MySQL Group Replication的RECOVERING状态深度理解

接收到报警通知，db3这台组复制成员故障down机了，等修复好，启动服务器，然后再启动mysql实例，进去查看状态，变成了RECOVERING，如下所示；

mysql> SELECT * FROM performance_schema.replication_group_members;
+---------------------------+--------------------------------------+----------------------+-------------+--------------+
| CHANNEL_NAME              | MEMBER_ID                            | MEMBER_HOST          | MEMBER_PORT | MEMBER_STATE |
+---------------------------+--------------------------------------+----------------------+-------------+--------------+
| group_replication_applier | 3381d155-d7d1-11e6-94f7-b8ca3af6e36c | hch_test_dbm2_121_71 |        3317 | ONLINE       |
| group_replication_applier | 664b9ce9-d7de-11e6-9e8c-18a99b763071 | hch_test_web_1_24    |        3317 | ONLINE       |
| group_replication_applier | 84dba8ff-d7d2-11e6-aa9a-18a99b76310d | bpe_service          |        3317 | RECOVERING   |
+---------------------------+--------------------------------------+----------------------+-------------+--------------+
3 rows in set (0.00 sec)


mysql> 

这种状态有点类似mongodb的sendary的RECOVERING状态，对于刚刚出现的MySQL Group Replication新技术，遇到这种情况，这种情况怎么办？

---

在mongodb的这种模式下，一般是secondary从primary库上不停复制数据，所以我们遵循这种思路，去看下db2的数据目录，看到有很多relay-bin的recovery日志在应用执行，如下所示relay日志包记录

-rw-r----- 1 mysql mysql  383 1月  17 16:08 bpe_service-relay-bin-group_replication_applier.000019
-rw-r----- 1 mysql mysql  502 1月  17 16:11 bpe_service-relay-bin-group_replication_applier.000020
-rw-r----- 1 mysql mysql  502 1月  17 16:23 bpe_service-relay-bin-group_replication_applier.000021
-rw-r----- 1 mysql mysql  421 1月  17 16:23 bpe_service-relay-bin-group_replication_applier.000022
-rw-r----- 1 mysql mysql  228 1月  17 16:23 bpe_service-relay-bin-group_replication_applier.index
-rw-r----- 1 mysql mysql  312 1月  17 16:24 bpe_service-relay-bin-group_replication_recovery.000007
-rw-r----- 1 mysql mysql 1.1G 1月  17 16:24 bpe_service-relay-bin-group_replication_recovery.000008
-rw-r----- 1 mysql mysql  391 1月  17 16:24 bpe_service-relay-bin-group_replication_recovery.000009
-rw-r----- 1 mysql mysql  312 1月  17 16:24 bpe_service-relay-bin-group_replication_recovery.000010
-rw-r----- 1 mysql mysql 1.1G 1月  17 16:24 bpe_service-relay-bin-group_replication_recovery.000011
-rw-r----- 1 mysql mysql  391 1月  17 16:24 bpe_service-relay-bin-group_replication_recovery.000012
-rw-r----- 1 mysql mysql  312 1月  17 16:24 bpe_service-relay-bin-group_replication_recovery.000013
-rw-r----- 1 mysql mysql 1.1G 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000014
-rw-r----- 1 mysql mysql  391 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000015
-rw-r----- 1 mysql mysql  312 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000016
-rw-r----- 1 mysql mysql 1.1G 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000017
-rw-r----- 1 mysql mysql  391 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000018
-rw-r----- 1 mysql mysql  312 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000019
-rw-r----- 1 mysql mysql 1.1G 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000020
-rw-r----- 1 mysql mysql  391 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000021
-rw-r----- 1 mysql mysql  312 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000022
-rw-r----- 1 mysql mysql 1.1G 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000023
-rw-r----- 1 mysql mysql  391 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000024
-rw-r----- 1 mysql mysql  312 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.000025
-rw-r----- 1 mysql mysql 181M 1月  17 16:26 bpe_service-relay-bin-group_replication_recovery.000026
-rw-r----- 1 mysql mysql 1.2K 1月  17 16:25 bpe_service-relay-bin-group_replication_recovery.index
drwxr-x--- 2 mysql mysql 4.0K 1月  17 14:11 business_db

---

然后在mysql窗口，可以看到几个线程，其中有一个就是Reading event from the relay log线程，就是表示正在应用relay日志：

mysql> show full processlist;
+----+-------------+-----------+------+---------+------+--------------------------------------------------------+-----------------------+
| Id | User        | Host      | db   | Command | Time | State                                                  | Info                  |
+----+-------------+-----------+------+---------+------+--------------------------------------------------------+-----------------------+
| 28 | system user |           | NULL | Connect |  701 | Suspending                                             | NULL                  |
| 31 | system user |           | NULL | Connect |  701 | Slave has read all relay log; waiting for more updates | NULL                  |
| 35 | system user |           | NULL | Connect |  699 | Waiting for master to send event                       | NULL                  |
| 36 | system user |           | NULL | Connect | 4519 | Reading event from the relay log                       | NULL                  |
| 38 | root        | localhost | NULL | Query   |    0 | starting                                               | show full processlist |
+----+-------------+-----------+------+---------+------+--------------------------------------------------------+-----------------------+
5 rows in set (0.00 sec)


mysql> 

---

再过一会儿，数据同步完成之后，bpe_service就会从RECOVERING变成ONLINE状态。这个更加证实了，数据已经同步完成。

mysql> SELECT * FROM performance_schema.replication_group_members;
+---------------------------+--------------------------------------+----------------------+-------------+--------------+
| CHANNEL_NAME              | MEMBER_ID                            | MEMBER_HOST          | MEMBER_PORT | MEMBER_STATE |
+---------------------------+--------------------------------------+----------------------+-------------+--------------+
| group_replication_applier | 3381d155-d7d1-11e6-94f7-b8ca3af6e36c | hch_test_dbm2_121_71 |        3317 | ONLINE       |
| group_replication_applier | 664b9ce9-d7de-11e6-9e8c-18a99b763071 | hch_test_web_1_24    |        3317 | ONLINE       |
| group_replication_applier | 84dba8ff-d7d2-11e6-aa9a-18a99b76310d | bpe_service          |        3317 | ONLINE       |
+---------------------------+--------------------------------------+----------------------+-------------+--------------+
3 rows in set (0.00 sec)


mysql>

---

去查看后台error log日志显示，刚开始启动，在做**CHANGE MASTER TO FOR CHANNEL ‘group_replication_applier’ executed’**的时候，没有取到值，然后看后面同步完后，就会执行加入组复制成员的命令

“CHANGE MASTER TO FOR CHANNEL 'group_replication_recovery' executed'. Previous state master_host='hch_test_web_1_24', master_port= 3317, master_log_file='', master_log_pos= 4, master_bind=''. New state master_host='', master_port= 0, master_log_file='', master_log_pos= 4, master_bind=''”，

在执行结束后，最后告诉我们“‘This server was declared online within the replication group’”已经加入到了组成员，而且变成了实时的ONLINE状态：

2017-01-17T08:23:15.710146Z 4 [Note] Plugin group_replication reported: 'auto_increment_increment is reset to 1'
2017-01-17T08:23:15.710200Z 4 [Note] Plugin group_replication reported: 'auto_increment_offset is reset to 1'
2017-01-17T08:23:15.710401Z 20 [Note] Error reading relay log event for channel 'group_replication_applier': slave SQL thread was killed
2017-01-17T08:23:15.711212Z 17 [Note] Plugin group_replication reported: 'The group replication applier thread was killed'
2017-01-17T08:23:42.141239Z 4 [Note] Plugin group_replication reported: 'Group communication SSL configuration: group_replication_ssl_mode: "DISABLED"'
2017-01-17T08:23:42.141451Z 4 [Note] Plugin group_replication reported: '[GCS] Added automatically IP ranges 127.0.0.1/8,192.168.121.111/23 to the whitelist'
2017-01-17T08:23:42.141990Z 4 [Note] Plugin group_replication reported: '[GCS] Translated 'db2' to 192.168.121.111'
2017-01-17T08:23:42.142174Z 4 [Note] Plugin group_replication reported: '[GCS] SSL was not enabled'
2017-01-17T08:23:42.142220Z 4 [Note] Plugin group_replication reported: 'Initialized group communication with configuration: group_replication_group_name: "e4668cea-d7ca-11e6-86b5-18a99b76310d"; group_replication_local_address: "db2:24902"; group_replication_group_seeds: "db1:24901,db2:24902,db3:24903"; group_replication_bootstrap_group: false; group_replication_poll_spin_loops: 0; group_replication_compression_threshold: 1000000; group_replication_ip_whitelist: "AUTOMATIC"'
2017-01-17T08:23:42.142936Z 28 [Note] 'CHANGE MASTER TO FOR CHANNEL 'group_replication_applier' executed'. Previous state master_host='', master_port= 0, master_log_file='', master_log_pos= 4, master_bind=''. New state master_host='', master_port= 0, master_log_file='', master_log_pos= 4, master_bind=''.

-- #（1）尝试启动默认的组复制，结果信息不对称，启动失败，然后开始初始化Group Replication
2017-01-17T08:23:42.166956Z 31 [Note] Slave SQL thread for channel 'group_replication_applier' initialized, starting replication in log 'FIRST' at position 0, relay log './bpe_service-relay-bin-group_replication_applier.000019' position: 4
2017-01-17T08:23:42.166960Z 4 [Note] Plugin group_replication reported: 'Group Replication applier module successfully initialized!'
2017-01-17T08:23:42.167064Z 4 [Note] Plugin group_replication reported: 'auto_increment_increment is set to 7'
2017-01-17T08:23:42.167079Z 4 [Note] Plugin group_replication reported: 'auto_increment_offset is set to 12002'
2017-01-17T08:23:42.167219Z 0 [Note] Plugin group_replication reported: 'state 4257 action xa_init'
2017-01-17T08:23:42.167292Z 0 [Note] Plugin group_replication reported: 'Successfully bound to 0.0.0.0:24902 (socket=49).'
2017-01-17T08:23:42.167341Z 0 [Note] Plugin group_replication reported: 'Successfully set listen backlog to 32 (socket=49)!'
2017-01-17T08:23:42.167357Z 0 [Note] Plugin group_replication reported: 'Successfully unblocked socket (socket=49)!'
2017-01-17T08:23:42.167400Z 0 [Note] Plugin group_replication reported: 'connecting to db2 24902'
2017-01-17T08:23:42.167585Z 0 [Note] Plugin group_replication reported: 'Ready to accept incoming connections on 0.0.0.0:24902 (socket=49)!'
2017-01-17T08:23:42.174427Z 0 [Note] Plugin group_replication reported: 'client connected to db2 24902 fd 50'
2017-01-17T08:23:42.174807Z 0 [Note] Plugin group_replication reported: 'connecting to db2 24902'
2017-01-17T08:23:42.175008Z 0 [Note] Plugin group_replication reported: 'client connected to db2 24902 fd 63'
2017-01-17T08:23:42.175234Z 0 [Note] Plugin group_replication reported: 'connecting to db2 24902'
2017-01-17T08:23:42.175411Z 0 [Note] Plugin group_replication reported: 'client connected to db2 24902 fd 65'
2017-01-17T08:23:42.175614Z 0 [Note] Plugin group_replication reported: 'connecting to db2 24902'
2017-01-17T08:23:42.175799Z 0 [Note] Plugin group_replication reported: 'client connected to db2 24902 fd 67'
2017-01-17T08:23:42.176006Z 0 [Note] Plugin group_replication reported: 'connecting to db2 24902'
2017-01-17T08:23:42.176176Z 0 [Note] Plugin group_replication reported: 'client connected to db2 24902 fd 69'
2017-01-17T08:23:42.176377Z 0 [Note] Plugin group_replication reported: 'connecting to db2 24902'
2017-01-17T08:23:42.176563Z 0 [Note] Plugin group_replication reported: 'client connected to db2 24902 fd 71'
2017-01-17T08:23:42.176876Z 0 [Note] Plugin group_replication reported: 'connecting to db1 24901'
2017-01-17T08:23:42.177312Z 0 [Note] Plugin group_replication reported: 'client connected to db1 24901 fd 73'
2017-01-17T08:23:43.463474Z 0 [Note] Plugin group_replication reported: 'state 4257 action xa_snapshot'
2017-01-17T08:23:43.463969Z 0 [Note] Plugin group_replication reported: 'new state x_recover'
2017-01-17T08:23:43.464000Z 0 [Note] Plugin group_replication reported: 'state 4277 action xa_complete'
2017-01-17T08:23:43.464275Z 0 [Note] Plugin group_replication reported: 'new state x_run'
2017-01-17T08:23:44.619112Z 0 [Note] Plugin group_replication reported: 'Starting group replication recovery with view_id 14844510167669342:17' 

--#（2） 开始进行组内数据recovery
2017-01-17T08:23:44.619451Z 34 [Note] Plugin group_replication reported: 'Establishing group recovery connection with a possible donor. Attempt 1/10'
2017-01-17T08:23:44.624181Z 34 [Note] 'CHANGE MASTER TO FOR CHANNEL 'group_replication_recovery' executed'. Previous state master_host='', master_port= 0, master_log_file='', master_log_pos= 4, master_bind=''. New state master_host='hch_test_web_1_24', master_port= 3317, master_log_file='', master_log_pos= 4, master_bind=''.
2017-01-17T08:23:44.628853Z 34 [Note] Plugin group_replication reported: 'Establishing connection to a group replication recovery donor 664b9ce9-d7de-11e6-9e8c-18a99b763071 at hch_test_web_1_24 port: 3317.'
2017-01-17T08:23:44.629156Z 35 [Warning] Storing MySQL user name or password information in the master info repository is not secure and is therefore not recommended. Please consider using the USER and PASSWORD connection options for START SLAVE; see the 'START SLAVE Syntax' in the MySQL Manual for more information.
2017-01-17T08:23:44.635870Z 36 [Note] Slave SQL thread for channel 'group_replication_recovery' initialized, starting replication in log 'FIRST' at position 0, relay log './bpe_service-relay-bin-group_replication_recovery.000001' position: 4
2017-01-17T08:23:44.681897Z 35 [Note] Slave I/O thread for channel 'group_replication_recovery': connected to master 'repl@hch_test_web_1_24:3317',replication started in log 'FIRST' at position 4
2017-01-17T08:43:34.219952Z 34 [Note] Plugin group_replication reported: 'Terminating existing group replication donor connection and purging the corresponding logs.'
2017-01-17T08:43:34.220609Z 36 [Note] Slave SQL thread for channel 'group_replication_recovery' exiting, replication stopped in log 'binlog.000042' at position 185511578
2017-01-17T08:43:34.226869Z 35 [Note] Slave I/O thread killed while reading event for channel 'group_replication_recovery'
2017-01-17T08:43:34.226886Z 35 [Note] Slave I/O thread exiting for channel 'group_replication_recovery', read up to log 'binlog.000042', position 185511578
2017-01-17T08:43:34.269973Z 34 [Note] 'CHANGE MASTER TO FOR CHANNEL 'group_replication_recovery' executed'. Previous state master_host='hch_test_web_1_24', master_port= 3317, master_log_file='', master_log_pos= 4, master_bind=''. New state master_host='', master_port= 0, master_log_file='', master_log_pos= 4, master_bind=''.

-- # （3）已经数据恢复完成，加入组，成为组成员
2017-01-17T08:43:34.290920Z 0 [Note] Plugin group_replication reported: 'This server was declared online within the replication group'

小结：mysql group replication的使用来说，有些理念比如mongodb的分片以及primary-secondary的思路理念可以参考下，这样理解起来就会比较容易一些。