[Linux-cluster] lv and its snapshot can be mounted at the same time in one node?

[Rocty Wang]

Hi, all

	As we know, snapshot of lv(logical volume) is the important
technology implementation of data "backup" and "recovery". And now I use
lvm2's snapshot(also should insmod the dm-snapshot module in kernel) as the
result to deal with data.

	But when I tested XFS or ext3 file system in the standalone LVM2
under Linux, found that if you want to use the snapshot data, you must
umount the original lv firstly and mount the snapshot volume to finish that.

     Due to the cluster project, I use GFS as our storage solution, but when
I do the following operations, it seems that the lv and its snapshot can't
be mounted at the same time? 

Is that any method to mount lv and its snapshot at the same time? Use
lock_gulm?

Or if any file system can support this feature?

Any help will be appreciated.


*****operations:

OS: Linux FC2 (kernel upgraded to 2.6.8.1)
Linux Cluster project: (DEVEL.1098174560 (built Oct 19 2004 16:30:29))
GFS: DEVEL.1098174560 (built Oct 19 2004 16:30:43)
LVM2 version:     2.00.26-cvs (2004-09-29)
dm Library version: 1.00.20-ioctl-cvs (2004-07-03)
Driver version:  4.1.0

***I have create a vg as named "vg2" firstly

# lvcreate -L 8G -n test2 vg2
  Logical volume "test2" created

# lvcreate -s -L 2G -n snap2 /dev/vg2/test2
  Logical volume "snap2" created

# gfs_mkfs -p lock_dlm -t cluster1:myfs -j 2 /dev/vg2/test2

This will destroy any data on /dev/vg2/test2.
Are you sure you want to proceed? [y/n] y

Device:                    /dev/vg2/test2
Blocksize:                 4096
Filesystem Size:           2031440
Journals:                  2
Resource Groups:           32
Locking Protocol:          lock_dlm
Lock Table:                cluster1:myfs

Syncing...
All Done

# mount -t gfs /dev/vg2/test2 /home/test2

GFS: fsid=cluster1:myfs.0: Joined cluster. Now mounting FS...
GFS: fsid=cluster1:myfs.0: jid=0: Trying to acquire journal lock...
GFS: fsid=cluster1:myfs.0: jid=0: Looking at journal...
GFS: fsid=cluster1:myfs.0: jid=0: Done
GFS: fsid=cluster1:myfs.0: jid=1: Trying to acquire journal lock...
GFS: fsid=cluster1:myfs.0: jid=1: Looking at journal...
GFS: fsid=cluster1:myfs.0: jid=1: Done

# mount -t gfs /dev/vg2/snap2 /home/snap2

FAT: bogus number of reserved sectors
VFS: Can't find a valid FAT filesystem on dev dm-3.
GFS: not a GFS filesystem
mount: wrong fs type, bad option, bad superblock on /dev/vg2/snap2,
       or too many mounted file systems