This section outlines the procedure for starting NDB Cluster replication using a single replication channel.
Start the MySQL replication master server by issuing this command:
mysqld --ndbcluster --server-id=
In the previous statement,
idis this server's unique ID (see Section 18.6.2, “General Requirements for NDB Cluster Replication”). This starts the server's mysqld process with binary logging enabled using the proper logging format.Note
You can also start the master with
--binlog-format=MIXED, in which case row-based replication is used automatically when replicating between clusters.
STATEMENTbased binary logging is not supported for NDB Cluster Replication (see Section 18.6.2, “General Requirements for NDB Cluster Replication”).
Start the MySQL replication slave server as shown here:
mysqld --ndbcluster --server-id=
In the command just shown,
idis the slave server's unique ID. It is not necessary to enable logging on the replication slave.Note
You should use the
--skip-slave-startoption with this command or else you should include
skip-slave-startin the slave server's
my.cnffile, unless you want replication to begin immediately. With the use of this option, the start of replication is delayed until the appropriate
START SLAVEstatement has been issued, as explained in Step 4 below.
It is necessary to synchronize the slave server with the master server's replication binary log. If binary logging has not previously been running on the master, run the following statement on the slave:
CHANGE MASTER TO->
This instructs the slave to begin reading the master's binary log from the log's starting point. Otherwise—that is, if you are loading data from the master using a backup—see Section 18.6.8, “Implementing Failover with NDB Cluster Replication”, for information on how to obtain the correct values to use for
MASTER_LOG_POSin such cases.
Finally, you must instruct the slave to begin applying replication by issuing this command from the mysql client on the replication slave:
This also initiates the transmission of replication data from the master to the slave.
It is also possible to use two replication channels, in a manner similar to the procedure described in the next section; the differences between this and using a single replication channel are covered in Section 18.6.7, “Using Two Replication Channels for NDB Cluster Replication”.
It is also possible to improve cluster replication performance by
enabling batched updates.
This can be accomplished by setting the
variable on the slave mysqld processes.
Normally, updates are applied as soon as they are received.
However, the use of batching causes updates to be applied in 32 KB
batches, which can result in higher throughput and less CPU usage,
particularly where individual updates are relatively small.
Slave batching works on a per-epoch basis; updates belonging to more than one transaction can be sent as part of the same batch.
All outstanding updates are applied when the end of an epoch is reached, even if the updates total less than 32 KB.
Batching can be turned on and off at runtime. To activate it at runtime, you can use either of these two statements:
SET GLOBAL slave_allow_batching = 1; SET GLOBAL slave_allow_batching = ON;
If a particular batch causes problems (such as a statement whose effects do not appear to be replicated correctly), slave batching can be deactivated using either of the following statements:
SET GLOBAL slave_allow_batching = 0; SET GLOBAL slave_allow_batching = OFF;
You can check whether slave batching is currently being used by
means of an appropriate
VARIABLES statement, like this one:
SHOW VARIABLES LIKE 'slave%';+---------------------------+-------+ | Variable_name | Value | +---------------------------+-------+ | slave_allow_batching | ON | | slave_compressed_protocol | OFF | | slave_load_tmpdir | /tmp | | slave_net_timeout | 3600 | | slave_skip_errors | OFF | | slave_transaction_retries | 10 | +---------------------------+-------+ 6 rows in set (0.00 sec)