This section provides information about basic requirements for performing backup and restore operations using MySQL Cluster Manager.
Requirements for MySQL Cluster backup. Basic requirements for performing MySQL backups using MySQL Cluster Manager are minimal. At least one data node in each node group must be running, and there must be sufficient disk space on the node file systems. Partial backups are not supported.
Requirements for MySQL Cluster restore. In general, the following requirements apply when you try to restore a MySQL Cluster using MySQL Cluster Manager:
A complete restore requires that all data nodes are up and running, and that all files belonging to a given backup are available.
A partial restore is possible, but must be specified as such. This can be accomplished using the
restore clusterclient command with its
--skip-nodeidoption. See Section 126.96.36.199, “Partial restore—missing images” for details.
In the event that data nodes have been added to the cluster since the backup was taken, only those data nodes for which backup files exist are restored. In such cases data is not automatically distributed to the new nodes, and, following the restore, you must redistribute the data manually by issuing an
ALTER ONLINE TABLE ... REORGANIZE PARTITIONstatement in the mysql client for each
NDBtable in the cluster. See Adding NDB Cluster Data Nodes Online: Basic procedure, for details.
To restore a backup created by MySQL Cluster Manager to a cluster with fewer data nodes, you need to restore first the logical backup of the metadata of the NDB tables (available in backups created by MySQL Cluster Manager 1.4.1 and after) using the mysqldump utility and then restore the table data using the ndb_restore program. See Section 188.8.131.52, “Restoring a Backup to a Cluster with Fewer Data Nodes” for details.