Comments on: MySQL SLAVE, obey your MASTER!

Question: MySQL SLAVE, obey your MASTER!

ph00dz — Tue, 29 May 2007 08:29:08 -0800

MySQL, why won't you replicate properly?

Replication doesn't work.

We set up replication on two windows machines running mysql, got everything all synced up, set up the master/slave relationship, and everything worked fine for about two days.

Then it just stops... and the slave doesn't want to catch up with the master. Basic tests seem to indicate that the connection is still live -- we can still see the master's status from the slave, but it won't just grab all the data it needs to be current.

Here are the mysql servers:
master: 4.1 (NT)
slave: 5.0 (Advanced Server 2003)

What can we do to diagnose the problem?

By: cmiller

cmiller — Tue, 29 May 2007 09:05:27 -0800

I know something of MySQL server. FWIW, replicating from a lower-version master to a higher-version slave is totally fine. That's supported.

My first question would be, what are the thread states?

How many "host_name-relay-bin.nnnnnn" files do you have?

I assume you've read these sections of the manual: Replication Implementation Details.

By: mendel

mendel — Tue, 29 May 2007 09:16:43 -0800

MySQL logs replication problems into the server log on the slave. Check that log.

Also, please share the output of SHOW MASTER STATUS on the master and SHOW SLAVE STATUS on the slave with us -- otherwise it's all crazy guessing.

Mixed-version replication is always slightly hazardous, though, in either direction.

By: ph00dz

ph00dz — Tue, 29 May 2007 09:17:14 -0800

When I run show processlist the relevant line is:

Command: Binlog Dump
Status: Has sent all binlog to slave; waiting for binlog to be updated

There are 7 of those bin files in the data dir on the slave.

By: ph00dz

ph00dz — Tue, 29 May 2007 09:22:45 -0800

Here's the master status:
mysql> show master status \G;
*************************** 1. row ***************************
File: td_server-bin.000003
Position: 960993582
Binlog_Do_DB:
Binlog_Ignore_DB:
1 row in set (0.04 sec)

Here's the slave status:
mysql> show slave status \G;
*************************** 1. row ***************************
Slave_IO_State: Waiting for master to send event
Master_Host: 192.168.2.102
Master_User: slave_user
Master_Port: 3306
Connect_Retry: 60
Master_Log_File: td_server-bin.000003
Read_Master_Log_Pos: 960986863
Relay_Log_File: tdesigns_backup-relay-bin.000002
Relay_Log_Pos: 54975987
Relay_Master_Log_File: td_server-bin.000003
Slave_IO_Running: Yes
Slave_SQL_Running: No
Replicate_Do_DB: traffic
Replicate_Ignore_DB:
Replicate_Do_Table:
Replicate_Ignore_Table:
Replicate_Wild_Do_Table:
Replicate_Wild_Ignore_Table:
Last_Errno: 1062
Last_Error: Error 'Duplicate entry '216555' for key 1' on query. Default database: 'traffic'. Query: 'INSERT INTO td_lookuplog (id, dateAdded, epid, cuid, ipAddress) VALUES (', NOW(), '1262', '2057113', '192.168.2.136')'
Skip_Counter: 0
Exec_Master_Log_Pos: 544987956
Relay_Log_Space: 475701546
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: NULL
1 row in set (0.00 sec)

... so, the question is -- does that error kill things?

By: jenkinsEar

jenkinsEar — Tue, 29 May 2007 09:26:01 -0800

looks like it. Did you try to slurp in the same binlog twice somehow?

By: ph00dz

ph00dz — Tue, 29 May 2007 09:39:07 -0800

Sweet. Deleted the offending row and everything got all synced back up.

Yay! Thanks!