[sheepdog-users] [sheepdog users] cluster snapshot VS qemu-img
namei.unix at gmail.com
Mon Aug 18 15:31:50 CEST 2014
On Mon, Aug 18, 2014 at 11:55:23AM +0200, Valerio Pachera wrote:
> 2014-08-18 11:11 GMT+02:00 Liu Yuan <namei.unix at gmail.com>:
> > > Restore: 6'01''
> > We use only at most nr_of_nodes * 2 threads for snapshot save, considering
> > not
> > draining the cluster too much.
> > Ok clear.
> > I guess qemu-img use more threads and it doesn't calculating the
> > fingerprint
> > to de-duplicate.
> Repeating for the second time 'cluster snapshot save' with '-m' option, it
> takes 3'37''.
> I guess because fingerprints were already calculated.
> Import is also a bit slower:
> 5'05'' for 'cluster snaphsot load -m'
For now, '-m' can only apply to 'snapshot save'. 'snapshot load' is always
More information about the sheepdog-users