Re: snapshot destruction making IO extremely slow

[David Sterba <dsterba@suse.cz>]

On Wed, Mar 25, 2015 at 07:38:20AM -0400, Rich Freeman wrote:
> On Wed, Mar 25, 2015 at 6:55 AM, Marc Cousin <cousinmarc@gmail.com> wrote:
> > On 25/03/2015 02:19, David Sterba wrote:
> >> as it reads the pre/post snapshots and deletes them if the diff is
> >> empty. This adds some IO stress.
> >
> > I couldn't find a clear explanation in the documentation. Does it mean
> > that when there is absolutely no difference between two snapshots, one
> > of them is deleted ? And that snapper does a diff between them to
> > determine that ?
> >
> 
> It seems like there should be some supported way of doing a diff on
> two btrfs subvolumes.

> The problem is that we don't have any functionality in kernel space to
> do this (that I'm aware of), and we don't expose the necessary
> information to userspace for it to do this smartly (again, as far as
> I'm aware).

> Maybe there would be some way to do it using btrfs send and parsing the output.

If the subvolumes are read-only than the lightweight send (ioctl with
flag bit set BTRFS_SEND_FLAG_NO_FILE_DATA) could be used and I think
this was the expected usecase.