Slave: received end packet from server, apparent master shutdown:

数据库的一台从产生大量的如下错误： 140121 13:56:07 [Note] Slave: received end packet from server, apparent master shutdown:
140121 13:56:07 [Note] Slave I/O thread: Failed reading log event, reconnecting to retry, log ‘mysql-bin.003062′ at postion 220648071
140121 13:56:09 [Note] Slave: received end packet from server, apparent master shutdown:
140121 13:56:09 [Note] Slave I/O thread: Failed reading log event, reconnecting to retry, log ‘mysql-bin.003062′ at postion 220718334
140121 13:56:10 [Note] Slave: received end packet from server, apparent master shutdown:
140121 13:56:10 [Note] Slave I/O thread: Failed reading log event, reconnecting to retry, log ‘mysql-bin.003062′ at postion 221074239
140121 13:56:12 [Note] Slave: received end packet from server, apparent master shutdown:
原来slave上的my.cnf 配置文件是一台台拷贝的，忘记修改server-id值，导致server-id值相同，切记，切记，server-id值一定保持唯一。
参考：http://nettedfish.sinaapp.com/blog/2013/07/24/how-to-generate-unique-server_id/

这种情况下复制会左右摇摆。当两个从库的server-id相同时，如果从库1已经连接上主库，此时从库2也需要连接到主库，发现之前有server-id相同的连接，就会先注销该连接，然后重新注册。 两台从库不停的注册，不停的注销，会产生很多relay log文件，查看从库状态会看到relay log文件名不停改变，从库的复制状态一会是yes一会是正在连接中