Introducing TokuMX 1.4: Major improvements to MongoDB sharding and replication
Today we are thrilled to announce the availability of the latest release (1.4.0) of TokuMX, our high-performance distribution of MongoDB.
For those of you just joining us, TokuMX is an open-source distribution of MongoDB that shares much of the code, and all of the driver protocols, data model, and programming interface with MongoDB, but which stores all data on disk in Tokutek’s industrial-strength, write-optimized, ACID and MVCC Fractal Tree indexes. In short, it’s a drop-in replacement server for MongoDB applications that gets the most out of your hardware resources and adds great compression, document-level concurrency, transactional semantics, and operational ease-of-use and reliability to your database. Read more about TokuMX.
TokuMX 1.4 is full of features and optimizations that we’re really excited about, mostly centered around making the clustering experience (with either replication or sharding, or both) smoother, faster, and easier to administer. Some of the improvements are fixes to strange behaviors we had in previous versions that were detected by us or our customers, and some are just things that we’ve wanted to do for a long time, ever since we started TokuMX and decided to take responsibility for the whole stack.
The full release notes are here (and large), but we wanted to highlight some of the more impactful ones. So over the coming few days, we’re going to publish some impressive new benchmarks and a series of short posts introducing these features, showing what we did and why, what it means for you, and how to take the most advantage of them:
- Primary keys
- Partitioned oplog
- Optimized updates
- Smaller, faster sharded clusters
- Faster chunk migrations in sharded clusters