byu-awslogin 0.15.2

An aws-adfs spinoff that fits BYU's needs

AWSLOGIN

Python script for CLI and SDK access to AWS via ADFS while requiring MFA access using https://duo.com/

History and Purpose

BYU used to use the great aws-adfs CLI tool to login to our AWS accounts. It worked great, especially the DUO 2FA support. Eventually, we decided to write our own similar tool but make it BYU-specific so that we could tailor it to our needs (which basically means hard-code certain BYU-specific things) and remove some of the required parameters. Since this tool will be used by BYU employees only we had that option. We then morphed it a little more for our use cases. This isn't something that you could use outside of BYU, sorry.

DUO 2FA Requirements

In order for Duo 2FA to work properly Automatic Push needs to be enabled.

Installation

• Python 3.6+ is recommended as python2 is EOL January 2020.
• It is highly recommended to use an application like Pipx to install and use python cli applications.
• Follow the pipx installation documentation then simply run pipx install byu_awslogin
• Experimental Binaries are available on the releases page. These are new and in testing Releases
• See the installation options For additional options page for step by step instructions for installing in various environments

Upgrading

If you already have byu_awslogin install and are looking to upgrade simply run

pip3 install --upgrade byu_awslogin or pip install --upgrade byu_awslogin as appropriate for your python installation

Usage

awslogin defaults to the default profile in your ~/.aws/config and ~/.aws/credentials files. *If you already have a default profile you want to save in your ~/.aws files make sure to do that before running awslogin.*

Once you're logged in, you can execute commands using the AWS CLI or AWS SDK. Try running aws s3 ls.

Currently, AWS temporary credentials are only valid for 1 hour. We cache your ADFS session, however, so you can just re-run awslogin again to get a new set of AWS credentials without logging in again to ADFS. Your ADFS login session is valid for 8 hours, after which time you'll be required to login to ADFS again to obtain a new session.

To switch accounts after you've already authenticated to an account, just run awslogin again and select a new account/role combination.

To use it:

• Run awslogin and it will prompt you for the AWS account and role to use.
• Run awslogin --account <account name> --role <role name> to skip the prompting for account and name. You could specify just one of the arguments as well.
• Run awslogin --profile <profile name> to specifiy an alternative profile
• Run awslogin --region <region name> to specify a different region. The default region is us-west-2.
• Run awslogin --status for the current status of the default profile
• Run awslogin --status --profile dev for the current status of the dev profile
• Run awslogin --status --profile all for the current status of the all profiles
• Run awslogin --logout to logout of a cached ADFS session
• Run awslogin --version to display the running version of awslogin
• Run awslogin --help for full help message

Bash or ZSH Completion

Bash:

• Run the following: _AWSLOGIN_COMPLETE=source awslogin > ~/_awslogin Then add source /path/to/_awslogin to .bashrc

ZSH:

• Run the following: _AWSLOGIN_COMPLETE=source_zsh awslogin > ~/_awslogin Then add source /path/to/_awslogin to .zshrc

Alternatively put the _awslogin in your /etc/bash_completion.d or similiar directory (~/.zfunc) to load at shells startup

To test if it works run awslogin at least once for the account and role cache to populate. On next login awslogin -a [TAB][TAB] should output available accounts and awslogin -a {some account} -r [TAB][TAB] should output available roles for the selected account

Limitation: Accounts and Role completion at the CLI is loaded from a cache file. This file will be updated anytime awslogin is ran.

Reporting bugs or requesting features

• Enter an issue on the github repo.
• Or, even better if you can, fix the issue and make a pull request.

Deploying changes

• Update the version in pyproject.toml and __version__.py
• Commit the changes and push.
• Build binaries
• Create a new release (add binaries and sha256sums.txt) with the version number and Github Actions will build, test and publish

TODO

• Write tests
  • Write more tests to increase overall coverage