Btrfs devel - Jan 2011 - raw fs backup/copy

Is there a reason one can not mount a btrfs and it''s byte-for-byte
copy in parallel, or is the driver just acting silly?

Mini example:
# dd if=/dev/zero of=fs1 bs=1 count=1 seek=$((1024*1024*1024-1))
# mkfs.btrfs fs1
# cp --sparse=always fs1 fs2
# mkdir 1 2
# mount fs1 1 -o loop
# mount fs2 2 -o loop
# touch 1/1
# ls 2
1
# cat /proc/mounts
/dev/loop0 /tmp/test/1 btrfs rw,relatime 0 0
/dev/loop1 /tmp/test/2 btrfs rw,relatime 0 0

--
ivan
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On 01/06/11 19:35, Hugo Mills wrote:
> On Thu, Jan 06, 2011 at 05:36:48PM +0100, Ivan Labáth wrote:
>> Is there a reason one can not mount a btrfs and it''s
byte-for-byte
>> copy in parallel, or is the driver just acting silly?
> 
>    Probably because both filesystems have identical UUIDs (and labels)
> and so will (I suspect) end up being identified as the same filesystem
> by the btrfs code.
> 
>    Hugo.
> 
The reason I ask is, that I had a mirror RAID (btrfs on LVM on LUKS on RAID)
across 2 disks. Then I snatched one to take with me when going away for
holidays, copied some files on to it and when I came back, I wanted to get
to those files. The problem was, I couldn''t mount the fs I took with
me,
because a similar fs was already mounted. As it wasn''t the root fs,
I could unmount it and copy the files on some other fs
(though one could use pivot_root, if it was the root fs).

It seems to me there are cases, where one would want to mount two copies of
originally the same filesystem - think of RAID, VMs, distributing fs images
on a network, LVM snapshots, backups, etc.

--
ivan
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html