Requirements

First, you must clone this repo to some directory of your choosing. We will call this <K9DB_DIR>. You should also initialize its git submodules.

git submodule init
git submodule update

After cloning, you have multiple options for building and running K9DB:

1. By using our provided docker containers: Dockerfile.release packages a ready-to-run K9db server (recommended for trying out K9db), while Dockerfile.dev is for building K9db from source, running tests, and running experiments. On arm64 machines (e.g. Apple Silicon Macs), use Dockerfile.dev.arm64 instead of Dockerfile.dev: a lightweight development image for building K9db from source and running tests (it does not include the experiments pipeline).
2. By installing the requirements and building K9db yourself.
3. By using our provided scripts if you are running on google cloud (recommend if you are trying to reproduce the results from our paper). Consult the experiments section for more details on this.

Currently, you cannot build K9db directly on MacOS. You must instead use the docker containers. Apple Silicon Macs (M series) are supported: the docker images build and run as native arm64 Linux containers (Dockerfile.release is multi-arch, and Dockerfile.dev.arm64 is provided for development).