unable to parse addrs in 'ceph-mon.ceph'

I am trying to deploy multi node cluster as described in helm documentation. But ceph-osd pods are failing with following errors.

+ exec /usr/local/bin/forego start -f /etc/forego/ceph/Procfile
2017-08-08 21:13:11 /entrypoint.sh: SUCCESS
forego | starting ceph-1.1 on port 5000
ceph-1.1 | starting osd.1 at :/0 osd_data /var/lib/ceph/osd/ceph-1 /var/lib/ceph/osd/ceph-1//journal
ceph-1.1 | server name not found: ceph-mon.ceph (Name or service not known)
ceph-1.1 | unable to parse addrs in 'ceph-mon.ceph'
root@kube-master:~#

I have added dns entries in /etc/resolv.conf in kube-controller-manager-kube-master and also update in all nodes /etc/resolv.conf and /etc/network/interface. Able to resolve ceph-mon.ceph from all nodes. However, it shows different ips when I try to reach it from different ceph-osd nodes.

root@kube-master:~# kubectl -n ceph get pods
NAME READY STATUS RESTARTS AGE
ceph-mds-4005287839-stkrw 1/1 Running 1 36m
ceph-mon-check-1448197789-pkvn7 1/1 Running 1 36m
ceph-mon-n60lh 1/1 Running 1 36m
ceph-mon-zmtl6 1/1 Running 1 36m
ceph-osd-438rc 0/1 CrashLoopBackOff 16 36m
ceph-osd-ztwmp 0/1 CrashLoopBackOff 16 36m
root@kube-master:~#

Can you please help me to isolate the problem?