puppetlabs / pupperware

Container fun time lives here.
Other
183 stars 67 forks source link
pupperware

pupperware

Run a container-based deployment of Puppet Infrastructure.

To get started, you will need an installation of Docker Compose on the host on which you will run your Puppet Infrastructure.

Running Puppet Infrastructure in Kubernetes is also a very viable option. To get started with that, you will need a running K8s cluster with Helm deployed.

We've been developing our own Helm chart which can get you up & running fast. You can find it here. It's hosted as a Helm chart here and published in the fantastic Helm Hub and Artifact Hub. The latter will allow you to make use of it by just adding the repo in your configured Helm repos.

Generally, containers included here follow Dockerfile best practices.

Required versions

Provisioning

Once you have Docker Compose installed, you can start the stack on Linux or OSX with:

    export ADDITIONAL_COMPOSE_SERVICES_PATH=${PWD}/gem/lib/pupperware/compose-services
    export COMPOSE_FILE=${ADDITIONAL_COMPOSE_SERVICES_PATH}/postgres.yml:${ADDITIONAL_COMPOSE_SERVICES_PATH}/puppetdb.yml:${ADDITIONAL_COMPOSE_SERVICES_PATH}/puppet.yml
    DNS_ALT_NAMES=host.example.com docker-compose up -d

With the environment variables exported, the stack can be torn down with:

    docker-compose down --volumes

The value of DNS_ALT_NAMES must list all the names, as a comma-separated list, under which the Puppet server in the stack can be reached from agents. It will have puppet prepended to it as that name is used by PuppetDB to communicate with the Puppet server. The value of DNS_ALT_NAMES only has an effect the first time you start the stack, as it is placed into the server's SSL certificate. If you need to change it after that, you will need to properly revoke the server's certificate and restart the stack with the changed DNS_ALT_NAMES value.

When you first start the Puppet Infrastructure, the stack will create a number of Docker volumes to store the persistent data that should survive the restart of your infrastructure. The actual location on disk of these volumes may be examined with the docker inspect command. The following volumes include:

Container Versions

By default, the puppetserver and puppetdb containers will use the latest tag. PUPPETSERVER_IMAGE and PUPPETDB_IMAGE environment variables have been added to the compose file to easily select different image repos / pin versions if you need to by setting those on the command line, or in a .env file.

Pupperware on Windows with WSL2 (Unsupported)

Complete instructions for provisiong a server with WSL2 support are in README-windows.md

Creating the stack from PowerShell is nearly identical to other platforms, aside from how environment variables are declared:

PS> $ENV:DNS_ALT_NAMES = 'host.example.com'
PS> $ENV:ADDITIONAL_COMPOSE_SERVICES_PATH="${PWD}/gem/lib/pupperware/compose-services"
PS> $ENV:COMPOSE_FILE="${ENV:ADDITIONAL_COMPOSE_SERVICES_PATH}\postgres.yml;${ENV:ADDITIONAL_COMPOSE_SERVICES_PATH}\puppetdb.yml;${ENV:ADDITIONAL_COMPOSE_SERVICES_PATH}\puppet.yml"

PS> docker-compose up
Creating network "pupperware_default" with the default driver
Creating volume "pupperware_puppetserver-code" with default driver
Creating volume "pupperware_puppetserver-config" with default driver
Creating volume "pupperware_puppetserver-data" with default driver
Creating volume "pupperware_puppetdb-ssl" with default driver
Creating volume "pupperware_puppetdb-postgres" with default driver
Creating pupperware_postgres_1 ...

Creating pupperware_puppet_1   ...

Creating pupperware_puppet_1   ... done

Creating pupperware_postgres_1 ... done

Creating pupperware_puppetdb_1 ...

Creating pupperware_puppetdb_1 ... done

...

To delete the stack:

PS> docker-compose down
Removing network pupperware_default
...

Managing the stack

The script bin/puppet (or bin\puppet.ps1 on Windows) makes it easy to run puppet commands on the puppet master. For example, ./bin/puppet config print autosign --section master prints the current setting for autosigning, which is true by default. In a similar manner, any other task that you would perform on a puppet master by running puppet x y z ... can be achieved against the stack by running ./bin/puppet x y z ....

There is also a similar script providing easy access to puppetserver commands. This is particularly useful for CA and cert management via the ca subcommand.

Changing postgresql password

The postgresql instance uses password authentication for communication with the puppetdb instance. If you need to change the postgresql password, you'll need to do the following:

Running tests

Running tests locally

This repo contains some simple tests that can be run with RSpec. To run these tests you need to have Ruby, Docker, and Docker Compose installed on the machine where you're running the tests. The tests depend on the 'rspec' and 'json' rubygems. The tests are known to run on at least ruby 1.9.3-p551 and as new as ruby 2.4.3p205.

NOTE These tests will start and stop the cluster running from the current checkout of Pupperware, so be careful where you run them from.

To run the tests:

  1. bundle install --with test
  2. bundle exec rspec spec

Containers

The containers used in pupperware are generated based on dockerfiles in the repos for puppetserver and puppetdb. Published containers can be found on dockerhub.

Analytics Data Collection

The Puppet owned containers run in the pupperware stack collect usage data. You can opt out of providing this data.

What data is collected?

Why does pupperware collect data?

We collect data to help us understand how the containers are used and make decisions about upcoming changes.

How can I opt out of pupperware container data collection?

Create a .env file in this directory with the contents:

PUPPERWARE_ANALYTICS_ENABLED=false

This file is in the .gitignore file and will not be managed or changed by pupperware.

License

See LICENSE file.

Issue Tracking

Please report any issues as GitHub issues in this repo.

Contact us!

If you have questions or comments about pupperware, feel free to send a message to the puppet-users mailing list or reach out in the #puppet channel in the puppet community slack.