Description:
Hi,
when running CHANGE MASTER TO MASTER_HOST='a new host', MySQL adds a line in the error logs with the previous state of replication. Here is an example of such line.
2018-07-23T14:47:37.674972Z 6 [Note] 'CHANGE MASTER TO FOR CHANNEL '' executed'. Previous state master_host='127.0.0.1', master_port= 16745, master_log_file='mysql-bin.000001', master_log_pos= 26783, master_bind=''. New state master_host='aa', master_port= 16745, master_log_file='', master_log_pos= 4, master_bind=''.
However, if the SQL Thread was stopped before the IO Thread, the log above is wrong. I would expect the log to have data about the SQL Thread position, but it has data about the IO Thread position.
Many thanks for looking into that,
JFG
How to repeat:
slave1 [localhost] {msandbox} ((none)) > SELECT @@version;
+------------+
| @@version |
+------------+
| 5.7.22-log |
+------------+
1 row in set (0.00 sec)
slave1 [localhost] {msandbox} ((none)) > STOP SLAVE SQL_THREAD;
Query OK, 0 rows affected (0.00 sec)
slave1 [localhost] {msandbox} ((none)) > STOP SLAVE; SHOW SLAVE STATUS\G
Query OK, 0 rows affected (0.00 sec)
*************************** 1. row ***************************
Slave_IO_State:
Master_Host: 127.0.0.1
Master_User: rsandbox
Master_Port: 16745
Connect_Retry: 60
Master_Log_File: mysql-bin.000001
Read_Master_Log_Pos: 26783
Relay_Log_File: mysql-relay.000003
Relay_Log_Pos: 2678
Relay_Master_Log_File: mysql-bin.000001
Slave_IO_Running: No
Slave_SQL_Running: No
[...]
Exec_Master_Log_Pos: 19447
Relay_Log_Space: 27365
Until_Condition: None
[...]
1 row in set (0.00 sec)
slave1 [localhost] {msandbox} ((none)) > CHANGE MASTER TO MASTER_HOST = 'aa';
Query OK, 0 rows affected (0.01 sec)
When running above commands, I have the following in the logs:
2018-07-23T14:47:01.450764Z 8 [Note] Error reading relay log event for channel '': slave SQL thread was killed
2018-07-23T14:47:29.904618Z 7 [Note] Slave I/O thread killed while reading event for channel ''
2018-07-23T14:47:29.904655Z 7 [Note] Slave I/O thread exiting for channel '', read up to log 'mysql-bin.000001', position 26783
2018-07-23T14:47:37.674972Z 6 [Note] 'CHANGE MASTER TO FOR CHANNEL '' executed'. Previous state master_host='127.0.0.1', master_port= 16745, master_log_file='mysql-bin.000001', master_log_pos= 26783, master_bind=''. New state master_host='aa', master_port= 16745, master_log_file='', master_log_pos= 4, master_bind=''.
But the position of the SQL Thread is _not_ "master_log_file='mysql-bin.000001', master_log_pos= 26783", but it is (as shown by the SHOW SLAVE STATUS\G):
Relay_Master_Log_File: mysql-bin.000001
Exec_Master_Log_Pos: 19447
So I would expect "master_log_file='mysql-bin.000001', master_log_pos= 19447" in the logs.
Description: Hi, when running CHANGE MASTER TO MASTER_HOST='a new host', MySQL adds a line in the error logs with the previous state of replication. Here is an example of such line. 2018-07-23T14:47:37.674972Z 6 [Note] 'CHANGE MASTER TO FOR CHANNEL '' executed'. Previous state master_host='127.0.0.1', master_port= 16745, master_log_file='mysql-bin.000001', master_log_pos= 26783, master_bind=''. New state master_host='aa', master_port= 16745, master_log_file='', master_log_pos= 4, master_bind=''. However, if the SQL Thread was stopped before the IO Thread, the log above is wrong. I would expect the log to have data about the SQL Thread position, but it has data about the IO Thread position. Many thanks for looking into that, JFG How to repeat: slave1 [localhost] {msandbox} ((none)) > SELECT @@version; +------------+ | @@version | +------------+ | 5.7.22-log | +------------+ 1 row in set (0.00 sec) slave1 [localhost] {msandbox} ((none)) > STOP SLAVE SQL_THREAD; Query OK, 0 rows affected (0.00 sec) slave1 [localhost] {msandbox} ((none)) > STOP SLAVE; SHOW SLAVE STATUS\G Query OK, 0 rows affected (0.00 sec) *************************** 1. row *************************** Slave_IO_State: Master_Host: 127.0.0.1 Master_User: rsandbox Master_Port: 16745 Connect_Retry: 60 Master_Log_File: mysql-bin.000001 Read_Master_Log_Pos: 26783 Relay_Log_File: mysql-relay.000003 Relay_Log_Pos: 2678 Relay_Master_Log_File: mysql-bin.000001 Slave_IO_Running: No Slave_SQL_Running: No [...] Exec_Master_Log_Pos: 19447 Relay_Log_Space: 27365 Until_Condition: None [...] 1 row in set (0.00 sec) slave1 [localhost] {msandbox} ((none)) > CHANGE MASTER TO MASTER_HOST = 'aa'; Query OK, 0 rows affected (0.01 sec) When running above commands, I have the following in the logs: 2018-07-23T14:47:01.450764Z 8 [Note] Error reading relay log event for channel '': slave SQL thread was killed 2018-07-23T14:47:29.904618Z 7 [Note] Slave I/O thread killed while reading event for channel '' 2018-07-23T14:47:29.904655Z 7 [Note] Slave I/O thread exiting for channel '', read up to log 'mysql-bin.000001', position 26783 2018-07-23T14:47:37.674972Z 6 [Note] 'CHANGE MASTER TO FOR CHANNEL '' executed'. Previous state master_host='127.0.0.1', master_port= 16745, master_log_file='mysql-bin.000001', master_log_pos= 26783, master_bind=''. New state master_host='aa', master_port= 16745, master_log_file='', master_log_pos= 4, master_bind=''. But the position of the SQL Thread is _not_ "master_log_file='mysql-bin.000001', master_log_pos= 26783", but it is (as shown by the SHOW SLAVE STATUS\G): Relay_Master_Log_File: mysql-bin.000001 Exec_Master_Log_Pos: 19447 So I would expect "master_log_file='mysql-bin.000001', master_log_pos= 19447" in the logs.