On 18/09/2014 13:50, m.channappa.negalur at accenture.com wrote: > Even after setting replication size 3 , my data is not getting replicated on all the 3 nodes. > > Example: > root at Cephadmin:/home/oss# ceph osd map storage check1 > osdmap e122 pool 'storage' (9) object 'check1' -> pg 9.7c9c5619 (9.1) -> up ([0,2,1], p0) acting ([0,2,1], p0) pg 9.7c9c5619 (9.1) -> up ([0,2,1], p0) acting ([0,2,1], p0) Right here it says your data is being replicated for that PG across osd.0 osd.1 osd.3 ([0,2,1]) so yes your data is being replicated across the three nodes. > but here if I shutdown my 2 nodes I will be unable to access data. In actual scenario I should be able to access / write data as my other 3rd node is up (if my understanding is correct). Please let me know where I am wrong. Where are your mons situated? If you have 3 mons across 3 nodes once two are shut down you'll only have 1 mon left, 1/3 will fail quorum and so the cluster will stop taking data to prevent split-brain scenarios. For 2 nodes to be down and the cluster to continue to operate you'd need a minimum of 5 mons or you'd need to move your mons away from your osd's. -Michael
 |