Jonathan Ellis details what’s new in Cassandra 0.6.5:
- Dynamic Snitch
- Use mlockall via JNA, if present, to prevent Linux from swapping out parts of the JVM
- Page within a single row during hinted handoff
- Faster UUIDType, LongType comparisons
- Log summary of dropped messages instead of spamming log
It’s interesting to hear that Cassandra had to go out of the VM in order to optimize its behavior by using
mlockall on OSes supporting it. Also learning about the dynamic snitch:
Cassandra has always been good at dealing with cluster members who are all the way dead, thanks to its failure detector. The dynamic snitch lets us also handle members who are only mostly dead, that is, are still responding but with impaired performance.
The dynamic Snitch incorporates real-time request latency into its closeness metric, and routes requests to nodes that respond the fastest, no matter where they are actually located.
Original title and link for this post: Cassandra 0.6.5: What is New? (published on the NoSQL blog: myNoSQL)