MySQL Cluster Manager 8.0 Release Notes  /  Changes in MySQL Cluster Manager 8.0.31 (2022-10-11, General Availability)

Changes in MySQL Cluster Manager 8.0.31 (2022-10-11, General Availability)

Functionality Added or Changed

  • Agent: Previously, an attempted cluster upgrade from the NDB 7.6 series to the 8.0 series that was rejected on the first start of a new ndb_mgmd node could not be rolled back using the upgrade cluster command. With this fix, the user can now perform the rollback in the situation to return the cluster to a stable state. (Bug #34432349)

  • Agent: If a set command was attempted with the same value as was already in use, for some configuration parameters, mcmd would still perform some unnecessary process or cluster restarts. With this fix, the command returns an error in the situation, telling the user that the command causes no changes to the cluster configuration. Also, a new option --retry, has been added to the set command, to be used when the user really needs to run the set command and force a restart of the relevant processes. (Bug #22588763)

  • Agent: The stop process command now accepts a node type or a list of node IDs as its argument for stopping multiple processes. See the command description for stop process for details. (Bug #16879795)

  • Client: Since NDB 8.0.31, the --core-file command-line-only attribute is no longer supported by ndb_mgmd. As a result, mcmd no longer supports configuring the option on ndb_mgmd nodes. (Bug #34629250)

  • Transparent Data Encryption (TDE) for NDB Cluster (see File System Encryption for NDB Cluster) is now supported by MySQL Cluster Manager. See Setting Up Encryption, for more information. (Bug #34122958, WL #15319)

Bugs Fixed

  • Agent: mcmd could hang when a packet larger than 16MB was sent between the client and mcmd. It was because mcmd miscalculated the packet size and thus failed to chunk the packet into multiple packets for proper transmission, and this fix corrects the problem. (Bug #34578419)

  • Agent: When a newly added data node failed to start, attempts to restart it again with start process and the --initial option always resulted in errors with the complaint Nodegroup unavailable. (Bug #34307445)

  • Client: mcmd used to accept some invalid configuration options silently and simply ignored them. With this fix, invalid configuration options are rejected with errors. (Bug #34450589)

  • The MySQL Cluster Manager installation included a number of unnecessary libraries, which have now been removed. (Bug #34456516)