HEALTH_WARN 1 osds down; 1 host (1 osds) down; clock skew detected on mon.comput

admin

HEALTH_WARN 1 osds down; 1 host (1 osds) down; clock skew detected on mon.compute02
OSD_DOWN 1 osds down
    osd.2 (root=default,host=compute03) is down
OSD_HOST_DOWN 1 host (1 osds) down
& D) U% ?" S: |$ n* c    host compute03 (root=default) (1 osds) is down
MON_CLOCK_SKEW clock skew detected on mon.compute02
; v# l6 M, n- i    mon.compute02 clock skew 0.4046s > max 0.05s (latency 0.0719596s)
+ V! x# V3 d0 K7 S. ]) F9 ]

admin

[root@compute01 deploy]# vim ceph.conf
[root@compute01 deploy]# history |grep admin
: v2 e0 J8 a) w* v1 D; l  A  121  ceph-deploy admin iscsi
  122  cat ceph.client.admin.keyring
1 Z9 o; h7 @( ~3 v  576  ceph-deploy admin compute0{1,2,3}
- J+ B3 j! B$ h6 g1 N  577  ceph-deploy --overwrite-conf admin compute0{1,2,3}
1005  history |grep admin
# }" U' F' ]8 M4 Z[root@compute01 deploy]# ceph-deploy --overwrite-conf admin compute0{1,2,3}
* G, `* v/ H3 U& j  r8 c. K$ A+ _: v[ceph_deploy.conf][DEBUG ] found configuration file at: /root/.cephdeploy.conf
[ceph_deploy.cli][INFO  ] Invoked (2.0.1): /usr/bin/ceph-deploy --overwrite-conf admin compute01 compute02 compute03
7 x% g! P$ L; w5 k- a[ceph_deploy.cli][INFO  ] ceph-deploy options:
& Q4 n7 T9 x* H6 e& r# B[ceph_deploy.cli][INFO  ]  username                      : None
[ceph_deploy.cli][INFO  ]  verbose                       : False
' e: x; B4 z- L  x[ceph_deploy.cli][INFO  ]  overwrite_conf                : True
[ceph_deploy.cli][INFO  ]  quiet                         : False
[ceph_deploy.cli][INFO  ]  cd_conf                       : <ceph_deploy.conf.cephdeploy.Conf instance at 0x7f9031dcc638>
[ceph_deploy.cli][INFO  ]  cluster                       : ceph
. L9 I, t1 [- ][ceph_deploy.cli][INFO  ]  client                        : ['compute01', 'compute02', 'compute03']
' t9 J; [% L; v6 ~7 L+ O. r  ?[ceph_deploy.cli][INFO  ]  func                          : <function admin at 0x7f9032664230>
! J6 K9 K8 Y! [% @7 p[ceph_deploy.cli][INFO  ]  ceph_conf                     : None
5 b3 A2 C6 E6 v[ceph_deploy.cli][INFO  ]  default_release               : False
* w6 G9 A7 M- Y$ W$ o! d[ceph_deploy.admin][DEBUG ] Pushing admin keys and conf to compute01
[compute01][DEBUG ] connected to host: compute01
[compute01][DEBUG ] detect platform information from remote host
[compute01][DEBUG ] detect machine type
[compute01][DEBUG ] write cluster configuration to /etc/ceph/{cluster}.conf
[ceph_deploy.admin][DEBUG ] Pushing admin keys and conf to compute02
* M2 ]/ ?$ H) |& r" L5 e5 i[compute02][DEBUG ] connected to host: compute02
* m5 s% Z9 z( ?1 [[compute02][DEBUG ] detect platform information from remote host
[compute02][DEBUG ] detect machine type
[compute02][DEBUG ] write cluster configuration to /etc/ceph/{cluster}.conf
[ceph_deploy.admin][DEBUG ] Pushing admin keys and conf to compute03
, T4 L9 |% K( e7 ?  X8 h0 P* b[compute03][DEBUG ] connected to host: compute03
5 F; U2 e- u4 k9 E2 p* S2 W9 d[compute03][DEBUG ] detect platform information from remote host
[compute03][DEBUG ] detect machine type
& V; V# X3 l4 `1 f9 m[compute03][DEBUG ] write cluster configuration to /etc/ceph/{cluster}.conf
[root@compute01 deploy]# systemctl restart ceph-mon.target

admin

第二步：修改ceph配置中的时间偏差阈值
1. 在admin部署节点修改配置参数：
/ O# z: N( S4 `
# vi ~/my-cluster/ceph.conf
在global字段下添加：
" `. f) V9 C0 Z
mon clock drift allowed = 2
- L% |5 }0 T/ M5 Jmon clock drift warn backoff = 30   
9 F, \, b) W3 L. k1 H2. 向需要同步的mon节点推送配置文件：

' V; r% G7 L7 h4 y6 S# ceph-deploy --overwrite-conf config push node{1..3}
7 ]+ e1 h! V2 S4 V这里是向node1 node2 node3推送，也可以后跟其它不连续节点
  O$ E) D2 N8 m" {5 n  w3. 重启mon服务（centos7环境下）

# systemctl restart ceph-mon.target
4.验证：
) u6 l) r* F( g) G% n% I& q
! }7 S" G5 `0 A4 [# ?: O/ X# ceph -s
显示health_ok说明问题解决