Verifying HA

Simulate a scenario where a secondary node is stopped and test read requests.
1. Log in to secondary database 1.
```
/usr/local/mysql/bin/mysql -uroot -p123456 -S /data/mysql/run/mysqls.sock
```
2. Stop the database.
```
shutdown;
```
3. Log in to the primary database and view the group member information.
```
/usr/local/mysql/bin/mysql -uroot -p123456 -S /data/mysql/run/mysqlm.sock
select * from performance_schema.replication_group_members;
```
4. Execute the read request script.
```
sh /home/read_6447.sh
```
  The 127.0.0.1:3307 database service is stopped. Connections to the read port are routed only to 127.0.0.1:3308 as expected.
Simulate a scenario where two secondary nodes are stopped and test the routing of read requests.
1. Log in to secondary database 2.
```
/usr/local/mysql/bin/mysql -uroot -p123456 -S /data/mysql/run/mysqls2.sock
```
2. Stop the database.
```
shutdown;
```
3. Log in to the primary database and view the group member information.
```
/usr/local/mysql/bin/mysql -uroot -p123456 -S /data/mysql/run/mysqlm.sock
select * from performance_schema.replication_group_members;
```
4. Execute the read request script.
```
sh /home/read_6447.sh
```
  Both secondary servers are stopped. Connections to the read port are routed to the primary node (127.0.0.1:3306) as expected (implemented by using the round-robin-with-fallback algorithm).

Simulate a primary role switchover.

Start secondary databases 1 and 2.

/usr/local/mysql/bin/mysqld_safe --defaults-file=/etc/my.cnf-slave-mgr &
/usr/local/mysql/bin/mysqld_safe --defaults-file=/etc/my.cnf-slave-2-mgr &

/usr/local/mysql/bin/mysql -uroot -p123456 -S /data/mysql/run/mysqls.sock
/usr/local/mysql/bin/mysql -uroot -p123456 -S /data/mysql/run/mysqls2.sock

Execute the following statement in secondary databases 1 and 2 separately to start MGR:
```
START GROUP_REPLICATION;
```
View the information about group members. You can view any member in the MGR group.
```
select * from performance_schema.replication_group_members;
```
Configure secondary database 2 as the primary and the other two databases as the secondaries.
```
select group_replication_set_as_primary ('member_id of secondary database 2');
```
Execute the read request script.
```
sh /home/read_6446.sh
```

Execute the write request script.

Open the file.
```
vim /home/write_6446.sh
```

Replace the file content with the following:

           
                #!/bin/bash
for i in {1001..2000}
do
/usr/local/mysql/bin/mysql -uroot -p123456 -P6446 -h127.0.0.1 -Dsysbench -e "insert into test1(id,report_serverid) values($i,@@server_id)"
done

Press Esc, type :wq!, and press Enter to save the file and exit.

Execute the write request script.

           
                sh /home/write_6446.sh

Confirm the test data.

           
                /usr/local/mysql/bin/mysql -uroot -p123456 -P6446 -h127.0.0.1 -Dsysbench -e "select * from test1"

Execute the read request script.
```
sh /home/read_6447.sh
```
All read and write requests of the write port 6446 are routed to the primary node 127.0.0.1:3308, and all read requests of the read port 6447 are routed to secondary nodes 127.0.0.1:3306 and 127.0.0.1:3307, which meets the expectation.

Parent topic: Deploying MySQL Router