Igniters, I'm happy to announce that the 5th alpha version of Ignite 3 is out!
On top of the functionality that was previously released, Alpha 5 adds the following major features: - Pluggable storages: ability to choose a specific storage for a table (RocksDB based storage, Page memory persistent and in-memory storage) with some known limitations. - Compute API (A simple remote job execution): The first phase of Compute API design and implementation. Of course, with known limitations. - Data colocation: The colocation key concept replaces the affinity key concept. DDL introduces COLOCATE BY clause. Colocated job execution. - Open API for the Ignite REST endpoints: A Specification to generate a client for any language + auto-generated docs for REST API. - Ignite REPL: The Ignite CLI as a REPL with autocompletion and improved UX. - Cluster lifecycle: It introduces cluster initialization logic and allows to specify cluster management and meta storage groups. Improved node join protocol. - Local and distributed recovery: Now it is possible to restart a cluster/node without data loss. - Data rebalance improvements (in progress and could be excluded from the release), including dynamically changing the number of partition replicas. - Robust client connection with seamless reconnection support and retry policies. - Java API for SQL: A simplified API for executing SQL queries on a cluster. Code examples have been added for the new features, you can check them out here: https://github.com/apache/ignite-3/tree/3.0.0-alpha5/examples The best way to try the release out is to go through the Getting Started Guide: https://ignite.apache.org/docs/3.0.0-alpha If there are any questions, issues, or thoughts, please do not hesitate to reply to this email. Ignite Community is welcoming any feedback and will consider it in future development.