Ceph系统报时间不同步的问题解决

我的ceph集群运行了一段时间后，报如下警告：

# ceph -s
    cluster c6e7e7d9-2b91-4550-80b0-6fa46d0644f6
     health HEALTH_WARN
            clock skew detected on mon.c
            896 pgs stuck inactive
            896 pgs stuck unclean
            noscrub flag(s) set
            Monitor clock skew detected 
     monmap e1: 5 mons at {a=101.71.4.11:6789/0,b=101.71.4.12:6789/0,c=101.71.4.13:6789/0,d=101.71.4.14:6789/0,e=101.71.4.15:6789/0}
            election epoch 28, quorum 0,1,2,3,4 a,b,c,d,e
     osdmap e1616: 240 osds: 216 up, 216 in
            flags noscrub
      pgmap v16891: 4992 pgs, 18 pools, 1093 GB data, 38340 objects
            5446 GB used, 361 TB / 386 TB avail
                4096 active+clean
                 896 creating

核心信息就是 clock skew detected on mon.c
                  Monitor clock skew detected......

我采用如下办法解决问题：
1. 才每一台MON机器上执行如下命令关闭ntpd服务

service ntpd stop

2. 执行ntpupdate命令进行时间信息同步

[root@gnop029-ct-zhejiang_wenzhou-16-14 ~]# ntpdate us.pool.ntp.org
 5 Dec 16:27:20 ntpdate[30359]: adjust time server 209.118.204.201 offset 0.000712 sec

3. 重新启动ntpd服务

service ntpd start

4.重新启动ceph -s后，发现集群不再报时间问题：

[root@gnop029-ct-zhejiang_wenzhou-16-14 ~]# ceph -s
    cluster c6e7e7d9-2b91-4550-80b0-6fa46d0644f6
     health HEALTH_WARN
            896 pgs stuck inactive
            896 pgs stuck unclean
            noscrub flag(s) set
     monmap e1: 5 mons at {a=101.71.4.11:6789/0,b=101.71.4.12:6789/0,c=101.71.4.13:6789/0,d=101.71.4.14:6789/0,e=101.71.4.15:6789/0}
            election epoch 28, quorum 0,1,2,3,4 a,b,c,d,e
     osdmap e1616: 240 osds: 216 up, 216 in
            flags noscrub
      pgmap v16891: 4992 pgs, 18 pools, 1093 GB data, 38340 objects
            5446 GB used, 361 TB / 386 TB avail
                4096 active+clean
                 896 creating

---------------------------------------------------------------------

可以通过crontab命令将动作放置到周期任务中
crontab -e 
增加如下信息

10 * * * * /usr/sbin/ntpdate us.pool.ntp.org

再重新启动服务

service crond restart
Stopping crond:                                            [  OK  ]
Starting crond:                                            [  OK  ]

转载于:https://my.oschina.net/myspaceNUAA/blog/539674