OAuth + JupyterHub Authenticator = OAuthenticator
OAuthenticator currently supports the following authentication services:
A generic implementation, which you can use with any provider, is also available.
For an example docker image using OAuthenticator, see the examples directory.
Another example is using GitHub OAuth to spawn each user's server in a separate docker container.
Install with pip:
pip3 install oauthenticator
Or clone the repo and do a dev install:
git clone https://github.com/jupyterhub/oauthenticator.git
cd oauthenticator
pip3 install -e .
The first step is to tell JupyterHub to use your chosen OAuthenticator. Each
authenticator is provided in a submodule of oauthenticator
, and each
authenticator has a variant with Local
(e.g. LocalGitHubOAuthenticator
),
which will map OAuth usernames onto local system usernames.
In jupyterhub_config.py
, add:
from oauthenticator.github import GitHubOAuthenticator
c.JupyterHub.authenticator_class = GitHubOAuthenticator
All OAuthenticators require setting a callback URL, client ID, and client secret. You will generally get these when you register your OAuth application with your OAuth provider. Provider-specific details are available in sections below. When registering your oauth application with your provider, you will probably need to specify a callback URL. The callback URL should look like:
http[s]://[your-host]/hub/oauth_callback
where [your-host]
is where your server will be running. Such as
example.com:8000
.
When JupyterHub runs, these values will be retrieved from the environment variables:
$OAUTH_CALLBACK_URL
$OAUTH_CLIENT_ID
$OAUTH_CLIENT_SECRET
You can also set these values in your configuration file, jupyterhub_config.py
:
c.MyOAuthenticator.oauth_callback_url = 'http[s]://[your-host]/hub/oauth_callback'
c.MyOAuthenticator.client_id = 'your-client-id'
c.MyOAuthenticator.client_secret = 'your-client-secret'
- Requires:
PyJWT>=1.5.3
> pip3 install PyJWT
- BE SURE TO SET THE
AAD_TENANT_ID
environment variable
> export AAD_TENANT_ID='{AAD-TENANT-ID}'
- Sample code is provided for you in
examples > azuread > sample_jupyter_config.py
- Just add the code below to your
jupyterhub_config.py
file - Making sure to replace the values in
'{}'
with your APP, TENANT, DOMAIN, etc. values
Follow this link to create an AAD APP
CLIENT_ID === Azure
Application ID
- found inAzure portal --> AD --> App Registrations --> App
TENANT_ID === Azure
Directory ID
- found inAzure portal --> AD --> Properties
jupyterhub_config.py:
import os
from oauthenticator.azuread import AzureAdOAuthenticator
c.JupyterHub.authenticator_class = AzureAdOAuthenticator
c.Application.log_level = 'DEBUG'
c.AzureAdOAuthenticator.tenant_id = os.environ.get('AAD_TENANT_ID')
c.AzureAdOAuthenticator.oauth_callback_url = 'http://{your-domain}/hub/oauth_callback'
c.AzureAdOAuthenticator.client_id = '{AAD-APP-CLIENT-ID}'
c.AzureAdOAuthenticator.client_secret = '{AAD-APP-CLIENT-SECRET}'
sudo jupyterhub -f ./path/to/jupyterhub_config.py
See
run.sh
for an example
First, you'll need to create a GitHub OAuth application.
Then, add the following to your jupyterhub_config.py
file:
from oauthenticator.github import GitHubOAuthenticator
c.JupyterHub.authenticator_class = GitHubOAuthenticator
You can also use LocalGitHubOAuthenticator
to map GitHub accounts onto local users.
You can use your own Github Enterprise instance by setting the GITHUB_HOST
environment variable.
You can set GITHUB_HTTP
environment variable to true or anything if your GitHub Enterprise supports http only.
GitHub allows expanded capabilities by adding GitHub-Specific Scopes to the requested token.
First, you'll need to create a GitLab OAuth application.
Then, add the following to your jupyterhub_config.py
file:
from oauthenticator.gitlab import GitLabOAuthenticator
c.JupyterHub.authenticator_class = GitLabOAuthenticator
You can also use LocalGitLabOAuthenticator
to map GitLab accounts onto local users.
You can use your own GitLab CE/EE instance by setting the GITLAB_HOST
environment
flag.
Visit https://console.developers.google.com to set up an OAuth client ID and secret. See Google's documentation on how to create OAUth 2.0 client credentials. The Authorized JavaScript origins
should be set to to your hub's public address while Authorized redirect URIs
should be set to the same but followed by /hub/oauth_callback
.
Then, add the following to your jupyterhub_config.py
file:
from oauthenticator.google import GoogleOAuthenticator
c.JupyterHub.authenticator_class = GoogleOAuthenticator
By default, any domain is allowed to login but you can restrict authorized domains with a list (recommended):
c.GoogleOAuthenticator.hosted_domain = ['mycollege.edu', 'mycompany.com']
You can customize the sign in button text (optional):
c.GoogleOAuthenticator.login_service = 'My College'
In case you have an OpenShift deployment with OAuth properly configured (see the
following sections for a quick reference), you should set the client ID and
secret by the environment variables OAUTH_CLIENT_ID
, OAUTH_CLIENT_SECRET
and
OAUTH_CALLBACK_URL
. The OpenShift API URL can be specified by setting the
variable OPENSHIFT_URL
.
The OAUTH_CALLBACK_URL
should match http[s]://[your-app-route]/hub/oauth_callback
As a cluster admin, you can create a global OAuth client in your OpenShift cluster creating a new OAuthClient object using the API:
$ oc create -f - <<EOF
apiVersion: v1
kind: OAuthClient
metadata:
name: <OAUTH_CLIENT_ID>
redirectURIs:
- <OUAUTH_CALLBACK_URL>
secret: <OAUTH_SECRET>
EOF
As a project member, you can use the Service Accounts as OAuth Clients scenario. This gives you the possibility of defining clients associated with service accounts. You just need to create the service account with the proper annotations:
$ oc create -f - <<EOF
apiVersion: v1
kind: ServiceAccount
metadata:
name: <name>
annotations:
serviceaccounts.openshift.io/oauth-redirecturi.1: '<OUAUTH_CALLBACK_URL>'
EOF
In this scenario your OAUTH_CLIENT_ID
will be system:serviceaccount:<serviceaccount_namespace>:<serviceaccount_name>
,
the OAUTH_CLIENT_SECRET is the API token of the service account (oc sa get-token <serviceaccount_name>
)
and the OAUTH_CALLBACK_URL is the value of the annotation serviceaccounts.openshift.io/oauth-redirecturi.1
.
More details can be found in the upstream documentation.
Okpy is an auto-grading tool that
is widely used in UC Berkeley EECS and Data Science courses. This authenticator
enhances its support for Jupyter Notebook by enabling students to authenticate with
the Hub first and saving relevant user states
to the env
(the feature is redacted until a secure state saving mechanism is developed).
If you want to authenticate your Hub using OkpyAuthenticator, you need to specify
the authenticator class in your jupyterhub_config.py
file:
from oauthenticator.okpy import OkpyOAuthenticator
c.JupyterHub.authenticator_class = OkpyOAuthenticator
and set your OAUTH_
environment variables.
Visit https://developers.globus.org/ to set up your app. Ensure Native App is unchecked and make sure the callback URL looks like:
https://[your-host]/hub/oauth_callback
Set scopes for authorization and transfer. The defaults include:
openid profile urn:globus:auth:scope:transfer.api.globus.org:all
Set the above settings in your jupyterhub_config
:
# Tell JupyterHub to create system accounts
from oauthenticator.globus import LocalGlobusOAuthenticator
c.JupyterHub.authenticator_class = LocalGlobusOAuthenticator
c.LocalGlobusOAuthenticator.enable_auth_state = True
c.LocalGlobusOAuthenticator.oauth_callback_url = 'https://[your-host]/hub/oauth_callback'
c.LocalGlobusOAuthenticator.client_id = '[your app client id]'
c.LocalGlobusOAuthenticator.client_secret = '[your app client secret]'
Alternatively you can set env variables for the following: OAUTH_CALLBACK_URL
, OAUTH_CLIENT_ID
,
and OAUTH_CLIENT_SECRET
. Setting JUPYTERHUB_CRYPT_KEY
is required, and can be generated
with OpenSSL: openssl rand -hex 32
You are all set by this point! Be sure to check below for tweaking settings related to User Identity, Transfer, and additional security.
By default, all users are restricted to their Globus IDs (example@globusid.org) with the default Jupyterhub config:
c.GlobusOAuthenticator.identity_provider = 'globusid.org'
If you want to use a Linked Identity such as malcolm@universityofindependence.edu
,
go to your App Developer page and set
Required Identity Provider for your app to <Your University>
, and set the
following in the config:
c.GlobusOAuthenticator.identity_provider = 'universityofindependence.edu'
The default configuration will automatically setup user environments with tokens,
allowing them to start up python notebooks and initiate Globus Transfers. If you
want to transfer data onto your JupyterHub server, it's suggested you install
Globus Connect Server and add the globus_local_endpoint
uuid below. If you want
to change other behavior, you can modify the defaults below:
# Allow Refresh Tokens in user notebooks. Disallow these for increased security,
# allow them for better usability.
c.LocalGlobusOAuthenticator.allow_refresh_tokens = True
# Default scopes are below if unspecified. Add a custom transfer server if you have one.
c.LocalGlobusOAuthenticator.scope = ['openid', 'profile', 'urn:globus:auth:scope:transfer.api.globus.org:all']
# Default tokens excluded from being passed into the spawner environment
c.LocalGlobusOAuthenticator.exclude_tokens = ['auth.globus.org']
# If the JupyterHub server is an endpoint, for convenience the endpoint id can be
# set here. It will show up in the notebook kernel for all users as 'GLOBUS_LOCAL_ENDPOINT'.
c.LocalGlobusOAuthenticator.globus_local_endpoint = '<Your Local JupyterHub UUID>'
# Set a custom logout URL for your identity provider
c.LocalGlobusOAuthenticator.logout_redirect_url = 'https://auth.globus.org/v2/web/logout'
# For added security, revoke all service tokens when users logout. (Note: users must start
# a new server to get fresh tokens, logging out does not shut it down by default)
c.LocalGlobusOAuthenticator.revoke_tokens_on_logout = False
If you only want to authenticate users with their Globus IDs but don't want to
allow them to do transfers, you can remove urn:globus:auth:scope:transfer.api.globus.org:all
.
Conversely, you can add an additional scope for another transfer server if you wish.
Use c.GlobusOAuthenticator.exclude
to prevent tokens from being passed into a
users environment. By default, auth.globus.org
is excluded but transfer.api.globus.org
is allowed. If you want to disable transfers, modify c.GlobusOAuthenticator.scope
instead of c.GlobusOAuthenticator.exclude
to avoid procuring unnecessary tokens.
First install the OAuth2 Server Plugin for Moodle.
Use the GenericOAuthenticator
for Jupyterhub by editing your jupyterhub_config.py
accordingly:
from oauthenticator.generic import GenericOAuthenticator
c.JupyterHub.authenticator_class = GenericOAuthenticator
c.GenericOAuthenticator.oauth_callback_url = 'http://YOUR-JUPYTERHUB.com/hub/oauth_callback'
c.GenericOAuthenticator.client_id = 'MOODLE-CLIENT-ID'
c.GenericOAuthenticator.client_secret = 'MOODLE-CLIENT-SECRET-KEY'
c.GenericOAuthenticator.login_service = 'NAME-OF-SERVICE'
c.GenericOAuthenticator.userdata_url = 'http://YOUR-MOODLE-DOMAIN.com/local/oauth/user_info.php'
c.GenericOAuthenticator.token_url = 'http://YOUR-MOODLE-DOMAIN.com/local/oauth/token.php'
c.GenericOAuthenticator.userdata_method = 'POST'
c.GenericOAuthenticator.extra_params = {
'scope': 'user_info',
'client_id': 'MOODLE-CLIENT-ID',
'client_secret': 'MOODLE-CLIENT-SECRET-KEY'}
And set your environmental variable OAUTH2_AUTHORIZE_URL
to:
http://YOUR-MOODLE-DOMAIN.com/local/oauth/login.php?client_id=MOODLE-CLIENT-ID&response_type=code