⚡️ SSH into instances using fuzzy search.
Clone or download
pmazurek Merge pull request #35 from egidijus/documentation-dependency-conflicts
this is a guide on how to have several tools installed with pip witho…
Latest commit 3ec73ea Aug 6, 2018

README.md

Build Status

AWS Fuzzy Finder

aws-fuzzy-finder aims at one thing: making the process of finding the IPs and SSH'ing into your EC2 instances super fast and easy. It will connect with AWS, automatically grab all the instances you have access to, and present them to you in a fuzzy searchable way!

It is built on top of fzf binaries and boto3.

Installation

To install use the following command:

pip install aws-fuzzy-finder

This package uses boto to authenticate, so if you have your aws-cli or ansible configured and working, you can skip the following step, it will work out of the box.

if not, create ~/.aws/credentials file and make it look like this:

[default]
aws_access_key_id = your_key
aws_secret_access_key = your_secret
region = your_region_code

More information on alternative ways of configuring your AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY and AWS_DEFAULT_REGION variables can be found here: http://boto3.readthedocs.io/en/latest/guide/configuration.html

Settings

You will need to set the user you want to SSH with and the path to your ssh key. --ssh-user will default to ec2-user and --key-path will default to ~/.ssh/id_rsa so if you use defaults, you can skip this step.

If you want to use private IP's instead of public ones, use --private flag.

Either use the command line params, or you can append this to your ~/.bashrc to make the settings permamant:

export AWS_FUZZ_USER="your.user"
export AWS_FUZZ_KEY_PATH="~/.ssh/your_private_key"
export AWS_FUZZ_PRIVATE_IP='true' # Delete this one if you want to use public IP's

Remeber that every change to ~/.bashrc requires you to re-load it: source ~/.bashrc or restart terminal.

AWS_FUZZ_SSH_COMMAND_TEMPLATE - set this env var if you want to customize the ssh command , defaults to ssh {key} {user}{host}

Usage

To run, use the following command:

aws-fuzzy

To run using a different AWS profile, run the command as follows:

AWS_DEFAULT_PROFILE=profile_name aws-fuzzy

Enjoy!

Key bindings

It is very convenient to bind various aws-fuzzy profiles/settings to keys. This gives you even faster access to your instances. To achieve this, add this to your ~/.bashrc:

# This will bind the aws-fuzzy command to ctrl+f
bind  '"\C-f": "aws-fuzzy\e\C-e\er\C-m"'

# You can bind different settings to different keys
bind  '"\C-a": "AWS_DEFAULT_PROFILE=production aws-fuzzy --private\e\C-e\er\C-m"' 

Advanced usage

Sometimes you need to use only the IP of the instance. You can use this command to interactively pick IP's to use with other commands. To do so, add --ip-only as a parameter. Example usage:

$ echo "foo $(aws-fuzzy --ip-only) bar"
> foo 10.123.42.12 bar

Example to make ansible interactive:

$ ansible --become --ask-become-pass -v -i "$(aws-fuzzy --ip-only)," all -m shell -a "setenforce 0"

This will bring an interactive prompt, and the IP of the instance of your choice will be used. You can combine it with basically any command you want, sky is the limit now ;)

Cache

If you are managing lots of instances and downloading the data takes too long, you can use the built in cache. To enable it set the following variables in your .bashrc:

export AWS_FUZZ_USE_CACHE=yes
export AWS_FUZZ_CACHE_EXPIRY=3600  # expiry time in seconds

To invalidate cache and refresh data, run with --no-cache param Cache will be stored as a file in ~/.aws_fuzzy_finder.cache.

Tunneling

If you have to access your instances through a gateway instance, use the --tunneling param. This will make the fuzzy find to run twice: first time you will pick the gateway to tunnel through, and the second time you choose is the instance you would like to SSH into.

Gateway must be allowed to access the instance with its own ssh key. You may set the user and key path spearately using --tunnel-user and --tunnel-key-path params. The key will be looked up ON the gateway instace.

Dependency conflicts/ Virtualenvs

It's 2018, Summer and you are using pip to install aws-fuzzy-finder along with awscli on the same machine, but it fails, you see errors about botocore, you become sad. You want to be able to use both tools in a convenient and easy way. You accept that some initial configuration is OK, and a happy life afterwards is great, you are no longer sad.

Firstly we will need, python setup tools and virtualenv as provided by your distribution. (I assume you are using a Linux).

We will modify our pip config, so that packages installed with pip are always installed locally to / for the user, because you probably will not be running aws-fuzzy-finder as a system service.

Create a file in your home >> .config >> pip directory:

~/.config/pip/pip.conf

add this to the file

[install]
user = yes
no-binary = :all:

Then we will create a directory where we will store all the virtual environments for all tools that we might want to install like this.

mkdir -p ~/.virtualenvs

Now let's install awscli:

cd ~/.virtualenvs
virtualenv --system-site-packages -p python3 awscli
. awscli/bin/activate
pip install awscli
deactivate

Now let's install aws-fuzzy-finder:

cd ~/.virtualenvs
virtualenv --system-site-packages -p python3 aws-fuzzy-finder
. aws-fuzzy-finder/bin/activate
pip install aws-fuzzy-finder
deactivate

Now, in your ~/.local/bin/ directory, you should see aws "binary" and a aws-fuzzy "binary". If you head -n1 ~/.local/bin/aws you should see something like:

#!/home/your-user-name/.virtualenvs/awscli/bin/python

This is essentially telling the "binary" to use the virtual environment we specifically created just for this tool.

No more conflicts!!!