fuku

Fuku is a (young) system to help manage web application infrastructure using AWS. It's aim is to replicate some of the simplicity of Heroku without the cost.

Requirements:

Please install the following requirements prior to installing Fuku:

• Python 3.6

• ssh

• ssh-agent

• gpg

• psql

• awscli

Installation

Standard pip install:

pip install fuku

Quickstart for new configurations

1. Configure your AWS credentials as usual (aws configure).

2. fuku profile ls to see available profiles.

   [optional] If you need multiple AWS profiles, create a new profile by editing the file ~/.aws/credentials and copying the default lines.

   fuku profile sl <profile> to select the active AWS profile for Fuku.

3. fuku profile bucket <bucket> to set a workspace for Fuku.

4. fuku region ls to see available regions.

   fuku region sl <region> to select a region.

5. fuku cluster mk <name> to create a cluster.

Quickstart for existing configurations

1. Configure your AWS credentials as usual (aws configure).

2. fuku profile ls to see available profiles.

   fuku profile sl <name> to select your profile.

3. fuku profile bucket <bucket> to set a workspace for Fuku.

4. fuku region ls to see available regions.

   fuku region sl <region> to select a region.

5. fuku cluster ls to see available clusters.

   fuku cluster sl <cluster> to select a cluster.

   Enter password to access cluster key file.

6. fuku app ls to see available apps.

   fuku app sl <app> to select an app.

7. fuku pg ls to see available DBs.

   fuku pg sl <db> to select a DB.

   Enter password to access PGPASS file.

At this point you have configured your session for a particular cluster, application, and database. For convenience it's best to cache the session for easy retrieval:

fuku session sv <name>

fuku session ld <name>

Downloading current DB

To download the current database:

fuku pg dump <dumpfile> <dump_path>

Upload DB

To overwrite a database with new content (CAUTION):

fuku pg restore <filename>

SSH into a node

To access one of the nodes in the cluster directly:

fuku node ssh <name>

Run an arbitrary command

To run a command:

fuku service run <task> <command>

This attaches to a running container from the specified task, then runs the provided command.

Logging

To control logs printed use the flag --log, it uses the available logging levels (CRITICAL, WARNING, INFO, DEBUG)

fuku --log=DEBUG <command>

By default the logs are set to WARNING.

Running fuku in Sub-processes

The default behaviour is to assume that one user is on a single app and/or DB instance at all time.

However if you need to spawn multiple processes running commands on different app and/or DB instance, you can use the --app or --db flags.

For example, we can run in parallel:

fuku --app=first_app service wait bg; fuku --app=first_app service run bg "./manage.py migrate";

fuku --app=second_app service wait bg; fuku --app=second_app service run bg "./manage.py migrate";