While we've worked hard to make updates to newer versions of Neo4j as pleasant as possible, these are a few things to be aware of, particularly when switching to a major revision.
Neo4j 3.3 Upgrade FAQUpgrade FAQs for Neo4j 3.2 and earlier
What is in 3.3?
- Write and read performance
- Write performance of transactional writes have improved with on average 50% for typical graphs compared to Neo4j 3.2, and with over 100% compared to Neo4j 3.1.
- Bulk writes at initial graph creation has an up to 40% smaller memory footprint
- The new Cypher Slotted Runtime results in faster queries while using one third of the memory compared to the 3.2 Cypher Runtime
- Least-connected load balancing strategy for clusters minimizes the effect on cluster performance when one server experiences heavy load
- Enterprise Security
- New support introduced for intra-cluster encryption, including multi-DC cluster communication encryption
- Kernel improvements
- Neo4j Enterprise now allows key configuration parameters to be changed on the fly, without needing to recycle a database instance.
- Local Schema Locks automatically narrow the scope of locks, avoiding the need for the database to take a global schema lock when creating or changing a schema object or constraint.
- Cache-hit ratio is now available at both a query and database level, as a new metric for helping to size one’s database cache.
Where are the Release Notes?Release notes for all Neo4j releases can be found here: http://neo4j.com/release-notes/
Is 3.3 an API-Breaking Release?No. Minor releases are a place to deprecate features and functionality, which includes public APIs. However, we will not remove deprecated functionality until a major release. Note: Upgrading from a version prior to Neo4j 3.0 means that the 3.0 Upgrade FAQs apply, and this would then be a breaking release. If you are upgrading from a 2.x release, see the Upgrade FAQ Archive.
A few things you should know about 3.3:
- The new Slotted Runtime will be used for most Cypher queries. Use the PROFILE or EXPLAIN keywords to see which runtime is being used if unsure.
- While Causal Clustering is the primary clustering architecture we are building on, HA still works as well.
- Causal Clustering is still available for server only, not embedded at this time.
- neo4j-shell is deprecated in favor of the new cypher-shell
- neo4j-backup is deprecated in favor of neo4j-admin backup
- Configuration settings that changed:
- dbms.allow_format_migration is now dbms.allow_upgrade