[pve-devel] [PATCH storage 1/2] fix #3873: btrfs: check for correct subvolume taking snapshot

[Fiona Ebner]

Am 09.07.24 um 13:51 schrieb Maximiliano Sandoval:
> Suppose we are doing a snapshot of disk 0 for VM 100. The
> dir_glob_foreach runs over $path=/subvolume/images/100, lists all
> snapshot names and appends their names to the path of the disk, e.g.
> /subvolume/images/vm-100-disk-0 at SNAP_NAME, but the original directory
> $path might contain a second disk `vm-100-disk-1` which is also listed
> by the dir_glib_foreach.
> 
> The patch skips images which reference other disks.
> 
> Signed-off-by: Maximiliano Sandoval <m.sandoval at proxmox.com>
> ---
>  src/PVE/Storage/BTRFSPlugin.pm | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/src/PVE/Storage/BTRFSPlugin.pm b/src/PVE/Storage/BTRFSPlugin.pm
> index 42815cb..dc0420d 100644
> --- a/src/PVE/Storage/BTRFSPlugin.pm
> +++ b/src/PVE/Storage/BTRFSPlugin.pm
> @@ -767,6 +767,9 @@ sub volume_export {
>  	push @$cmd, (map { "$path\@$_" } ($with_snapshots // [])->@*), $path;
>      } else {
>  	dir_glob_foreach(dirname($path), $BTRFS_VOL_REGEX, sub {

I feel like this would be a bit nicer (and slightly more robust) if it
used foreach_subvol() and would check the basename like done in
free_image(). Or even better, introduce a new helper that iterates over
all snapshots of a specific volume to not repeat ourselves here and in
free_image(). Since foreach_subvol() is only used once (i.e. in
free_image()), we could even replace that. There's another pre-existing
thing that's a bit confusing, which is that BTRFS_VOL_REGEX only matches
snapshot volumes and not the actual volume itself AFAICS.

> +	    if (index($path, $_[1]) < 0) {
> +		return
> +	    }
>  	    push @$cmd, "$path\@$_[2]" if !(defined($snapshot) && $_[2] eq $snapshot);
>  	});
>      }