Z
Z
zoxa2013-01-23 23:39:02
MySQL
zoxa, 2013-01-23 23:39:02

Problem with MySQL replication Master<->Master -> slave. What could be the reason?

I'm trying to add a Slave to the current implementation to the combat Master <-> Master.
I made a backup yesterday and raised Slave replication via SSL to Master2. Slave began to catch up with the Masters and in an hour reduced the difference from 8.5 to 7 hours. And then later partially stopped:
Ie . IO_THREAD all the time received new binlogs and position from Master2,
and SQL_THREAD, although it was launched, stopped at one binlog and position and did not want to move further.
as a result, IO_THREAD outstripped SQL_THREAD by as much as 10 binlog files
Symptoms:
1. in SHOW SLAVE STATUS -> Last_ERROR_* everything is empty
2. in MySQL logs are empty
3. in SHOW PROCESSLIST all two processes. IO_THREAD and my SHOW PROCESSLIST
Tried to stop \ start replication and IO_THREAD and SQL_THREAD separately - 0 effect. We get logs from the master, but does not want to move.
I do not understand where to move from this situation?
Tell me which way is best to dig?

Answer the question

In order to leave comments, you need to log in

1 answer(s)
E
EugeneOZ, 2013-01-24
@EugeneOZ

You can look here: www.codership.com/content/using-galera-cluster

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question