Ceph client cannot connect to cluster problem resolution

1. Description of the problem

after doing the iptables strategy today and restarting one of the machines in the cluster, the input ceph-s discovers the following conditions:

[[email protected] ~]# ceph-s2015-09-10 13:50:57.688516 7f6a6b8cc700  0 monclient (Hunting): Authenticate timed out AF ter 3002015-09-10 13:50:57.688553 7f6a6b8cc700  0 librados:client.admin authentication error (+) Connection timed O Uterror Connecting to Cluster:timedout

2. Problem Analysis

Previously, the cluster was working well and later, due to the addition of iptables-related policies, it was not possible to connect to the cluster via Ceph client and Get
cluster status. The first response is that the relevant IP or port number is not blocked by the policy, causing the packet to fail to reach the relevant process normally. The port number for the Ceph-mon process was not found by the command NETSTAT-NTPL.

3. Handling Methods

in the policy script, add the following:

#添加允许访问的IP地址allowip = (58.220.*.*) #开放ceph Mon port number 6789$ipt-a input-p TCP--dport 6789-j ACCEPT

Execute command NETSTAT-NTPL after re-execution

After executing the ceph-s command

Ceph client cannot connect to cluster problem resolution