linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* dedup and receive -p
@ 2016-03-14 12:39 Sylvain Joyeux
  2016-03-15  0:49 ` Duncan
  0 siblings, 1 reply; 2+ messages in thread
From: Sylvain Joyeux @ 2016-03-14 12:39 UTC (permalink / raw)
  To: linux-btrfs

I was trying to find a definitive information about this, but could not ...

AFAIK,  defrag breaks CoW and send/receive -p. I was wondering whether
deduplication would break it too, i.e. if doing a send/receive to
transfer a subvolume, running dedup, and then using said subvolum as a
parent with send -p would work or not.

Sylvain

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: dedup and receive -p
  2016-03-14 12:39 dedup and receive -p Sylvain Joyeux
@ 2016-03-15  0:49 ` Duncan
  0 siblings, 0 replies; 2+ messages in thread
From: Duncan @ 2016-03-15  0:49 UTC (permalink / raw)
  To: linux-btrfs

Sylvain Joyeux posted on Mon, 14 Mar 2016 09:39:51 -0300 as excerpted:

> I was trying to find a definitive information about this, but could not
> ...
> 
> AFAIK,  defrag breaks CoW and send/receive -p. I was wondering whether
> deduplication would break it too, i.e. if doing a send/receive to
> transfer a subvolume, running dedup, and then using said subvolum as a
> parent with send -p would work or not.

I believe it would... if done as you posted.  But this seems blindingly 
obvious to me so I think either the scenario you posted didn't actually 
match what you intended in your head, or there's a misunderstanding on 
either your part or mine.

The better method would be to do the dedup, take a read-only snapshot of 
the subvolume, send/receive it (of course send/receive already works with 
snapshots, but this is making it explicit), and keep that snapshot around 
to do the later send -p from.

You could then continue using the subvolume itself without worrying about 
changing it in ways incompatible with send -p, because you'd be changing 
the subvolume, not the read-only snapshot you took of it for the send, 
which would remain unchanged and thus could be used as the basis of a 
later send -p without a problem.

But like I said using the read-only snapshot for both the original send 
and as the parent in the later send -p seems blindingly obvious to me, so 
I think somewhere along the line some signals got crossed and I'm not 
sure if you didn't write what you actually intended, or if either you or 
I misunderstood something somewhere along the line.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2016-03-15  0:49 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-03-14 12:39 dedup and receive -p Sylvain Joyeux
2016-03-15  0:49 ` Duncan

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).