MongoDB 3.0 (dev version of 2.8) has been almost a year in the making and there are several prominent changes that are worth mentioning:
- Document level locking – changes to those large json documents (aka rows ) will work similarly to regular RDBMS. Previously you had to either queue or create a custom way of handling inconsistencies.
- Document has a hard limit of 16megs (reasonable) and that allows your app to create massive amount of data that will require sharding your servers. A new storage engine WiredTiger will provide better compression for storage, allegedly up to 80% , shed those Gigs of data Foursquare!
- Besides other significant improvements, a new tool is introduced named Ops Manager, uncertain if it would be available to all or MMS subscribers.
- Looking into MongoDB Jira , there is a large set of bugs that have been fixed, better error handling in GEO lookups, syncing issues between nodes and others.
A few notes:
In order to take advantage of a new pluggable storage engine wiredTiger you need to explicitly configure it as such:
mongod --storageEngine wiredTiger
Overall great job!