[ale] And now for my latest problem: stuck mount points with no mount

[Michael B. Trausch]

This is another weird one.

I have several filesystems on my server, all of which are in LVM.
Sometimes, I will have a filesystem that will randomly become read-only.
The normal solution, of course, would be to unmount and remount
read-write.  However, I am apparently unable to do that.

That is, I can unmount the filesystem, and the filesystem disappears
from the mount point it was assigned to, but the mount point becomes
"stuck".  When the filesystem is unmounted, I can no longer remount it:

mbt at saffron:/srv/vm/trausch.us/spicerack.trausch.us$ sudo mount /srv/vm/trausch.us/spicerack.trausch.us/rootfs
[sudo] password for mbt: 
mount: /dev/mapper/saffron.data-spicerack.trausch.us already mounted or /srv/vm/trausch.us/spicerack.trausch.us/rootfs busy

But:

mbt at saffron:/srv/vm/trausch.us/spicerack.trausch.us$ sudo umount /dev/mapper/saffron.data-spicerack.trausch.us
umount: /dev/mapper/saffron.data-spicerack.trausch.us: not mounted
mbt at saffron:/srv/vm/trausch.us/spicerack.trausch.us$ 

Also, I am unable to move the ex-mount point directory out of the way so
that I can remount the filesystem and continue my work:

mbt at saffron:/srv/vm/trausch.us/spicerack.trausch.us$ mv rootfs rootfs.old.2
mv: cannot move `rootfs' to `rootfs.old.2': Device or resource busy

The LV is also still shown as open:

mbt at saffron:/srv/vm/trausch.us/spicerack.trausch.us$ sudo lvdisplay /dev/mapper/saffron.data-spicerack.trausch.us
  --- Logical volume ---
  LV Name                /dev/saffron.data/spicerack.trausch.us
  VG Name                saffron.data
  LV UUID                9clr0W-CaLC-Kogg-B8cp-hlGE-15BY-nx7m3O
  LV Write Access        read/write
  LV Status              available
  # open                 1
  LV Size                9.00 GB
  Current LE             2304
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     256
  Block device           252:18

There is nothing in dmesg to indicate that a filesystem panic occurred,
or that the kernel saw fit to otherwise remount the filesystem read-only
in any automatic fashion.  Also, I can find no indication of any such
occurrences in any of the log files.

This system is running Ubuntu Karmic Server, several LXC containers, and
all ext4 filesystems.

Has anyone seen anything like this before?

	--- Mike

-- 
Blog:  http://mike.trausch.us/blog/
Misc. Software:  http://mike.trausch.us/software/

“The greater danger for most of us lies not in setting our aim too
high and falling short; but in setting our aim too low, and achieving
our mark.” —Michelangelo