Enabling High Availability for Kafka Schema Registry

You can enable high availability by installing multiple instances of schema registry on the same cluster.

About this task

One node in the cluster acts as the primary instance. Only the primary instance can publish writes to the underlying Kafka log. The primary node can also manage read requests.

All other secondary nodes manage only read requests. These nodes serve registration requests indirectly by forwarding them to the current primary node and returning any response supplied by the primary node.

Procedure

  1. Install schema registry on several nodes. For installation instructions, see Installing Kafka Schema Registry.
  2. Configure high availability for the ensemble of ZooKeeper nodes. You can find the steps for this in the open source documentation, section Clustered (Multi-Server) Setup.
  3. Modify the schema registry configuration file on each node to specify the kafkastore.connection.url property with the ZooKeeper nodes that form the HA ensemble.
  4. Run the configure.sh -R command on each schema registry node.