Environment Setup


This tutorial walks through the process of setting up a virtual environment useful for core development of the Hyperledger Sawtooth distributed ledger using Vagrant and VirtualBox. At the end, you will have an environment capable of running a validator, transaction processor, and other components. You will have cloned the Sawtooth repository.


The following tools are required:

These tools are available for the following platforms:

  • Linux
  • Windows
  • macOS


Vagrant is a tool which installs and configures virtual development environments. It allows development teams to easily specify and share consistent virtual machine configurations.

The dev tools directory (path ‘’sawtooth-core/tools’‘) of the sawtooth-core repository contains a Vagrant configuration which is specifically tailored to Sawtooth development. A new developer with installed copies of Vagrant and VirtualBox can clone the sawtooth-core repository and have a functional VM which can run validators within a few minutes.

A quick introduction to using dev tools is available as a quickstart in this Developer’s Guide.

Layout of Dev Tools

The dev tools directory is structured as follows:

Vagrantfile is the main configuration file for Vagrant. It is ruby code executed by Vagrant and is executed every time a Vagrant command is executed.
The bootstrap.d directory contains a set of bash scripts which are executed in order during the provisioning step of ‘vagrant up’. These scripts are always executed.

The guest-files directory contains configuration files which are used by the bootstrap.d scripts.

There is also a local-env.sh script which contains environment specific variables for Sawtooth development, such as PYTHONPATH.

The plugins directory contains bash scripts which can be easily configured to execute during the provisioning step of ‘vagrant up’. These scripts run after bootstrap.d scripts.
This scripts directory contains scripts which are sometimes useful to the developer after provisioning has been completed and the developer has a shell in the virtual machine. For example, there are scripts which help build Ubuntu packages.
The tests directory includes tests run within the Vagrant environment. These are in addition to the unit and integration tests found in the tests directory of sawtooth-core and sawtooth-validator.
The win-bin directory includes scripts for running Sawtooth natively under Windows and is not used in the Vagrant environment.

Layout in the Virtual Machine

A convenient feature of Vagrant is the easy sharing of directories between the host machine and the guest virtual machine. By default the Vagrant configuration directory is mounted under /vagrant in the virtual machine.

In the dev tools configuration a /project mount point is also defined which provides access to the Sawtooth repositories.

Step One: Clone Repository

You’ll need to have git installed in order to clone the Sawtooth source code repository. You can find up-to-date installation instructions here:


When checking out Sawtooth on Windows for use with Vagrant, you should take steps to ensure that Windows-style CRLF line endings are not added to the code. The bash scripts used by your Vagrant VM will not run correctly with CRLF line endings. Git uses a configuration setting, core.autocrlf, to control whether or not LF-style line endings are automatically converted to CRLF-style line endings. This setting should be set so that CRLFs are not introduced into your repository.

Open up a terminal and run the following:

% cd $HOME
% mkdir project
% cd project
% git clone https://github.com/hyperledger/sawtooth-core.git


On a Windows environment, the suggested version of the last command above is:

C:\> git clone https://github.com/hyperledger/sawtooth-core.git
--config core.autocrlf=false

Configure Proxy (Optional)

If you are behind a network proxy, follow these steps before continuing:

  1. Set the following environment variables:
  • http_proxy
  • https_proxy

If you are using the Bash shell, run the following commands:


The example URLs and port numbers used below are examples only. Please substitute the actual URL, with actual port numbers, used in your environment. Contact your network administrator for the information if necessary.

% export http_proxy=http://example-proxy-server.com:3128
% export https_proxy=https://example-proxy-server.com:3129

If you are using Windows, run the following commands:

% set http_proxy=http://example-proxy-server.com:3128
% set https_proxy=https://example-proxy-server.com:3129
  1. Install the vagrant-proxyconf plugin:
% cd sawtooth-core/tools
% vagrant plugin install vagrant-proxyconf

Build and Run Virtual Machine

In order to start the Vagrant VM, run:

% cd sawtooth-core/tools
% vagrant up


We have encountered an intermittent problem on Windows hosts which presents as an ‘Operation not permitted’ error in the Vagrant startup output. If you encounter this error, perform a ‘vagrant destroy’ and then run ‘vagrant up’ again.

Downloading the Vagrant box file, booting the VM, and running through the bootstrap scripts will take several minutes.

Once the ‘vagrant up’ command has finished executing, run:

% vagrant ssh

By default, Vagrant sets up ssh keys so that users can log into the VM without setting up additional accounts or credentials. The logged in user, vagrant (uid 1000), also has permissions to execute sudo with no password required. Any number of vagrant ssh sessions can be established from the host.


Occasionally, the configuration files used to create the Vagrant environment change and this can result in Vagrant asking for a password when doing vagrant up or vagrant ssh. This is usually a sign that your environment is out of date. If this happens, please follow the instructions below to reset your environment. This is especially common when switching between versions of Sawtooth, eg. 0.7 -> 0.8.

Reset The Environment (Optional)

If the VM needs to be reset for any reason, it can be returned to the default state by running the following commands from the sawtooth-core/tools directory on the host:

% vagrant destroy
% vagrant up


Vagrant destroy will delete all contents within the VM. However, /vagrant and /project are shared with the host and will be preserved.

Build Components

Some of the components of Sawtooth depend partially on code that must first be built. This includes generating protobuf classes for each language. To build all of the components of Sawtooth within Vagrant, do:

$ /project/sawtooth-core/bin/build_all

Building Individual Components

If you need to build a component related to a single language or componenent, you can save time by running the build command for that component rather than build all. For example, to build only Python components, run:

$ /project/sawtooth-core/bin/build_python

Running Tests (Optional)

The automated tests for python and all other languages rely on docker to ensure reproducibility. To run the automated tests for python, first run the following:

$ /project/sawtooth-core/bin/build_all -l python

This will create docker images for all the python components and run build_python inside a build container. You can then run the automated tests for python components, while excluding java and javascript components, with:

$ /project/sawtooth-core/bin/run_tests -x java_sdk -x javascript_sdk


The run_tests command provides the -x flag to allow you to exclude various components from tests. You can also specify which tests to run with the -m flag. Run the command run_tests -h for help.

If you are not behind a proxy, you can build and test everything Sawtooth has to offer with:

$ /project/sawtooth-core/bin/build_all
$ /project/sawtooth-core/bin/run_tests