-kernel
I see the same, though I have not tried to chroot.
It happens when I have run mozilla from a NFS mounted fs.
Both client and server is 2.4.16, but I have had this problem with older
kernels. NFSv3 support is compiled in both kernels.
lsof and fuser reports that no files are in use, but I'm unable to
unmount the filesystem. Btw. The filesystem is mounted ro. (and exported
ro). The exported filesystem is a reiser-fs.
I would really be greatfull if this issue was to be resolved.
Thanks in advance
Anders Fugmann
On 12/10/2001 02:57 AM, Daniel Freedman wrote:
> Hi,
>
> It seems like I can generate reproducible stale NFS mounts by mounting
> a partition, chroot'ing into that mount, immediately exiting the
> chroot, and then finding myself unable to unmount the NFS partition.
> I'm pretty sure I've confirmed that nothing is using the partition
> (both with fuser and lsof) and even tried to force umount the
> partition (which seems like it should definitely umount it, rather
> than returning with the same "device is busy" errors), to no avail.
> The only method which I've used that seems to be able to get rid of
> this NFS mount, is to reboot the NFS client, and clearly that's not a
> good one at all. If I'm missing something obvious here, my apologies
> in advance. Also, if there's any further information I can provide,
> I'd be happy to help. The dump of my procedure follows this message.
>
> Thanks again and take care,
> Daniel
>
>
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2001-12-10 12:07 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2001-12-10 1:57 NFS stale mount after chroot Daniel Freedman
2001-12-10 2:18 ` Daniel Freedman
2001-12-10 12:06 ` Anders Peter Fugmann
-- strict thread matches above, loose matches on Subject: below --
2001-12-10 3:42 Tyler BIRD
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox