On 03/09/2014 12:19 PM, Pieter Koorts wrote:
Hello, Just a general question really. What is the recommended node size for Ceph with storage clusters? The Ceph documentation does say use more smaller nodes rather than fewer large nodes but what constitutes to large in terms of Ceph? Is it 16 OSD or more like 32 OSD?
I always recommend: Make sure that one machine that goes down is 10% at max, but try to aim for <5%.
Where does Ceph tail off in performance when having to many OSDs in servers?
When your Journal isn't able to keep up. If you use SSDs for journaling, use 6 OSDs per SSD at max.
For example: (generalisation) 7x 16 Drive/OSD Servers (with MON's mixed in) or 10x 12 Drive/OSD Servers (with MON's mixed in)
This one, since one node failure is 10% You would need two Journal SSDs in this case. Wido
Regards Pieter _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
-- Wido den Hollander 42on B.V. Phone: +31 (0)20 700 9902 Skype: contact42on _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
