zookeeper意外异常导致袜子打开时关闭

Question

我在Cloudera Hadoop 2.0集群中遇到了不少重大问题，这与zookeeper上每天发生多次的以下错误不谋而合。

我找不到造成这种情况的根本原因。

感谢您的帮助。

2016-04-11 14:48:30,872 ERROR org.apache.zookeeper.server.quorum.LearnerHandler: Unexpected exception causing shutdown while sock still open 2016-04-11 14:48:49,584 ERROR org.apache.zookeeper.server.quorum.LearnerHandler: Unexpected exception causing shutdown while sock still open 2016-04-11 14:49:07,239 ERROR org.apache.zookeeper.server.quorum.LearnerHandler: Unexpected exception causing shutdown while sock still open 2016-04-11 14:49:25,291 ERROR org.apache.zookeeper.server.quorum.LearnerHandler: Unexpected exception causing shutdown while sock still open 2016-04-11 14:49:42,779 ERROR org.apache.zookeeper.server.quorum.LearnerHandler: Unexpected exception causing shutdown while sock still open 2016-04-11 14:50:00,613 ERROR org.apache.zookeeper.server.quorum.LearnerHandler: Unexpected exception causing shutdown while sock still open 2016-04-11 14:50:17,976 ERROR org.apache.zookeeper.server.quorum.LearnerHandler: Unexpected exception causing shutdown while sock still open 2016-04-11 14:50:35,957 ERROR org.apache.zookeeper.server.quorum.LearnerHandler: Unexpected exception causing shutdown while sock still open 2016-04-11 14:50:54,676 ERROR org.apache.zookeeper.server.quorum.LearnerHandler: Unexpected exception causing shutdown while sock still open

Answer 1

这是两层问题：

上述错误发生在领导人选举进行太久的时候。

领导人选举耗时过长的原因是3个动物园管理员服务器中的一个服务器上的数据损坏。

一旦/var/lib/zookeeper中的数据文件被吹走，zookeepper重新启动，领导人选举成功，从而解决了上述问题。