MGR deployment
Scene Description:
Use three servers to build a simple MGR cluster using MySQL 5.7.24 version, the server list is:
192.168.1.147 192.168.1.148 192.168.1.149
1, using the common configuration file to start the MySQL service, MGR plug-in installation (performing all nodes)
INSTALL PLUGIN group_replication SONAME 'group_replication.so';
2, MySQL adjust each node configuration file and restart the MySQL service (all nodes perform)
## config master server-id = 17218228149 log-bin = mysql-bin master_info_repository = TABLE binlog_format = ROW expire_logs_days = 7 sync_binlog = 1 gtid_mode = on enforce-gtid-consistency = true binlog_rows_query_log_events = on
binlog_checksum = NONE ## config slave skip-slave-start slave-parallel-workers = 8 slave-parallel-type = LOGICAL_CLOCK slave_preserve_commit_order = 1 log_slave_updates = 1 report_host = 192.168.1.149 ## config relay log relay-log = relay-log relay_log_recovery = ON sync_relay_log = 0 relay_log_info_repository = TABLE ## config group replication transaction_write_set_extraction = XXHASH64 loose-group_replication_group_name = "aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaa01" loose-group_replication_start_on_boot = OFF loose-group_replication_bootstrap_group = OFF loose-group_replication_local_address = "192.168.1.149:33581" loose-group_replication_group_seeds = "192.168.1.147:33581,192.168.1.148:33581,192.168.1.149:33581" loose-group_replication_ip_whitelist = "192.168.1.147,192.168.1.148,192.168.1.149"
MGR plug-in is not installed before adding MGR configuration will get an error in the configuration file.
MGR configuration instructions: above the yellow part of many parameters not mentioned in the configuration file, but is likely to lead to deployment failure, especially loose-group_replication_ip_whitelist parameters.
3, configuration MGR communication account, and clean up the MASTER (all nodes perform)
CREATE USER repl@'%' IDENTIFIED BY 'repl'; GRANT REPLICATION SLAVE ON *.* TO repl@'%'; RESET MASTER;
4, create MGR dependent replication environment (perform all nodes)
CHANGE MASTER TO MASTER_USER='repl', MASTER_PASSWORD='repl' FOR CHANNEL 'group_replication_recovery';
5, on the master node starts MGR (executed on node 192.168.1.147)
SET GLOBAL group_replication_bootstrap_group=ON; START group_replication; SET GLOBAL group_replication_bootstrap_group=off;
6, MGR started on the secondary node (node 192.168.1.148 and executed on a node 192.168.1.149)
START group_replication;
7, after the building is complete, you can use the following statement to view the status:
View status of each node ## the SELECT * the FROM performance_schema.replication_group_members; ## MGR see the current mode (single-master or multi-master) the SELECT @@ group_replication_single_primary_mode;