This section describes how to shut down an entire Ceph cluster. Make sure that your Ceph cluster is healthy before proceeding with the following steps.
Stop all clients that write to Ceph.
Shut down the RADOS Gateway nodes:
salt 'rgw*' system.poweroff
Set the Ceph OSD flags:
ceph osd set noout
ceph osd set nobackfill
ceph osd set norecover
ceph osd set norebalance
ceph osd set nodown
ceph osd set pause
Shut down the Ceph OSD nodes one by one:
salt 'osd01*' system.poweroff
salt 'osd02*' system.poweroff
salt 'osd03*' system.poweroff
...
Shut down the Ceph Monitor nodes one by one:
salt 'cmn01*' system.poweroff
salt 'cmn02*' system.poweroff
salt 'cmn03*' system.poweroff
...