gistfile1.txt

The script reads fro CWD, so keep best to start from known clean.

#ceph-deploy -v new  vm-ds-0{1,2,3}

This creates the new initial monitor members and bootstraps the config

#ceph-deploy -v  mon create  vm-ds-0{1,2,3}

This puts on the monitor keyrings etc - creates the initial monitor map
{wait for a minute or two, keys get generated so needs some entropy}

#ceph-deploy -v gatherkeys vm-ds-01

Now you should be able to run 

#ceph -s
It should return something about the cluster health (proably WARN as there are no OSDs yet).. this is good. If you get an auth error, see me.

Now to install the OSDs. This step creates a GPT labeled disk and one XFS partition per disk. Disks used are /dev/sd{b,c,d,e,f,g}

for n in 1 2 3 4 5 6  
do
 ceph-deploy -v  osd create --zap-disk vm-ds-0$n:/dev/sdb vm-ds-0$n:/dev/sdc vm-ds-0$n:/dev/sdd vm-ds-0$n:/dev/sde vm-ds-0$n:/dev/sdf vm-ds-0$n:/dev/sdg
done

This takes a while.. got for a brew


When done, you should see something like

#ceph -s

root@vm-ds-01:~# ceph -s
   health HEALTH_OK
   monmap e1: 3 mons at {vm-ds-01=192.168.2.1:6789/0,vm-ds-02=192.168.2.2:6789/0,vm-ds-03=192.168.2.3:6789/0}, election epoch 16, quorum 0,1,2 vm-ds-01,vm-ds-02,vm-ds-03
   osdmap e2787: 36 osds: 36 up, 36 in
    pgmap v325238: 1392 pgs: 1392 active+clean; 73940 MB data, 221 GB used, 66778 GB / 66999 GB avail; 0B/s rd, 144KB/s wr, 44op/s
   mdsmap e1: 0/0/1 up