3

Specifically I have a set of InnoDB tables and we are backing these up into MyISAM tables to provide full text search. The slave DB routinely falls behind the Master. I am looking for advice on diagnostics or commands that can be used to keep track of the lag. Both machines are on the same network. Many thanks.

IanMulvany
  • 133
  • 4

2 Answers2

8

SHOW SLAVE STATUS is your command

mysql> SHOW SLAVE STATUS\G
*************************** 1. row ***************************
               Slave_IO_State: Waiting for master to send event
                  Master_Host: localhost
                  Master_User: root
                  Master_Port: 3306
                Connect_Retry: 3
              Master_Log_File: gbichot-bin.005
          Read_Master_Log_Pos: 79
               Relay_Log_File: gbichot-relay-bin.005
                Relay_Log_Pos: 548
        Relay_Master_Log_File: gbichot-bin.005
             Slave_IO_Running: Yes
            Slave_SQL_Running: Yes
              Replicate_Do_DB:
          Replicate_Ignore_DB:
                   Last_Errno: 0
                   Last_Error:
                 Skip_Counter: 0
          Exec_Master_Log_Pos: 79
              Relay_Log_Space: 552
              Until_Condition: None
               Until_Log_File:
                Until_Log_Pos: 0
           Master_SSL_Allowed: No
           Master_SSL_CA_File:
           Master_SSL_CA_Path:
              Master_SSL_Cert:
            Master_SSL_Cipher:
               Master_SSL_Key:
        Seconds_Behind_Master: 8
Master_SSL_Verify_Server_Cert: No
                Last_IO_Errno: 0
                Last_IO_Error:
               Last_SQL_Errno: 0
               Last_SQL_Error:
  Replicate_Ignore_Server_Ids: 0

The relevant parts are:

               Slave_IO_State: Waiting for master to send event

"Waiting for master..." is what you usually want to see

              Master_Log_File: gbichot-bin.005
          Read_Master_Log_Pos: 79
          Exec_Master_Log_Pos: 79

This, combined with "SHOW MASTER STATUS" on the master server lets you compare how much has been transferred and properly replicated

             Slave_IO_Running: Yes
            Slave_SQL_Running: Yes

Is replication running?

        Seconds_Behind_Master: 8

Lag estimation

                Last_IO_Errno: 0
                Last_IO_Error:
               Last_SQL_Errno: 0
               Last_SQL_Error:

If there's an error, this show what's wrong. Also check your error log and SHOW PROCESSLIST for more info

Jorge Bernal
  • 454
  • 1
  • 3
  • 9
  • What actually is the difference between Slave_IO_Running and Slave_SQL_Running? I always check for Yes and Yes, but what actually do imply? – Coops Jul 31 '09 at 15:39
  • 1
    Replication in MySQL works in two threads. The IO thread "brings" data from the master, and the SQL thread applies it on the slave. – Jorge Bernal Aug 01 '09 at 17:35
2

nagios's plugin does that

-bash-3.2# ~nagios/libexec/check_mysql -h | grep check-slave
 -S, --check-slave
-bash-3.2#
alexus
  • 13,112
  • 32
  • 117
  • 174
  • 1
    Good grief... Just yesterday I wrote a Nagios plugin from scratch to pull the slave info. Thanks! – Insyte Jul 31 '09 at 16:30