MySQL5.6 新特性之GTID【转】

转自

MySQL5.6 新特性之GTID - jyzhou - 博客园
http://www.cnblogs.com/zhoujinyi/p/4717951.html

背景：

      MySQL5.6在5.5的基础上增加了一些改进，本文章先对其中一个一个比较大的改进"GTID"进行说明。

概念：

      GTID即全局事务ID（global transaction identifier），GTID实际上是由UUID+TID组成的。其中UUID是一个MySQL实例的唯一标识。TID代表了该实例上已经提交的事务数量，并且随着事务提交单调递增，所以GTID能够保证每个MySQL实例事务的执行（不会重复执行同一个事务，并且会补全没有执行的事务）。下面是一个GTID的具体形式：

4e659069-3cd8-11e5-9a49-001c4270714e:1-77

更具体的说明见官方说明。

GTID意义：

      引入GTID的意义是什么？

      1)因为清楚了GTID的格式，所以通过UUID可以知道这个事务在哪个实例上提交的。

      2)通过GTID可以极方便的进行复制结构上的故障转移，新主设置。很好的解决了下面这个图（图来自高性能MySQL第10章）的问题。

上面图的意思是：Server1(Master)崩溃，根据从上show slave status获得Master_log_File/Read_Master_Log_Pos的值，Server2(Slave)已经跟上了主，Server3(Slave)没有跟上主。这时要是把Server2提升为主，Server3变成Server2的从。这时在Server3上执行change的时候需要做一些计算，这里就不做说明了，具体的说明见高性能MySQL第10章，相对来说是比较麻烦的。

这个问题在5.6的GTID出现后，就显得非常的简单。由于同一事务的GTID在所有节点上的值一致，那么根据Server3当前停止点的GTID就能定位到Server2上的GTID。甚至由于MASTER_AUTO_POSITION功能的出现，我们都不需要知道GTID的具体值，直接使用CHANGE MASTER TO MASTER_HOST='xxx'， MASTER_AUTO_POSITION命令就可以直接完成failover的工作。

原理：

      从服务器连接到主服务器之后，把自己执行过的GTID(Executed_Gtid_Set)<SQL线程> 、获取到的GTID(Retrieved_Gtid_Set）<IO线程>发给主服务器，主服务器把从服务器缺少的GTID及对应的transactions发过去补全即可。当主服务器挂掉的时候，找出同步最成功的那台从服务器，直接把它提升为主即可。如果硬要指定某一台不是最新的从服务器提升为主， 先change到同步最成功的那台从服务器， 等把GTID全部补全了，就可以把它提升为主了。

测试：

1）复制环境的搭建：具体的复制搭建的步骤可以在网上搜索

因为支持GTID，所以5.6多了几个参数：

mysql> show variables like '%gtid%';
+---------------------------------+-----------+
| Variable_name                   | Value     |
+---------------------------------+-----------+
| binlog_gtid_simple_recovery     | OFF       |
| enforce_gtid_consistency        | OFF       |
| gtid_deployment_step            | OFF       |
| gtid_executed                   |           |
| gtid_mode                       | OFF       |
| gtid_next                       | AUTOMATIC |
| gtid_owned                      |           |
| gtid_purged                     |           |
| simplified_binlog_gtid_recovery | OFF       |
+---------------------------------+-----------+

主从环境的搭建和5.5没有什么区别，唯一需要注意的是：开启GTID需要启用这三个参数：

#GTID
gtid_mode = on
enforce_gtid_consistency = 1
log_slave_updates   = 1

任意一个参数不开启则都会报错：

2015-08-09 02:33:57 6512 [ERROR] --gtid-mode=ON or UPGRADE_STEP_1 or UPGRADE_STEP_2 requires --log-bin and --log-slave-updates
2015-08-09 02:33:57 6512 [ERROR] Aborting2015-08-09 02:39:58 9860 [ERROR] --gtid-mode=ON or UPGRADE_STEP_1 requires --enforce-gtid-consistency
2015-08-09 02:39:58 9860 [ERROR] Aborting

具体的方法可以参考官方文档。

三个实例开启之后(3306、3307、3308)，执行change的时候也要注意：

各个实例的uuid：

3306:
mysql> select @@server_uuid;
+--------------------------------------+
| @@server_uuid                        |
+--------------------------------------+
| 4e659069-3cd8-11e5-9a49-001c4270714e |
+--------------------------------------+3307:
mysql> select @@server_uuid;
+--------------------------------------+
| @@server_uuid                        |
+--------------------------------------+
| 041d0e65-3cde-11e5-9a6e-001c4270714e |
+--------------------------------------+3308:
mysql> select @@server_uuid;
+--------------------------------------+
| @@server_uuid                        |
+--------------------------------------+
| 081ccacf-3ce4-11e5-9a95-001c4270714e |
+--------------------------------------+

使用5.6之前的主从change：

mysql> change master to master_host='127.0.0.1',master_user='rep',master_password='rep',master_log_file='mysql-bin3306.000001',master_log_pos=151,/*master_auto_position=1*/;

报错：

ERROR 1776 (HY000): Parameters MASTER_LOG_FILE, MASTER_LOG_POS, RELAY_LOG_FILE and RELAY_LOG_POS cannot be set when MASTER_AUTO_POSITION is active.

当使用 MASTER_AUTO_POSITION 参数的时候，MASTER_LOG_FILE，MASTER_LOG_POS参数不能使用。

使用5.6之后的主从change：

mysql> change master to master_host='127.0.0.1',master_user='rep',master_password='rep',master_port=3306,master_auto_position=1;

在执行上面的命令的时候会报错2个warnings，主要的原因是复制账号安全的问题，相关的信息可以看这里。

从总体上看来，由于要支持GTID，所以不需要手工确定主服务器的MASTER_LOG_FILE及MASTER_LOG_POS。要是不需要GTID则需要指定FILE和POS。在2个从上执行上面命令，到此主从环境搭建完成。GTID的主从完成之后可以通过show processlist查看：

mysql> show processlist\G;
*************************** 1. row ***************************Id: 38User: repHost: localhost:52321db: NULLCommand: Binlog Dump GTID   #通过GTID复制Time: 48State: Master has sent all binlog to slave; waiting for binlog to be updatedInfo: NULLRows_sent: 0
Rows_examined: 0

2）测试复制的故障转移

server1(3306)挂了，服务器起不来了。需要把其中的一个从设置为主，另一个设置为其的从库：

server2(3307):

              Master_Log_File: mysql-bin3306.000002Read_Master_Log_Pos: 4156773Exec_Master_Log_Pos: 4156773

server3(3308):

              Master_Log_File: mysql-bin3306.000001Read_Master_Log_Pos: 83795320Exec_Master_Log_Pos: 83795320

相比之下server2完成的事务要比server3更接近或则等于server1，现在需要把server3设置为server2的从库。

在MySQL5.6之前，这里的计算会很麻烦，要计算之前主库的log_pos和当前要设置成主库的log_pos，很有可能出错。所以出现了一些高可用性的工具如MHA，MMM等解决问题。

在MySQL5.6之后，很简单的解决了这个难题。因为同一事务的GTID在所有节点上的值一致，那么根据server3当前停止点的GTID就能定位到server2上的GTID，所以直接在server3上执行change即可：

mysql> stop slave;
Query OK, 0 rows affected (0.02 sec)

#千万不要执行 reset master，否则会从最先的GTID上开始执行。mysql> change master to master_host='127.0.0.1',master_user='rep',master_password='rep',master_port=3307,master_auto_position=1; #指定到另一个比较接近主的从上。
Query OK, 0 rows affected, 2 warnings (0.02 sec)mysql> start slave;  #成功的切换到新主
Query OK, 0 rows affected (0.03 sec)

主从结构已经变更，server2是Master，server3是Slave。因为不需要计算pos的值，所以通过GTID很简单的解决了这个问题。

3）跳过复制错误：gtid_next、gtid_purged

① 从服务器跳过一个错误的事务:

mysql> show slave status\G;
*************************** 1. row ***************************Slave_IO_State: Waiting for master to send eventMaster_Host: 127.0.0.1Master_User: repMaster_Port: 3306Connect_Retry: 60Master_Log_File: mysql-bin3306.000001Read_Master_Log_Pos: 38260944Relay_Log_File: mysqld-relay-bin3307.000002Relay_Log_Pos: 369Relay_Master_Log_File: mysql-bin3306.000001Slave_IO_Running: YesSlave_SQL_Running: NoReplicate_Do_DB: Replicate_Ignore_DB: Replicate_Do_Table: Replicate_Ignore_Table: Replicate_Wild_Do_Table: Replicate_Wild_Ignore_Table: Last_Errno: 1008Last_Error: Error 'Can't drop database 'mablevi'; database doesn't exist' on query. Default database: 'mablevi'. Query: 'drop database mablevi'Skip_Counter: 0Exec_Master_Log_Pos: 151Relay_Log_Space: 38261371Until_Condition: NoneUntil_Log_File: Until_Log_Pos: 0Master_SSL_Allowed: NoMaster_SSL_CA_File: Master_SSL_CA_Path: Master_SSL_Cert: Master_SSL_Cipher: Master_SSL_Key: Seconds_Behind_Master: NULL
Master_SSL_Verify_Server_Cert: NoLast_IO_Errno: 0Last_IO_Error: Last_SQL_Errno: 1008Last_SQL_Error: Error 'Can't drop database 'mablevi'; database doesn't exist' on query. Default database: 'mablevi'. Query: 'drop database mablevi'Replicate_Ignore_Server_Ids: Master_Server_Id: 1Master_UUID: 4e659069-3cd8-11e5-9a49-001c4270714eMaster_Info_File: mysql.slave_master_infoSQL_Delay: 0  #通过在change的时候指定，如：change master to master_delay=600，延迟10分钟同步。SQL_Remaining_Delay: NULLSlave_SQL_Running_State: Master_Retry_Count: 86400Master_Bind: Last_IO_Error_Timestamp: Last_SQL_Error_Timestamp: 150810 23:38:39Master_SSL_Crl: Master_SSL_Crlpath: Retrieved_Gtid_Set: 4e659069-3cd8-11e5-9a49-001c4270714e:1-48Executed_Gtid_Set: Auto_Position: 1

在MySQL5.6之前，只需要执行：

mysql> set global sql_slave_skip_counter=1; 

跳过一个错误的事务，就可以继续进行复制了。但在MySQL5.6之后则不行：

mysql> set global sql_slave_skip_counter=1;
ERROR 1858 (HY000): sql_slave_skip_counter can not be set when the server is running with @@GLOBAL.GTID_MODE = ON. Instead, for each transaction that you want to skip, generate an empty transaction with the same GTID as the transaction

分析：因为是通过GTID来进行复制的，也需要跳过这个事务从而继续复制，这个事务可以到主上的binlog里面查看：因为不知道找哪个GTID上出错，所以也不知道如何跳过哪个GTID。但在show slave status里的信息里可以找到在执行Master里的POS:151

Exec_Master_Log_Pos: 151

的时候报错，所以通过mysqlbinlog找到了GTID：

# at 151
#150810 22:57:45 server id 1  end_log_pos 199 CRC32 0x5e14d88f     GTID [commit=yes]
SET @@SESSION.GTID_NEXT= '4e659069-3cd8-11e5-9a49-001c4270714e:1'/*!*/;

找到这个GTID之后执行：必须按照下面顺序执行

mysql> stop slave;
Query OK, 0 rows affected (0.01 sec)mysql> set session gtid_next='4e659069-3cd8-11e5-9a49-001c4270714e:1';  #在session里设置gtid_next，即跳过这个GTID
Query OK, 0 rows affected (0.01 sec)mysql> begin;      #开启一个事务
Query OK, 0 rows affected (0.00 sec)mysql> commit;
Query OK, 0 rows affected (0.01 sec)mysql> SET SESSION GTID_NEXT = AUTOMATIC;   #把gtid_next设置回来
Query OK, 0 rows affected (0.00 sec)mysql> start slave;  #开启复制
Query OK, 0 rows affected (0.01 sec)

查看复制状态：

mysql> show slave status\G;
*************************** 1. row ***************************Slave_IO_State: Waiting for master to send eventMaster_Host: 127.0.0.1Master_User: repMaster_Port: 3306Connect_Retry: 60Master_Log_File: mysql-bin3306.000001Read_Master_Log_Pos: 38260944Relay_Log_File: mysqld-relay-bin3307.000003Relay_Log_Pos: 716Relay_Master_Log_File: mysql-bin3306.000001Slave_IO_Running: YesSlave_SQL_Running: YesReplicate_Do_DB: Replicate_Ignore_DB: Replicate_Do_Table: Replicate_Ignore_Table: Replicate_Wild_Do_Table: Replicate_Wild_Ignore_Table: Last_Errno: 0Last_Error: Skip_Counter: 0Exec_Master_Log_Pos: 38260944Relay_Log_Space: 38261936Until_Condition: NoneUntil_Log_File: Until_Log_Pos: 0Master_SSL_Allowed: NoMaster_SSL_CA_File: Master_SSL_CA_Path: Master_SSL_Cert: Master_SSL_Cipher: Master_SSL_Key: Seconds_Behind_Master: 0
Master_SSL_Verify_Server_Cert: NoLast_IO_Errno: 0Last_IO_Error: Last_SQL_Errno: 0Last_SQL_Error: Replicate_Ignore_Server_Ids: Master_Server_Id: 1Master_UUID: 4e659069-3cd8-11e5-9a49-001c4270714eMaster_Info_File: mysql.slave_master_infoSQL_Delay: 0  #延迟同步SQL_Remaining_Delay: NULLSlave_SQL_Running_State: Slave has read all relay log; waiting for the slave I/O thread to update itMaster_Retry_Count: 86400Master_Bind: Last_IO_Error_Timestamp: Last_SQL_Error_Timestamp: Master_SSL_Crl: Master_SSL_Crlpath: Retrieved_Gtid_Set: 4e659069-3cd8-11e5-9a49-001c4270714e:1-48Executed_Gtid_Set: 4e659069-3cd8-11e5-9a49-001c4270714e:1-48Auto_Position: 1

在此成功跳过了错误，同步继续。可以通过这个办法来处理复制失败的问题，这里还有个例子，有兴趣的可以看一下（从服务器中跳过一条语句/事务）：

mysql > stop slave;
Query OK, 0 ROWS affected (0.05 sec)
mysql > CHANGE master TO MASTER_DELAY=600;
Query OK, 0 ROWS affected (0.27 sec)
mysql > START slave;
Query OK, 0 ROWS affected, 1 warning (0.06 sec)master 原本是正常的， 然后意外地执行了 truncate table:mysql > INSERT INTO t SET title='c';
Query OK, 1 ROW affected (0.03 sec)
mysql > INSERT INTO t SET title='d';
Query OK, 1 ROW affected (0.05 sec)mysql > SHOW master STATUS \G
*************************** 1. ROW ***************************File: black-bin.000001POSITION: 2817Binlog_Do_DB: Binlog_Ignore_DB: 
Executed_Gtid_Set: 0c005b76-d3c7-11e2-a27d-274c063b18c4:1-10
1 ROW IN SET (0.00 sec)mysql > TRUNCATE TABLE t;
Query OK, 0 ROWS affected (0.15 sec)mysql > SHOW master STATUS \G
*************************** 1. ROW ***************************File: black-bin.000001POSITION: 2948Binlog_Do_DB: Binlog_Ignore_DB: 
Executed_Gtid_Set: 0c005b76-d3c7-11e2-a27d-274c063b18c4:1-11
1 ROW IN SET (0.00 sec)slave有延迟， 虽然已经获取到了gtid及对应的events, 但是并未执行：mysql > SHOW slave STATUS \G
*************************** 1. ROW ***************************Slave_IO_State: Waiting FOR master TO send event
.......
.......SQL_Delay: 600SQL_Remaining_Delay: 565Slave_SQL_Running_State: Waiting until MASTER_DELAY seconds after master executed eventMaster_Retry_Count: 86400Master_Bind: Last_IO_Error_Timestamp: Last_SQL_Error_Timestamp: Master_SSL_Crl: Master_SSL_Crlpath: Retrieved_Gtid_Set: 0c005b76-d3c7-11e2-a27d-274c063b18c4:9-11Executed_Gtid_Set: 0c005b76-d3c7-11e2-a27d-274c063b18c4:1-8Auto_Position: 1
1 ROW IN SET (0.00 sec)要想办法在slave中跳过 GTID：0c005b76-d3c7-11e2-a27d-274c063b18c4:11， 也就是那条truncate table语句 。
办法就是设置GTID_NEXT，然后提交一个空的事务。mysql > stop slave;
Query OK, 0 ROWS affected (0.03 sec)
mysql > SET session gtid_next='0c005b76-d3c7-11e2-a27d-274c063b18c4:11';
Query OK, 0 ROWS affected (0.00 sec)
mysql > BEGIN; commit;
Query OK, 0 ROWS affected (0.00 sec)
Query OK, 0 ROWS affected (0.01 sec)mysql >SET SESSION GTID_NEXT = AUTOMATIC;
Query OK, 0 ROWS affected (0.00 sec)
mysql > START slave;
Query OK, 0 ROWS affected, 1 warning (0.07 sec)查看复制状态
mysql > SHOW slave STATUS \G
*************************** 1. ROW ***************************Slave_IO_State: Waiting FOR master TO send event
.......
.......Retrieved_Gtid_Set: 0c005b76-d3c7-11e2-a27d-274c063b18c4:9-11Executed_Gtid_Set: 0c005b76-d3c7-11e2-a27d-274c063b18c4:1-11Auto_Position: 1
1 ROW IN SET (0.00 sec)mysql > SELECT * FROM t;
+----+-------+
| id | title |
+----+-------+
|  1 | a     |
|  2 | b     |
|  3 | c     |
|  4 | d     |
+----+-------+
4 ROWS IN SET (0.00 sec)成功跳过 truncate table, 当然此时主从的数据已经不一致了。

注意：通过GTID的复制都是没有指定MASTER_LOG_FILE和MASTER_LOG_POS的，所以通过GTID复制都是从最先开始的事务开始，除非在自己的binlog里面有执行过之前的记录，才会继续后面的执行。

② 要是事务日志被purge，再进行change：

mysql> show master logs;   
+----------------------+-----------+
| Log_name             | File_size |
+----------------------+-----------+
| mysql-bin3306.000001 |  38260944 |
+----------------------+-----------+
1 row in set (0.00 sec)mysql> flush logs;
Query OK, 0 rows affected (0.03 sec)mysql> show tables;
+---------------+
| Tables_in_mmm |
+---------------+
| patent_family |
| t1            |
| t2            |
+---------------+
3 rows in set (0.01 sec)mysql> create table t3(id int)engine = tokudb;
Query OK, 0 rows affected (0.02 sec)mysql> insert into t3 values(3),(4);
Query OK, 2 rows affected (0.05 sec)
Records: 2  Duplicates: 0  Warnings: 0mysql> flush logs;
Query OK, 0 rows affected (0.02 sec)mysql> create table ttt(id int)engine = tokudb;
Query OK, 0 rows affected (0.02 sec)mysql> insert into ttt values(1),(2),(3),(4),(5);
Query OK, 5 rows affected (0.03 sec)
Records: 5  Duplicates: 0  Warnings: 0mysql> show master logs;
+----------------------+-----------+
| Log_name             | File_size |
+----------------------+-----------+
| mysql-bin3306.000001 |  38260995 |
| mysql-bin3306.000002 |       656 |
| mysql-bin3306.000003 |       619 |
+----------------------+-----------+
3 rows in set (0.00 sec)mysql> purge binary logs to 'mysql-bin3306.000003';  #日志被purge
Query OK, 0 rows affected (0.02 sec)mysql> show master logs;   #日志被purge之后等下的binlog 
+----------------------+-----------+
| Log_name             | File_size |
+----------------------+-----------+
| mysql-bin3306.000003 |       619 |
+----------------------+--------3308登陆之后执行：mysql> change master to master_host='127.0.0.1',master_user='rep',master_password='rep',master_port=3306,master_auto_position=1;
Query OK, 0 rows affected, 2 warnings (0.04 sec)mysql> start slave;
Query OK, 0 rows affected (0.02 sec)mysql> show slave status\G;
*************************** 1. row ***************************Slave_IO_State: Master_Host: 127.0.0.1Master_User: repMaster_Port: 3306Connect_Retry: 60Master_Log_File: Read_Master_Log_Pos: 4Relay_Log_File: mysqld-relay-bin3308.000001Relay_Log_Pos: 4Relay_Master_Log_File: Slave_IO_Running: NoSlave_SQL_Running: YesReplicate_Do_DB: Replicate_Ignore_DB: Replicate_Do_Table: Replicate_Ignore_Table: Replicate_Wild_Do_Table: Replicate_Wild_Ignore_Table: Last_Errno: 0Last_Error: Skip_Counter: 0Exec_Master_Log_Pos: 0Relay_Log_Space: 151Until_Condition: NoneUntil_Log_File: Until_Log_Pos: 0Master_SSL_Allowed: NoMaster_SSL_CA_File: Master_SSL_CA_Path: Master_SSL_Cert: Master_SSL_Cipher: Master_SSL_Key: Seconds_Behind_Master: 0
Master_SSL_Verify_Server_Cert: NoLast_IO_Errno: 1236Last_IO_Error: Got fatal error 1236 from master when reading data from binary log: 'The slave is connecting using CHANGE MASTER TO MASTER_AUTO_POSITION = 1, but the master has purged binary logs containing GTIDs that the slave requires.'Last_SQL_Errno: 0Last_SQL_Error: Replicate_Ignore_Server_Ids: Master_Server_Id: 1Master_UUID: 4e659069-3cd8-11e5-9a49-001c4270714eMaster_Info_File: /var/lib/mysql3/master.infoSQL_Delay: 0SQL_Remaining_Delay: NULLSlave_SQL_Running_State: Slave has read all relay log; waiting for the slave I/O thread to update itMaster_Retry_Count: 86400Master_Bind: Last_IO_Error_Timestamp: 150811 00:02:50Last_SQL_Error_Timestamp: Master_SSL_Crl: Master_SSL_Crlpath: Retrieved_Gtid_Set: Executed_Gtid_Set: Auto_Position: 1

报错：

                Last_IO_Errno: 1236Last_IO_Error: Got fatal error 1236 from master when reading data from binary log: 'The slave is connecting using CHANGE MASTER TO MASTER_AUTO_POSITION = 1, but the master has purged binary logs containing GTIDs that the slave requires.'

这里需要解决的是：Slave如何跳过purge的部分，而不是在最先开始的事务执行。

在主上执行，查看被purge的GTID：
mysql> show global variables like 'gtid_purged';
+---------------+-------------------------------------------+
| Variable_name | Value                                     |
+---------------+-------------------------------------------+
| gtid_purged   | 4e659069-3cd8-11e5-9a49-001c4270714e:1-50 |
+---------------+-------------------------------------------+
1 row in set (0.00 sec)在从上执行，跳过这个GTID：
mysql> stop slave;
Query OK, 0 rows affected (0.00 sec)mysql> set global gtid_purged = '4e659069-3cd8-11e5-9a49-001c4270714e:1-50';
Query OK, 0 rows affected (0.02 sec)mysql> reset master;
Query OK, 0 rows affected (0.04 sec)mysql> start slave;
Query OK, 0 rows affected (0.01 sec)要是出现：
ERROR 1840 (HY000): @@GLOBAL.GTID_PURGED can only be set when @@GLOBAL.GTID_EXECUTED is empty.
则需要执行：
reset master;

到这从的同步就正常了。 

mysql> show slave status\G;
*************************** 1. row ***************************Slave_IO_State: Waiting for master to send eventMaster_Host: 127.0.0.1Master_User: repMaster_Port: 3306Connect_Retry: 60Master_Log_File: mysql-bin3306.000003Read_Master_Log_Pos: 619Relay_Log_File: mysqld-relay-bin3308.000002Relay_Log_Pos: 797Relay_Master_Log_File: mysql-bin3306.000003Slave_IO_Running: YesSlave_SQL_Running: YesReplicate_Do_DB: Replicate_Ignore_DB: Replicate_Do_Table: Replicate_Ignore_Table: Replicate_Wild_Do_Table: Replicate_Wild_Ignore_Table: Last_Errno: 0Last_Error: Skip_Counter: 0Exec_Master_Log_Pos: 619Relay_Log_Space: 1006Until_Condition: NoneUntil_Log_File: Until_Log_Pos: 0Master_SSL_Allowed: NoMaster_SSL_CA_File: Master_SSL_CA_Path: Master_SSL_Cert: Master_SSL_Cipher: Master_SSL_Key: Seconds_Behind_Master: 0
Master_SSL_Verify_Server_Cert: NoLast_IO_Errno: 0Last_IO_Error: Last_SQL_Errno: 0Last_SQL_Error: Replicate_Ignore_Server_Ids: Master_Server_Id: 1Master_UUID: 4e659069-3cd8-11e5-9a49-001c4270714eMaster_Info_File: /var/lib/mysql3/master.infoSQL_Delay: 0SQL_Remaining_Delay: NULLSlave_SQL_Running_State: Slave has read all relay log; waiting for the slave I/O thread to update itMaster_Retry_Count: 86400Master_Bind: Last_IO_Error_Timestamp: Last_SQL_Error_Timestamp: Master_SSL_Crl: Master_SSL_Crlpath: Retrieved_Gtid_Set: 4e659069-3cd8-11e5-9a49-001c4270714e:51-52Executed_Gtid_Set: 4e659069-3cd8-11e5-9a49-001c4270714e:1-52Auto_Position: 1
1 row in set (0.00 sec)mysql> use mmm
Database changed
mysql> show tables;
+---------------+
| Tables_in_mmm |
+---------------+
| ttt           |
+---------------+
1 row in set (0.00 sec)

③ 通过另一个从库恢复从库数据

比如一台从库误操作，数据丢失了，可以通过另一个从库来进行恢复：

slave2（3308）：
mysql> use mmm
Database changed
mysql> show tables;
+---------------+
| Tables_in_mmm |
+---------------+
| patent_family |
| t             |
| tt            |
+---------------+
3 rows in set (0.00 sec)mysql> truncate table tt;  #误操作，把记录删除了
Query OK, 0 rows affected (0.02 sec)mysql> show slave status\G;
*************************** 1. row ***************************Slave_IO_State: Waiting for master to send eventMaster_Host: 127.0.0.1Master_User: repMaster_Port: 3306Connect_Retry: 60Master_Log_File: mysql-bin3306.000001Read_Master_Log_Pos: 38260553Relay_Log_File: mysqld-relay-bin3308.000002Relay_Log_Pos: 38260771Relay_Master_Log_File: mysql-bin3306.000001Slave_IO_Running: YesSlave_SQL_Running: YesReplicate_Do_DB: Replicate_Ignore_DB: Replicate_Do_Table: Replicate_Ignore_Table: Replicate_Wild_Do_Table: Replicate_Wild_Ignore_Table: Last_Errno: 0Last_Error: Skip_Counter: 0Exec_Master_Log_Pos: 38260553Relay_Log_Space: 38260980Until_Condition: NoneUntil_Log_File: Until_Log_Pos: 0Master_SSL_Allowed: NoMaster_SSL_CA_File: Master_SSL_CA_Path: Master_SSL_Cert: Master_SSL_Cipher: Master_SSL_Key: Seconds_Behind_Master: 0
Master_SSL_Verify_Server_Cert: NoLast_IO_Errno: 0Last_IO_Error: Last_SQL_Errno: 0Last_SQL_Error: Replicate_Ignore_Server_Ids: Master_Server_Id: 1Master_UUID: 4e659069-3cd8-11e5-9a49-001c4270714eMaster_Info_File: /var/lib/mysql3/master.infoSQL_Delay: 0  #延迟同步SQL_Remaining_Delay: NULLSlave_SQL_Running_State: Slave has read all relay log; waiting for the slave I/O thread to update itMaster_Retry_Count: 86400Master_Bind: Last_IO_Error_Timestamp: Last_SQL_Error_Timestamp: Master_SSL_Crl: Master_SSL_Crlpath: Retrieved_Gtid_Set: 4e659069-3cd8-11e5-9a49-001c4270714e:1-46Executed_Gtid_Set: 081ccacf-3ce4-11e5-9a95-001c4270714e:1,  #多出了一个GTID(本身实例执行的事务)
4e659069-3cd8-11e5-9a49-001c4270714e:1-46Auto_Position: 1数据被误删除之后，最好停止复制：stop slave;恢复数据从slave1（3307）上备份数据，并还原到slave2（3308）中。
备份：
mysqldump  -uzjy -p123456 -h127.0.0.1 -P3307 --default-character-set=utf8 --set-gtid-purged=ON -B mmm > mmm1.sql在还原到slave2的时候需要在slave2上执行：reset master; 不然会报错：
ERROR 1840 (HY000) at line 24: @@GLOBAL.GTID_PURGED can only be set when @@GLOBAL.GTID_EXECUTED is empty.还原：
root@zjy:~# mysql -uzjy -p123456 -h127.0.0.1 -P3308 --default-character-set=utf8 < mmm.sql 开启同步：
mysql> start slave;
Query OK, 0 rows affected, 1 warning (0.03 sec)这时候你会发现误删除的数据已经被还原，并且复制也正常。因为根据GTID的原理，通过slave1的备份直接可以和Master进行同步。

这里备份注意的一点是：在备份开启GTID的实例里，需要指定 --set-gtid-purged参数，否则会报warning：

Warning: A partial dump from a server that has GTIDs will by default include the GTIDs of all transactions, even those that changed suppressed parts of the database. If you don't want to restore GTIDs, pass --set-gtid-purged=OFF. To make a complete dump, pass --all-databases --triggers --routines --events

备份文件里面会出现：

SET @@GLOBAL.GTID_PURGED='4e659069-3cd8-11e5-9a49-001c4270714e:1-483';

还原的时候会要求先在实例上reset master，不然会报错：

Warning: Using a password on the command line interface can be insecure.
ERROR 1840 (HY000) at line 24: @@GLOBAL.GTID_PURGED can only be set when @@GLOBAL.GTID_EXECUTED is empty.

指定--set-gtid-purged=ON参数，出现GTID_PURGED，直接还原的时候执行，从库不需要其他操作就可以直接change到主。关于GTID更多的信息可以到官方文档里查看。 

总结：

      GTID就是全局事务ID(global transaction identifier )，最初由google实现，官方MySQL在5.6才加入该功能。要是主从结构只有一台Master和一台Slave对于GTID来说就没有优势了，而对于2台主以上的结构优势异常明显，可以在数据不丢失的情况下切换新主。
      使用GTID需要注意的是：在构建主从复制之前，在一台将成为主的实例上进行一些操作（如数据清理等），通过GTID复制，这些在主从成立之前的操作也会被复制到从服务器上，引起复制失败。即：通过GTID复制都是从最先开始的事务日志开始，即使这些操作在复制之前执行。比如在server1上执行一些drop、delete的清理操作，接着在server2上执行change的操作，会使得server2也进行server1的清理操作。

 

参考文章：

http://dev.mysql.com/doc/refman/5.6/en/replication-gtids.html

http://www.cnblogs.com/cenalulu/p/4309009.html

http://mysqllover.com/?p=594

http://forum.z27315.com/topic/22781-mysql56%E4%B8%BB%E4%BB%8E%E5%A4%8D%E5%88%B6/

www.zhaokunyao.com/archives/4131