Consider the following environment. We have a storage array that
provides its own facilities for creating snapshots, replicas, and
clones. We create a Btrfs file system on a volume on one of these
storage arrays, and we mount a snapshot of that volume on the same
host. (The simple use-case is that the snapshot is being used to
restore a set of files.)
We've noticed two problems:
1. The dmesg log shows that the kernel recognizes the Btrfs object id
for both devices. For example:
[ 2920.923517] btrfs: device fsid 54c6e67c-fd51-4aed-a865-4ed4e1cd246a
devid 1 transid 7 /dev/dm-2
[ 2921.007179] btrfs: device fsid 54c6e67c-fd51-4aed-a865-4ed4e1cd246a devid 1
transid 6 /dev/dm-9
If a user process modifies the original file system, the changes also
appear on the snapshot. This is highly undesirable as we want to
preserve the integrity of the snapshot so that it reflects the state
of the volume when the snapshot was created.
2. We cannot mount the original volume read-write and mount the
snapshot read-only.
Is there a mount option for Btrfs that will allow us to mount the
Btrfs file system and ignore the UUID on the device or allow us to
assign a new (random) UUID for the volume? For example, xfs provides
a mount option "-o nouuid". Gfs2 allows mount option "-o
locktable=<UUID>" to specify a new locktable.
Thank you.
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
