1. 背景

前面文章已經介紹如何部署MySQL GR 的兩種模式:

在前面的部署裡面，因為是預先知道需要部署的組節點規模，所以我們能夠在my.cnf裡面體現出組成員資訊。但是，前面的部署並沒有體現出動態新增組節點。

這篇文章介紹如何動態地往一個組裡面新增新的成員。

前提: MySQL GR 3節點（s1、s2、s3）部署成功，模式定為Multi-Primary Mode，Single-Primary Mode也是一樣的處理。

目標: 不重啟已有GR節點的基礎上，新增新成員s4、s5，組成5節點的GR。

2. 部署過程

假設已經通過3節點的部署(部署參考上面的部落格)

2.1 線上修改已有GR節點配置

分別登陸s1、s2、s3，執行以下命令，修改GR配置:

mysql> set global group_replication_group_seeds="127.0.0.1:24901,127.0.0.1:24902,127.0.0.1:24903,127.0.0.1:24904,127.0.0.1:24905";
Query OK, 0 rows affected (0.00 sec)

mysql> show variables like '%group_replication_group_seeds%';
+-------------------------------+---------------------------------------------------------------------------------+
 

| Variable_name                 | Value                                                                           |
+-------------------------------+---------------------------------------------------------------------------------+
| group_replication_group_seeds | 127.0.0.1:24901,127.0.0.1:24902,127.0.0.1:24903,127.0.0.1:24904,127.0.0.1:24905 |
+-------------------------------+---------------------------------------------------------------------------------+
 

1 row in set (0.00 sec)

2.2 部署s4

安裝s4:

mysql-5.7/bin/mysqld --initialize-insecure --basedir=$PWD/mysql-5.7 --datadir=$PWD/data/s4

增加配置s4.cnf:

# server configuration
[mysqld]
datadir=/dba/mysql/data/s4
basedir=/dba/mysql/mysql-5.7/

port=24804
socket=/dba/mysql/data/s4/s4.sock

server_id=4
gtid_mode=ON
enforce_gtid_consistency=ON
master_info_repository=TABLE
relay_log_info_repository=TABLE
binlog_checksum=NONE
log_slave_updates=ON
log_bin=binlog
binlog_format=ROW

transaction_write_set_extraction=XXHASH64
loose-group_replication_group_name="aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa"
loose-group_replication_start_on_boot=off
loose-group_replication_local_address= "127.0.0.1:24904"
loose-group_replication_group_seeds= "127.0.0.1:24901,127.0.0.1:24902,127.0.0.1:24903,127.0.0.1:24904,127.0.0.1:24905"
loose-group_replication_bootstrap_group= off
loose-group_replication_single_primary_mode=false
loose-group_replication_enforce_update_everywhere_checks=true

啟動s4:

nohup mysql-5.7/bin/mysqld --defaults-file=data/s4/s4.cnf >data/s4/nohup.out 2>data/s4/nohup.out &

使用以下命令登陸s4:

mysql -uroot -h127.0.0.1 -P24804 --skip-password

修改root密碼:

SET SQL_LOG_BIN=0;
ALTER USER 'root'@'localhost' IDENTIFIED BY 'mysql';
SET SQL_LOG_BIN=1;

建立gr所需要使用者並授權:

SET SQL_LOG_BIN=0;
CREATE USER [email protected]'%';
GRANT REPLICATION SLAVE ON *.* TO [email protected]'%' IDENTIFIED BY 'rpl_pass';
FLUSH PRIVILEGES;
SET SQL_LOG_BIN=1;
CHANGE MASTER TO MASTER_USER='rpl_user', MASTER_PASSWORD='rpl_pass' FOR CHANNEL 'group_replication_recovery';

安裝組複製外掛並開啟外掛

INSTALL PLUGIN group_replication SONAME 'group_replication.so';

START GROUP_REPLICATION;

驗證s4已經加入GR叢集:

mysql> SELECT * FROM performance_schema.replication_group_members;
+---------------------------+--------------------------------------+-------------+-------------+--------------+
| CHANNEL_NAME              | MEMBER_ID                            | MEMBER_HOST | MEMBER_PORT | MEMBER_STATE |
+---------------------------+--------------------------------------+-------------+-------------+--------------+
| group_replication_applier | 933a5345-d321-11e6-b971-fa163ee40410 | yourhost  |       24801 | ONLINE       |
| group_replication_applier | 95155949-d325-11e6-b68e-fa163ee40410 | yourhost  |       24804 | ONLINE       |
| group_replication_applier | 9abe6249-d321-11e6-bb5c-fa163ee40410 | yourhost  |       24802 | ONLINE       |
| group_replication_applier | a1c89c34-d321-11e6-bbbc-fa163ee40410 | yourhost  |       24803 | ONLINE       |
+---------------------------+--------------------------------------+-------------+-------------+--------------+
4 rows in set (0.00 sec)

2.3 部署s5

與s4相同的部署即可以將s5加入GR叢集:

s5.cnf配置:

# server configuration
[mysqld]
datadir=/dba/mysql/data/s5
basedir=/dba/mysql/mysql-5.7/

port=24805
socket=/dba/mysql/data/s5/s5.sock

server_id=5
gtid_mode=ON
enforce_gtid_consistency=ON
master_info_repository=TABLE
relay_log_info_repository=TABLE
binlog_checksum=NONE
log_slave_updates=ON
log_bin=binlog
binlog_format=ROW

transaction_write_set_extraction=XXHASH64
loose-group_replication_group_name="aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa"
loose-group_replication_start_on_boot=off
loose-group_replication_local_address= "127.0.0.1:24905"
loose-group_replication_group_seeds= "127.0.0.1:24901,127.0.0.1:24902,127.0.0.1:24903,127.0.0.1:24904,127.0.0.1:24905"
loose-group_replication_bootstrap_group= off
loose-group_replication_single_primary_mode=false
loose-group_replication_enforce_update_everywhere_checks=true

啟動s5

nohup mysql-5.7/bin/mysqld --defaults-file=data/s5/s5.cnf >data/s5/nohup.out 2>data/s5/nohup.out &

其他後續步驟省略，依據s4的部署執行即可。

最後也成功將s5加入GR叢集:

mysql> SELECT * FROM performance_schema.replication_group_members;
+---------------------------+--------------------------------------+-------------+-------------+--------------+
| CHANNEL_NAME              | MEMBER_ID                            | MEMBER_HOST | MEMBER_PORT | MEMBER_STATE |
+---------------------------+--------------------------------------+-------------+-------------+--------------+
| group_replication_applier | 933a5345-d321-11e6-b971-fa163ee40410 | yourhost  |       24801 | ONLINE       |
| group_replication_applier | 95155949-d325-11e6-b68e-fa163ee40410 | yourhost  |       24804 | ONLINE       |
| group_replication_applier | 9abe6249-d321-11e6-bb5c-fa163ee40410 | yourhost  |       24802 | ONLINE       |
| group_replication_applier | a1c89c34-d321-11e6-bbbc-fa163ee40410 | yourhost  |       24803 | ONLINE       |
| group_replication_applier | e973e06c-d327-11e6-98bc-fa163ee40410 | yourhost  |       24805 | ONLINE       |
+---------------------------+--------------------------------------+-------------+-------------+--------------+
5 rows in set (0.01 sec)

2.4 後續操作

到這裡我們已經完成了s4和s5的動態新增，當然，為了下次s1、s2、s3重啟後gr配置仍然能夠生效，我們需要修改s1.cnf、s2.cnf和s3.cnf裡面的一個配置項為:

loose-group_replication_group_seeds= "127.0.0.1:24901,127.0.0.1:24902,127.0.0.1:24903,127.0.0.1:24904,127.0.0.1:24905"

這樣能夠確保下次重啟這些節點，GR的配置能夠是最新的配置。