MySQL Lists are EOL. Please join:

List:General Discussion« Previous MessageNext Message »
From:Paul DuBois Date:July 27 2004 6:59pm
Subject:Re: daisy-chaining replication
View as plain text  
At 14:31 -0400 7/27/04, Michael Dykman wrote:
>Hello,
>
>	I'm running a development build of 4.1.3 and trying to set up of chain
>of replication servers.  The first level of replication works
>beautifully but every effort to establish slaves on that first slave
>have been met with failure.
>
>here is the setup (log-bin is enabled on all 3 servers):
>
>  	server_a: is set up as primary master: all clients connect to this
>server and all data changes are made there.
>
>	server_b: is setup as client with a clean copy of the data (restarting
>server_a with 'skip-networking' to take the clean cut and gather the
>master status info, then restarting server_a with networking enabled
>again).  do the change master thing on server_b and start the slave, in
>due course 'show slave status' shows that it has caught up with server_a
>and all data is there
>
>	server_c: setup is a problem (though not necessarily 'the problem').
>having restarted server_b with skip-networking and stopped the slave, i
>take a cut of the data and get the master status info.  having put the
>data cut on server_c, executing
>CHANGE MASTER TO
>      MASTER_HOST='server_b',
>      MASTER_USER='replica_dude',
>      MASTER_PASSWORD='secret',
>      MASTER_LOG_FILE='server_b.000001',
>      MASTER_LOG_POS=79;
>  reports:
>         ERROR 1201 at line 1: Could not initialize master info
>         structure; more error messages can be found in the MySQL error
>         log
>
>contents of the error log:
>...
>  Failed to open the relay log './server_b-relay-bin.000001'
>(relay_log_pos 1011481184)
>040727 11:40:34  Could not find target log during relay log
>initialization
>040727 11:40:34  Failed to initialize the master info structure
>/usr/local/mysql/libexec/mysqld: ready for connections.
>Version: '4.1.3-beta-log'  socket: '/tmp/mysql.sock'  port: 3306
>040727 11:41:26  Failed to open the relay log
>'./server_b-relay-bin.000001' (relay_log_pos 1011481184)
>040727 11:41:26  Could not find target log during relay log
>initialization
>
>The file server_b-relay-bin.000001 certainly is there and is
>permissioned mysql:mysql, as are all the data files.
>
>the only really spooky thing I can find going on is on server_b (slave
>to server_a, intended master to server_c).  Having reenabled the slave
>and seeing it catch up to server_a (both 'show slave status' and
>empirical data tests confirm this) anther look at 'show master status'
>on server_b *still* reports:
>+---------------------+----------+--------------+------------------+
>| File                | Position | Binlog_Do_DB | Binlog_Ignore_DB |
>+---------------------+----------+--------------+------------------+
>| server_b-bin.000001 |       79 |              |                  |
>+---------------------+----------+--------------+------------------+
>
>it is not updating at all, even as replication data pours in.  I have
>reinstalled server_b as a slave 3 times now using the data from server_a
>and no matter how much replication data server_b has gathered from
>server_a post-install, the position is always 79 (!).
>
>Can anyone lend me a clue?

Did you start server b with --log-slave-updates so that it writes
the events that it receives from its master to its own binary log?
Just turning on binary logging on server b isn't enough.

-- 
Paul DuBois, MySQL Documentation Team
Madison, Wisconsin, USA
MySQL AB, www.mysql.com
Thread
daisy-chaining replicationMichael Dykman27 Jul
  • Re: daisy-chaining replicationPaul DuBois27 Jul
    • Re: daisy-chaining replicationMichael Dykman28 Jul
    • Re: daisy-chaining replicationMichael Dykman28 Jul