Urgently need some help solving lost space due to snapshots during receive op

[Fabian Keil]

Willem Jan Withagen <wjw at digiware.nl> wrote:

> On 10-8-2015 06:42, Da Rock wrote:
> > I'm trying to move a pool from one system to another - exact same hdd
> > and config, different in other areas. Both register same space
> > available, and so should be no issue. The old system is quite full -
> > still at least 5% free though.

> Are the versions of your FreeBSD also equal?
> 
> Because in newer version, the free space reservation is significantly
> bigger. So if disks are equal size, then with newer FreeBSDs you have
> less usable space available.
> Don't know the exact SVN commit where it happened. But there are more
> question on this topic in the list, and they all boil down to the same
> thing: ZFS needs more reserved space to be able to do certain things
> without freezing the system.
> 
> I've not read any suggestions that you can circumvent this setting.

You can increase the usable space by increasing vfs.zfs.spa_slop_shift:

fk at r500 ~ $sysctl vfs.zfs.spa_slop_shift
vfs.zfs.spa_slop_shift: 6
fk at r500 ~ $zfs list tank
NAME   USED  AVAIL  REFER  MOUNTPOINT
tank   190G  34.4G   136K  /tank
fk at r500 ~ $sudo sysctl vfs.zfs.spa_slop_shift=7
vfs.zfs.spa_slop_shift: 6 -> 7
fk at r500 ~ $zfs list tank
NAME   USED  AVAIL  REFER  MOUNTPOINT
tank   190G  36.2G   136K  /tank

Fabian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 181 bytes
Desc: OpenPGP digital signature
URL: <http://lists.freebsd.org/pipermail/freebsd-fs/attachments/20150810/ecd7f727/attachment.bin>