The latest update of Kong (0.8.3) ships with some bugfixes:
- Fixed an edge case that was preventing new nodes from properly registering themselves into the “nodes” table.
- Fixed a bug that was occurring when using OAuth 2.0 and Postgres.
- And more.
To update to the latest version, simply follow the installation instructions.
Cassandra or PostgreSQL?
With the release of Kong 0.8.x you can now finally decide to use either Cassandra or PostgreSQL for storing your data. Cassandra and Postgres are two different databases with different engines that serve different use-cases.
We recommend using Cassandra in multi-datacenter setups, since Cassandra’s model eventually replicates the data across every datacenter, while still being master-less (which means you can write and read from any node in any datacenter).
Postgres has an easier replication model than Cassandra and usually it’s a more familiar technology. We recommend using Postgres when planning to run Kong in a single datacenter since it’d be easier to get started. You can still replicate data to a slave server located in a different datacenter, but you would always have to write to the master server (which could mean extra latency if you are writing data from a Kong node located in a different datacenter). That said, there are some technologies that still allow to replicate a master-less Postgres in multi-DC environment.
Upgrading from < 0.8.x
If you are migrating from an older version of Kong < 0.8.x, then follow the migration instructions in order to make your Kong cluster functioning after updating to 0.8.x.
We are looking forward to seeing you at our next webinar – register for a hands-on demo of Kong.