Managing ceph is a little bit interesting because of how the data is referenced vs how it is actually used. One of the main headaches with ceph is making changes to osds while they are already running. I will go into detail on that later once I’ve had time to properly document it.
Heres how My cluster is configured
- ceph0-mon0 (Mon, Management)
- ceph0-mon1 (Mon)
- ceph0-mon2 (Mon)
- ceph0-n0de0 (OSD)
- ceph0-n0de1 (OSD)
- ceph0-n0de2 (OSD)
In /home/ceph on ceph0-mon0 you will have ceph.conf, this was generated by ceph-deploy originally. Edit this file to adjust your configuration and then run this to distribute that configuration to the monitors and nodes. I am using the increment naming convention. (http://ceph.com/docs/master/rados/deployment/ceph-deploy-admin/)
ceph-deploy –overwrite-conf config push ceph0-mon{0,1,2} ceph0-node{0,1,2}