zfs discuss - Aug 2010 - send/recv reads a lot from destination zpool

I just copied a snapshot from one zpool (let''s call is
"source") to
another one ("destination") on the same machine using zpool send/recv.
I''m wondering why this process is taking so much bandwidth reading from
"destination", while writing to it, or reading from "source"
did not?
At least this is what "zpool iostat 5" says.

Any idea, anyone?

I saw this the other day when doing an initial "auto sync" from one
Nexenta 3.0.3 node to another (using the ZFS/SSH method). I later tried it again
with a fresh destination pool and the read traffic was minimal. Sadly I
didn''t have an opportunity to do and investigation, but it
doesn''t fit my current model if how things do or should work (which is
troubling).

On 15 Aug 2010, at 15:50, Jerome Warnier <jwarnier at beeznest.net> wrote:
> I just copied a snapshot from one zpool (let''s call is
"source") to
> another one ("destination") on the same machine using zpool
send/recv.
> I''m wondering why this process is taking so much bandwidth reading
from
> "destination", while writing to it, or reading from
"source" did not?
> At least this is what "zpool iostat 5" says.
> 
> Any idea, anyone?
> 
> _______________________________________________
> zfs-discuss mailing list
> zfs-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss