3 upgrade the ceph cluster,
5 step one ordering: mon.a, osd.0, osd.1, mds.a
6 step two ordering: mon.b, mon.c, osd.2, osd.3
7 ceph expected to be healthy state after each step
12 wait-for-healthy: true
16 daemons: [mon.b, mon.c, mgr.x]
17 wait-for-healthy: true
21 daemons: [osd.0, osd.1]
22 wait-for-healthy: true
25 - ceph.restart: [mds.a]
31 daemons: [osd.2, osd.3]
32 wait-for-healthy: false
33 wait-for-osds-up: true