This guide describes the options for running Pravega for development, testing and in production. If you are interested on deploying Pravega in production, we recommend you to have a look to our Configuration and Provisioning Guide.
There are two modes for running Pravega.
- Standalone - Standalone mode is suitable for development and testing Pravega applications. It can either be run from the source code, from the distribution package or as a docker container.
- Distributed - Distributed mode runs each component separately on a single or multiple nodes. This is suitable for production in addition for development and testing. The deployment options in this mode include a manual installation, running in a docker swarm, or using the Kubernetes operators.
The following prerequisites are required for running Pravega in all modes.
- Java 11
The following prerequisites are required for running in production. These are only required for running in distributed mode.
- External HDFS 2.7
- Zookeeper 3.6.1
- Bookkeeper 4.11.1
For more details on the prerequisites and recommended configuration options for bookkeeper see the Manual Install Guide.
There are multiple options provided for running Pravega in different environments. Most of these use the installation package from a Pravega release. You can find the latest Pravega release on the GitHub Releases page.
- Local - Running Pravega locally is suitable for development and testing.
- Production - Multi-node installation suitable for running in production.
- Configuration and Provisioning Guide for production clusters.