[OmniOS-discuss] Clues for tracking down a drastic ZFS fs space difference?

[Paul B. Henson]

On Wed, Apr 29, 2015 at 04:23:27PM -0400, Chris Siebenmann wrote:

> (On the other hand, maybe 'deferred free' is not blocks that are going
> to be freed but instead blocks serving as indexes to etc, in which case
> this might be a significant number. I don't have data from anything
> else to cross-check against.)

Not sure; you could ask on the zfs developer list, but their first
response will probably be to update to a more current system :).

Here's zdb output from one of my production pools that's not
misbehaving. It's a backup for another box so it pretty much gets zfs
recv all day and it has async destroy enabled:

Blocks  LSIZE   PSIZE   ASIZE     avg    comp   %Total  Type
 4.55K  60.9M   30.1M   90.3M   19.8K    2.02     0.39  deferred free

FWIW :).