From the official blog:
- Initial sync from a secondary
- Slave delay
- Hidden servers
- Freeze e member
- Fast sync
Judging also by the number of replica set reported issues, I’d say a lot of effort have been put into addressing these before the next production ready version.
As a side note, it is interesting to note the different versioning “policies” used by some of the NoSQL projects:
Some very old projects that have been around and that were running in production for years are still at versions that make you think of early alpha stages: Hadoop 0.21 and yesterday’s HBase 0.90.0
Other NoSQL databases have announced their 1.0 version after years of development. Take for example Neo4j 1.0 or CouchDB 1.0
MongoDB, which indeed has very short release cycles, is already at version 1.6.x (production)/1.7.x (development) even if it has been around for at most 2 years. Redis, another quite young NoSQL database, is preparing these days to release version 2.2.
Probably the lesson here is never judge an open source project by its version number.
Original title and link: MongoDB Replica Sets: 5 Improvements for Upcoming 1.8 Version (NoSQL databases © myNoSQL)