Enable specific hub features

There are several extra features and services that can be enabled on the 2i2c JupyterHubs. These add extra functionality, connect with cloud services, etc. See the sections below for more details.

• Allow users to setup custom, free-form user profile choices
  • Allow members of a particular github team only to test custom images
• Restrict profile options based on JupyterHub groups (or GitHub teams)
  • The allowed_groups configuration
  • Enabling externally managed groups for GitHubOAuthenticator
  • Enabling this feature for other Authenticators
• Anonymize usernames with CILogon
  • Do you really need this?
  • How does it work?
  • Limitations
  • Authorization
  • Enabling anonymization
  • Longer term solution
  • Credit
• Make binderhub-ui hub
  • Generate a sample initial hub configuration
  • Double-check the generated config
  • Manually handle registry creation and login
• Setup object storage buckets
  • Testing access to buckets
  • Allowing public, readonly to buckets from outside the JupyterHub
  • Enable access logs for objects in a bucket
  • Allowing authenticated access to buckets from outside the JupyterHub
  • Granting access to cloud buckets in other cloud accounts / projects
• Enable user access to cloud features
  • How it works
  • Enabling specific cloud access permissions
• Enable stronger anti-crypto abuse features for a hub
  • Testing the execwhacker detector
  • Looking at logs to understand why a process was killed by execwhacker
  • Regenerating list of banned strings
  • Working on the banned strings generator
• Add support for daskhubs
  • To an existing hub
  • To an existing cluster
  • GCP
• Setup a dedicated nodepool for a hub on a shared cluster
  • Node type and minimum nodepool size considerations
• Make an ephemeral hub
  • Authentication with tmpauthenticator
  • No persistent home directory
  • (Optional) Sharing shared directories from another hub with an ephemeral hub
  • Image configuration in chart
  • Enable hook pre-puller & disable JupyterHub
  • Disabling home page customizations
  • Use nbgitpuller for distributing content
• Allow users to push to GitHub
• Pre-pull a content repository
  • Defining the init container
• Enable access to GPUs
  • Setting up GPU nodes
  • Setting up a GPU user profile
  • Testing
• Using JupyterHub as an identity provider
  • 1. Request required information
  • 2. Configure JupyterHub
  • 3. Respond with setup information
• Enable user image features
  • Mount object storage as files with FUSE
• Make an imagebuilding hub
  • Connect with jupyterhub-fancy-profiles
  • Setup the image registry
  • Setup the binderhub-service chart
• Configure the hub login page
  • Redirect the logged out page to a different URL
• Setup a database server per user
  • Setting up the backing disk for the database
  • Setup the database server
  • Cleanup created disks after use
• Allow nbgitpuller to pull from private GitHub repos
  • Configure the image
  • Setup GitHub app
  • Helm values configuration
  • Grant access to the private repo
  • Create the nbgitpuller link
• Use Rocker based R images
  • Why?
  • Picking the image to use
  • Hub configuration
  • Configuration without profileList
  • Configuration with profileList
  • User installed packages
• Setup a shared database for all users on the hub
  • AWS
• Configure per-user storage quotas
  • Creating a pre-provisioned disk
  • Enabling jupyterhub-home-nfs
  • Migrating existing home directories and switching to the new NFS server
  • Enforcing storage quotas
  • Increasing the size of the disk used by the NFS server
  • Troubleshooting
• Integrate community-documentation with Jupyter Book
  • Create the community documentation repo
  • Use a custom subdomain for the book
  • Add Binder/JupyterHub configuration to the book