Ceph Client报错

最近一段时间在研究Ceph，今天遇到了一个问题，准备写下来跟大家共享。欢迎大家多想我提提意见！

今天Ceph客户端系统出现了一些问题把它重启了，重启之后发现连接不到Ceph集群了，报下列信息：

# ceph --debug -s 
2016-09-01 07:23:31.907282 7f31c1234700  0 librados: client.admin authentication error (95) Operation not supported
Error connecting to cluster: Error

并且radosgw也是Failed的状态，如下：

[root@ceph-client ceph]# systemctl status ceph/*.service
● [email protected] - Ceph rados gateway
   Loaded: loaded (/usr/lib/systemd/system/[email protected]; enabled; vendor preset: disabled)
   Active: failed (Result: start-limit) since Thu 2016-09-01 07:00:44 UTC; 28min ago
  Process: 1099 ExecStart=/usr/bin/radosgw -f --cluster ${CLUSTER} --name client.%i --setuser ceph --setgroup ceph (code=exited, status=5)
 Main PID: 1099 (code=exited, status=5)


Sep 01 07:00:44 ceph-client systemd[1]: [email protected]: main process exited, code=exited, status=5/NOTINSTALLED
Sep 01 07:00:44 ceph-client systemd[1]: Unit [email protected] entered failed state.
Sep 01 07:00:44 ceph-client systemd[1]: [email protected] failed.
Sep 01 07:00:44 ceph-client systemd[1]: [email protected] holdoff time over, scheduling restart.
Sep 01 07:00:44 ceph-client systemd[1]: start request repeated too quickly for [email protected]
Sep 01 07:00:44 ceph-client systemd[1]: Failed to start Ceph rados gateway.
Sep 01 07:00:44 ceph-client systemd[1]: Unit [email protected] entered failed state.
Sep 01 07:00:44 ceph-client systemd[1]: [email protected] failed.

后来检查了/etc/ceph/ceph.conf文件，发现auth_client_required是None的状态。因为配置的ceph集群的认证方式是cephx，所以便将none改为了cephx，然后ceph -s成功：

[root@ceph-client ceph]# ceph -s
    cluster beca06a9-c4c3-443d-9d85-fbc6d620c173
     health HEALTH_OK
     monmap e1: 1 mons at {admin-node=10.0.3.88:6789/0}
            election epoch 9, quorum 0 admin-node
      fsmap e31: 1/1/1 up {0=node1=up:active}
     osdmap e136: 3 osds: 3 up, 3 in
            flags sortbitwise
      pgmap v60704: 220 pgs, 10 pools, 140 MB data, 240 objects
            20432 MB used, 130 GB / 149 GB avail
                 220 active+clean

重启[email protected]服务

[root@ceph-client ceph]# systemctl restart [email protected]
[root@ceph-client ceph]# systemctl status [email protected] 
● [email protected] - Ceph rados gateway
   Loaded: loaded (/usr/lib/systemd/system/[email protected]; enabled; vendor preset: disabled)
   Active: active (running) since Thu 2016-09-01 07:57:43 UTC; 6s ago
 Main PID: 3825 (radosgw)
   CGroup: /system.slice/system-ceph\x2dradosgw.slice/[email protected]
           └─3825 /usr/bin/radosgw -f --cluster ceph --name client.rgw.ceph-client --setuser ceph --setgroup ceph


Sep 01 07:57:43 ceph-client systemd[1]: Started Ceph rados gateway.
Sep 01 07:57:43 ceph-client systemd[1]: Starting Ceph rados gateway...