zookeeper配置了kerberos之后,zkCli.sh 连接认证死活不通过
连接命令: zkCli.sh
报错如下:
WatchedEvent state:SyncConnected type:None path:null 2017-08-21 10:11:42,054 [myid:] - ERROR [main-SendThread(localhost:2181):ZooKeeperSaslClient@308] - An error: (java.security.PrivilegedActionException: javax.security.sasl.SaslException: GSS initiate failed [Caused by GSSException: No valid credentials provided (Mechanism level: Server not found in Kerberos database (7) - LOOKING_UP_SERVER)]) occurred when evaluating Zookeeper Quorum Member‘s received SASL token. Zookeeper Client will go to AUTH_FAILED state. 2017-08-21 10:11:42,054 [myid:] - ERROR [main-SendThread(localhost:2181):ClientCnxn$SendThread@1072] - SASL authentication with Zookeeper Quorum member failed: javax.security.sasl.SaslException: An error: (java.security.PrivilegedActionException: javax.security.sasl.SaslException: GSS initiate failed [Caused by GSSException: No valid credentials provided (Mechanism level: Server not found in Kerberos database (7) - LOOKING_UP_SERVER)]) occurred when evaluating Zookeeper Quorum Member‘s received SASL token. Zookeeper Client will go to AUTH_FAILED state.
查了很久,终于在kdc的日志中找到了蛛丝马迹,如下
Aug 21 10:11:42 master krb5kdc[21935](info): TGS_REQ (6 etypes {18 17 16 23 1 3}) 192.168.100.144: LOOKING_UP_SERVER: authtime 0, zkcli@NETEASE.COM for zookeeper/localhost@NETEASE.COM, Server not found in Kerberos database
原因分析:1、在zookeeper的认证请求中,zookeeper端的默认principall应该是zookeeper/<hostname>@<realm>
2、当采用zkCli.sh 的方式请求中,默认的host应该是localhost
因此在kdc中才会发现客户端的请求和 zookeeper/localhost@NETEASE.COM 这个principal进行认证,但是在kerberos的database中却没有这个principal。
解决方法: 使用zkCli.sh -server host:port 访问。 同时zookeeper配置文件中sever部分的principal必须为zookeeper/<hostname>@<your realm>
原文地址:http://1992zhong.blog.51cto.com/3963309/1958018