MySQL and HeatWave

  • May 10, 2017

MySQL Cluster 7.6: The First Milestone

Matt Lord
Senior MySQL Product Manager

The first Development Milestone Release (DMR) of MySQL Cluster 7.6--7.6.2--is now available! You can see the full changelog for 7.6.2 here. But before we get into the specifics on 7.6.2, I'd  like to take a minute just to paint the bigger picture of where we want to go with MySQL Cluster over the long term. While we cannot and will not discuss or project version numbers and/or dates for any given feature, it's helpful to know what our bigger goals and longer term roadmap are with MySQL Cluster (see the standard Oracle Safe Harbor disclaimer).

Long Term Goals 

We've set some pretty lofty goals for the MySQL Cluster development in future major releases, with our primary focus areas being as follows--with some examples of each:

Cloud Native
  • Scaling data nodes up and down
  • Security
    – Data encrypted in transit and at rest
    – AuthN and AuthZ in all nodes (e.g. ndb_mgmd
  • Improved service registry and discovery
  • Improved orchestration and automation
    – Providing our own tooling
    – Operating well within third party frameworks: Ansible, SaltStack, Puppet, Chef, etc.

Greater horizontal and vertical scaling
  • Greater than 48 data nodes and 255 total nodes 
  • Efficient 100s TB sized clusters

Improved ease-of-use
  • Better integration with MySQL Enterprise Monitor 
  • Zero configuration installs 
  • Improved tooling
    – Official yum/apt repos (available now!) and official Docker containers
    – Official Init / Systemd scripts
  • More dynamic (eliminating node failures and manual restarts)
    – Eliminate internal hardcoded limits on buffers / caches
    – Make configuration changes settable in the running process
    – Make buffers/caches adaptive and able to scale up and down as needed

Improved performance
  • Multi-threaded geo-replication between clusters
  • Parallelized backups and restores 
  • Shared memory transporter

Improved data redundancy
  • Full support for 1-4 replicas  

Highlights for 7.6.2 

MySQL Cluster 7.6 is our first release where we begin implementing this bigger long term vision. With this big picture context in mind, here are some highlights of what's new in the first 7.6 DMR:

  • A new disk based table file format, which eliminates a number of historical issues. This is one factor that can allow much larger total cluster sizes, as you can move your cold/archive data to disk based tables.
  • Our first move towards a single flexible and dynamic memory allocation pool by the merging of the IndexMemory parameter and structure into DataMemory.
  • A significant overall performance improvement by removing some now unnecessary debug facilities (these are "JAMs" within NDB that dump debug information to trace files).
  • Various ndbinfo additions and other metadata improvements that pave the way for making MySQL Cluster much easier to monitor and manage (stay tuned for future announcements around MySQL Enterprise Monitor and MySQL Cluster). 
  • A new AutoReconnect feature in ClusterJ--the native NoSQL NDB API implementation available for Java--which makes it much easier to write zero downtime Java applications for MySQL Cluster.
  • A new bulk data load tool: ndb_import. This makes it far easier to work with large data sets and perform bulk data loads and migrations by providing a highly parallel and concurrent tool for importing data. 
  • The very beginning of our work to improve the Local Data Manager component and the Local Checkpoint Protocol behavior, both of which are critical to our goals of supporting 100s TB sized clusters. 

So as you can see, with this first 7.6 DMR we've begun to lay the groundwork for many of our long term goals. We're really excited about what's coming in future 7.6 DMRs and beyond! 

You can download 7.6.2 here (Development Releases tab). I would encourage you to try it out! If you feel that you've encountered a bug, please let us know by filing a bug report or opening a support ticket

THANK YOU for using MySQL! 


Safe Harbor Statement

This content is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions. The development, release, and timing of any features or functionality described for Oracle’s products remains at the sole discretion of Oracle.


Be the first to comment

Comments ( 0 )
Please enter your name.Please provide a valid email address.Please enter a comment.CAPTCHA challenge response provided was incorrect. Please try again.