Installing Ray

Ray supports Python 2 and Python 3 as well as MacOS and Linux. Windows support is planned for the future.

Latest stable version

You can install the latest stable version of Ray as follows.

pip install -U ray  # also recommended: ray[debug]

Latest Snapshots (Nightlies)

Here are links to the latest wheels (which are built for each commit on the master branch). To install these wheels, run the following command:

pip install -U [link to wheel]
Linux MacOS
Linux Python 3.7 MacOS Python 3.7
Linux Python 3.6 MacOS Python 3.6
Linux Python 3.5 MacOS Python 3.5
Linux Python 2.7 MacOS Python 2.7

Building Ray from Source

Installing from pip should be sufficient for most Ray users.

However, should you need to build from source, follow instructions below for both Linux and MacOS.

Dependencies

To build Ray, first install the following dependencies. We recommend using Anaconda.

For Ubuntu, run the following commands:

sudo apt-get update
sudo apt-get install -y build-essential curl unzip psmisc

# If you are not using Anaconda, you need the following.
sudo apt-get install python-dev  # For Python 2.
sudo apt-get install python3-dev  # For Python 3.

pip install cython==0.29.0

For MacOS, run the following commands:

brew update
brew install wget

pip install cython==0.29.0

If you are using Anaconda, you may also need to run the following.

conda install libgcc

Install Ray

Ray can be built from the repository as follows.

git clone https://github.com/ray-project/ray.git

# Install Bazel.
ray/ci/travis/install-bazel.sh

# Optionally build the dashboard (requires Node.js, see below for more information).
pushd ray/python/ray/dashboard/client
npm ci
npm run build
popd

# Install Ray.
cd ray/python
pip install -e . --verbose  # Add --user if you see a permission denied error.

[Optional] Dashboard support

If you would like to use the dashboard, you will additionally need to install Node.js and build the dashboard before installing Ray. The relevant build steps are included in the installation instructions above.

The dashboard also requires:

  • Python 3
  • aiohttp
  • psutil

The latter two can be installed via pip:

pip install aiohttp psutil

The dashboard is enabled by setting include_webui=True during initialization, i.e.

import ray
ray.init(include_webui=True)

Docker Source Images

Run the script to create Docker images.

cd ray
./build-docker.sh

This script creates several Docker images:

  • The ray-project/deploy image is a self-contained copy of code and binaries suitable for end users.
  • The ray-project/examples adds additional libraries for running examples.
  • The ray-project/base-deps image builds from Ubuntu Xenial and includes Anaconda and other basic dependencies and can serve as a starting point for developers.

Review images by listing them:

docker images

Output should look something like the following:

REPOSITORY                          TAG                 IMAGE ID            CREATED             SIZE
ray-project/examples                latest              7584bde65894        4 days ago          3.257 GB
ray-project/deploy                  latest              970966166c71        4 days ago          2.899 GB
ray-project/base-deps               latest              f45d66963151        4 days ago          2.649 GB
ubuntu                              xenial              f49eec89601e        3 weeks ago         129.5 MB

Launch Ray in Docker

Start out by launching the deployment container.

docker run --shm-size=<shm-size> -t -i ray-project/deploy

Replace <shm-size> with a limit appropriate for your system, for example 512M or 2G. The -t and -i options here are required to support interactive use of the container.

Note: Ray requires a large amount of shared memory because each object store keeps all of its objects in shared memory, so the amount of shared memory will limit the size of the object store.

You should now see a prompt that looks something like:

root@ebc78f68d100:/ray#

Test if the installation succeeded

To test if the installation was successful, try running some tests. This assumes that you’ve cloned the git repository.

python -m pytest -v python/ray/tests/test_mini.py

Troubleshooting installing Arrow

Some candidate possibilities.

You have a different version of Flatbuffers installed

Arrow pulls and builds its own copy of Flatbuffers, but if you already have Flatbuffers installed, Arrow may find the wrong version. If a directory like /usr/local/include/flatbuffers shows up in the output, this may be the problem. To solve it, get rid of the old version of flatbuffers.

There is some problem with Boost

If a message like Unable to find the requested Boost libraries appears when installing Arrow, there may be a problem with Boost. This can happen if you installed Boost using MacPorts. This is sometimes solved by using Brew instead.