Make sure to test that stuff. I've never had to modify the min_size on an EC pool before.
On Wed, Jul 12, 2017 at 11:12 AM Jake Grimmett <jog@xxxxxxxxxxxxxxxxx> wrote:
Hi David,
put that way, the docs make complete sense, thank you!
i.e. to allow writing to a 5+2 EC cluster with one node down:
default is:
# ceph osd pool get ecpool min_size
min_size: 7
to tolerate one node failure, set:
# ceph osd pool set ecpool min_size 6
set pool 1 min_size to 6
to tolerate two nodes failing, set:
# ceph osd pool set ecpool min_size 5
set pool 1 min_size to 5
thanks again!
Jake
On 12/07/17 14:36, David Turner wrote:
> As long as you have the 7 copies online if you're using 7+2 then you can
> still work and read to the EC pool. For EC pool, size is equivalently 9
> and min_size is 7.
>
> I have a 3 node cluster with 2+1 and I can restart 1 node at a time with
> host failure domain.
>
>
> On Wed, Jul 12, 2017, 6:34 AM Jake Grimmett <jog@xxxxxxxxxxxxxxxxx
> <mailto:jog@xxxxxxxxxxxxxxxxx>> wrote:
>
> Dear All,
>
> Quick question; is it possible to write to a degraded EC pool?
>
> i.e. is there an equivalent to this setting for a replicated pool..
>
> osd pool default size = 3
> osd pool default min size = 2
>
> My reason for asking, is that it would be nice if we could build a EC
> 7+2 cluster, and actively use the cluster while a node was off-line by
> setting osd noout.
>
> BTW, Currently testing the Luminous RC, it's looking really nice!
>
> thanks,
>
> Jake
>
>
> _______________________________________________
> ceph-users mailing list
> ceph-users@xxxxxxxxxxxxxx <mailto:ceph-users@xxxxxxxxxxxxxx>
> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
>
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
