Building a High-Availability MySQL Cluster
Pages: 1, 2, 3

Installing MySQL-HA-Cluster

Initially, MySQL didn't provide the NDB build in RPM form, but they did include source RPMs so you could build your own. These days, MySQL kindly has split the different binaries into different packages: MySQL-client-4.1.12-1.i386.rpm, MySQL-devel-4.1.12-1.i386.rpm, MySQL-ndb-extra-4.1.12-1.i386.rpm, MySQL-ndb-management-4.1.12-1.i386.rpm, MySQL-ndb-storage-4.1.12-1.i386.rpm, MySQL-ndb-tools-4.1.12-1.i386.rpm, and MySQL-server-4.1.12-1.i386.rpm.

Install MySQL-ndb-management on the management node. Install MySQL-ndb-storage, MySQL-ndb-tools, MySQL-NDB-extra, and the usual MySQL-client and MySQL-server on the database nodes (here, DB-A and DB-B).

On the management node, the configuration file (config.ini) resides in the directory /var/lib/mysql/mysql-cluster:

[ndbd default]
NoOfReplicas= 2 

[mgm]
HostName= 172.16.32.1

[ndbd]
HeartbeatIntervalDbDb=600
HeartbeatIntervalDbApi=600
HostName= 172.16.32.7
DataDir: /var/lib/mysql/mysql-cluster/
DataMemory = 100M
IndexMemory = 50M
MaxNoOfConcurrentTransactions = 500
MaxNoOfConcurrentOperations = 250000
TimeBetweenWatchDogCheck = 40000
MaxNoOfOrderedIndexes = 27000
MaxNoOfTables = 9000
MaxNoOfAttributes = 25000

[ndbd]
HeartbeatIntervalDbDb=600
HeartbeatIntervalDbApi=600
HostName = 172.16.32.107
DataDir: /var/lib/mysql/mysql-cluster/
DataMemory = 100M
IndexMemory = 50M
MaxNoOfConcurrentTransactions = 50
MaxNoOfConcurrentOperations = 250000
TimeBetweenWatchDogCheck = 40000
MaxNoOfOrderedIndexes = 27000
MaxNoOfTables = 9000
MaxNoOfAttributes = 25000

[mysqld]
[mysqld]
[mysqld]

This is the main config file that you need to get MySQL NDB cluster up and running. In this version, ndbd has two replicas. The management server listens on 172.16.32.1. In my work project, our current needs don't exceed a 100MB data segment and a 50MB index memory.

The three unnamed [mysqld] sections refer to the two mysqld processes on the different DBs and a possible third mysqld engine.

With this config in place, you can start the ndb_mgmd for other nodes to connect to.

Configuring the NDB nodes

Both data nodes have a file called Ndb.cfg in /var/lib/mysql/mysql-cluster (also symlinked to /var/lib/mysql/):

DB-A:/var/lib/mysql # cat Ndb.cfg
host=172.16.32.1:1186

The first time you start ndbd, run it with the --initial parameter.

Running ndb_mgm now should show two active ndb nodes:

DB-A:/var/lib/mysql # ndb_mgm
-- NDB Cluster -- Management Client --
ndb_mgm> show
Connected to Management Server at: 172.16.32.1:1186
Cluster Configuration
---------------------
[ndbd(NDB)]   2 node(s)
id=2  @172.16.32.7 (Version: 4.1.12, Nodegroup: 0)
id=3  @172.16.32.107 (Version: 4.1.12, Nodegroup: 0, Master)

[ndb_mgmd(MGM)] 1 node(s)
id=1  @172.16.32.1 (Version: 4.1.12)

If you can see this, you've accomplished the first part of the mission. Congratulations! You now have a working NDB engine cluster.

Pages: 1, 2, 3

Next Page