On the management node, create a file (config.ini) on the management node directory intended for datadir.
An example starting configuration can be found in the MySQL documentation.
The following sections and variables can be used for a basic deployment:
|
Parameter Name |
Description |
Default Value |
Suggested Value |
|
[ndb_mgmd default]\DataDir |
The Data Directory on the management node. |
N/A |
N/A |
|
[ndb_mgmd]\HostName |
The hostname of the management node. For multiple management servers duplicate the [ndb_mgmd] section. |
N/A |
N/A |
|
The number of fragment replicas. A minimum of two is required to guarantee the availability of data. |
2 |
Min 2 NoOfReplicas for redundancy. Max 4. |
|
|
This is the amount of memory available to data being stored in the NDB cluster. |
98M |
> 1Gb |
|
|
Sets the number of REDO log files per node. |
16 |
If checkpointing is slow increase this value. |
|
|
If this parameter is enabled then NDB will attempt to use O_DIRECT writes for storage operations. |
false |
Recommended for Linux systems when using EXT4. |
|
|
If enabled data node threads will be done using real-time scheduling. |
false |
Enabling real-time scheduling can increase Data Node throughput. |
|
|
Defines the time between global checkpoints in milliseconds. |
2000 |
< 3000 |
|
|
Sets the size of the buffer in which the REDO log is written. |
32MB |
Increase this value for systems with frequent updates to data. |
|
|
Only if using ndbmtd affects the number of execution threads on a Data Node. |
2 |
Depending on CPU > 16 |
|
|
Controls the maximum allows mean latency for disk access. Note:
Recommendation for FlashArray and controlling IO Subsystem pressure: QOS. |
0 |
0 |
|
|
[ndbd]\ HostName |
The Hostname or IP Address of the Data Node. |
N/A |
N/A |
|
[ndbd]\ NodeId |
(Optional) The Node ID of the Data Node. |
N/A |
N/A |
|
[ndbd]\ DataDir |
The Data directory where NDB data will be stored on a Data Node. |
N/A |
N/A |
|
The directory in which backups are placed. |
DataDir/FilesystemPath |
Path to FlashArray MySQL Cluster Backup Volume for Node for NFS share for Node. Each Node should have its own Backup directory. |
|
|
[mysqld]\ HostName |
The Hostname of a SQL Node. |
N/A |
N/A |
This is an example config.ini file:
[ndb_mgmd default]
# Directory for management node log files
DataDir=/var/lib/mysql-cluster
[ndb_mgmd]
# Management Node IP Address
HostName=10.21.227.36
[ndbd default]
NoOfReplicas=2 # Number of replicas
DataMemory=368G # Memory allocate for data storage
TimeBetweenGlobalCheckpoints=3000
[ndbd]
# Data Node
HostName=10.21.227.39
NodeId=2
DataDir=/usr/local/mysql/data
BackupDataDir=/usr/local/mysql/backup
[ndbd]
# Data Node
HostName=10.21.227.40
NodeId=3
DataDir=/usr/local/mysql/data
BackupDataDir=/usr/local/mysql/backup
[mysqld]
# SQL Node
HostName=10.21.227.37
[mysqld]
#SQL Node
HostName=10.21.227.38