- Overview
- Module Description - What the module does and why it is useful
- Setup - The basics of getting started with r10k
- Usage - Configuration options and additional functionality
- Reference - An under-the-hood peek at what the module is doing and how
- Limitations - OS compatibility, etc.
- Development - Guide for contributing to the module
This module was built to install and configure r10k. It has a base class to configure r10k to synchronize dynamic environments. It also has a series of lateral scripts and tools that assist in general workflow, that will be seperated into their own modules into the future.
This module is meant to manage the installation and configuration of r10k using multiple installation methods on multiple platforms.
Please refer to the official r10k docs for specific configuration patterns.
Instead of passing a single remote
, you can pass a puppet hash as the sources
parameter. This allows you to configure r10k with prefix support. This often used when multiple teams use seperate repos, or if hiera and puppet are distributed across two repos.
class { 'r10k':
sources => {
'webteam' => {
'remote' => 'ssh://[email protected]/webteam/somerepo.git',
'basedir' => "${::settings::confdir}/environments",
'prefix' => true,
},
'secteam' => {
'remote' => 'ssh://[email protected]/secteam/someotherrepo.git',
'basedir' => '/some/other/basedir',
'prefix' => true,
},
},
}
- Installation of the r10k
gem
- Installation of git
- Installation of ruby when not using an existing ruby stack i.e. when using
puppet_gem
- Management of the
r10k.yaml
in /etc - Installation and configuration of a sinatra app when using the webhook.
Gem installation is pinned to a default version in this module, the following chart shows the gem installation tested with the respective module version.
You can override this by passing the version
parameter.
Module Version | r10k Version |
---|---|
v4.0.0+ | |
v3.0.x | 1.5.1 |
v2.8.2 | 1.5.1 |
v2.7.x | 1.5.1 |
v2.6.5 | 1.4.1 |
v2.5.4 | 1.4.0 |
v2.4.4 | 1.3.5 |
v2.3.1 | 1.3.4 |
v2.3.0 | 1.3.2 |
v2.2.8 | 1.3.1 |
v2.2.x | 1.1.0 |
r10k connects via ssh and does so silently in the background, this typically requires ssh keys to be deployed in advance of configuring r10k. This includes the known host ( public ) key of the respective git server, and the user running r10k's private key used to authenticate git/ssh during background runs.
Here is an example of deploying the ssh keys needed for r10k to connect to a repo called puppet/control on a gitlab server. This is helpful when you need to automatically deploy new masters
#https://docs.puppetlabs.com/references/latest/type.html#sshkey
sshkey { 'your.internal.gitlab.server.com':
ensure => present,
type => 'ssh-rsa',
target => '/root/.ssh/known_hosts',
key => '...+dffsfHQ==',
}
# Resource git_webhook is provided by https://forge.puppet.com/abrader/gms
git_deploy_key { 'add_deploy_key_to_puppet_control':
ensure => present,
name => $::fqdn,
path => '/root/.ssh/id_dsa.pub',
token => hiera('gitlab_api_token'),
project_name => 'puppet/control',
server_url => 'http://your.internal.gitlab.server.com',
provider => 'gitlab',
}
A simple example of creating an ssh private key would use an exec to call yes y | ssh-keygen -t dsa -C "r10k" -f /root/.ssh/id_dsa -q -N ''
.
The example above shows using git_deploy_key
which would deploy that key to the remote git server via its api. This is often required in the programtic creation of compile masters.
Given r10k will likely be downloading your modules, often on the first server
it's run on, you will have to puppet apply
this module to bootstrap this
configuration and allow for ongoing management from there.
The simplest example of using it would be to declare a single remote that would be written to r10k.yaml.
class { 'r10k':
remote => '[email protected]:someuser/puppet.git',
}
This will configure /etc/r10k.yaml
and install the r10k gem after installing
ruby using the puppetlabs/ruby module.
It also supports installation via multiple providers, such as installation in the puppet_enterprise ruby stack in versions less than 3.8
Installing into the Puppet Enterprise ruby stack in PE 2015.x
class { 'r10k':
remote => '[email protected]:someuser/puppet.git',
provider => 'puppet_gem',
}
Note: It is recommended you migrate to using the pe_r10k
module which is basically
a clone of this modules features and file tickets for anything missing.
Installing using a proxy server
An mcollective agent is included in this module which can be used to do on demand synchronization. This mcollective application and agent can be installed on all masters using the following class Note: You must have mcollective already configured for this tool to work, Puppet Enterprise users will automatically have mcollective configured. This class does not restart the mcollective or pe-mcollective server on the nodes to which it is applied, so you may need to restart mcollective for it to see the newly installed r10k agent.
include r10k::mcollective
Using mco you can then trigger mcollective to call r10k using
mco r10k synchronize
You can sync an individual environment using:
mco r10k deploy <environment>
Note: This implies -p
You can sync an individual module using:
mco r10k deploy_module <module>
If you are required to run r10k
as a specific user, you can do so by passing
the user
parameter:
mco r10k synchronize user=r10k
To obtain the output of running the shell command, run the agent like this:
mco rpc r10k synchronize -v
An example post-receive hook is included in the files directory. This hook can automatically cause code to synchronize on your servers at time of push in git. More modern git systems use webhooks, for those see below.
Install the mco
command from the puppet enterprise installation directory i.e.
cd ~/puppet-enterprise-3.0.1-el-6-x86_64/packages/el-6-x86_64
sudo rpm -i pe-mcollective-client-2.2.4-2.pe.el6.noarch.rpm
Copy the peadmin mcollective configuration and private keys from the certificate authority (puppet master)
/var/lib/peadmin/.mcollective
/var/lib/peadmin/.mcollective.d/mcollective-public.pem
/var/lib/peadmin/.mcollective.d/peadmin-cacert.pem
/var/lib/peadmin/.mcollective.d/peadmin-cert.pem
/var/lib/peadmin/.mcollective.d/peadmin-private.pem
/var/lib/peadmin/.mcollective.d/peadmin-public.pem
Ensure you update the paths in ~/.mcollective when copying to new users whose name is not peadmin. Ideally mcollective will be used with more then just the peadmin user's certificate in the future. That said, if your git user does not have a home directory, you can rename .mcollective as /etc/client.cfg and copy the certs to somewhere that is readable by the respective user.
/home/gitolite/.mcollective
/home/gitolite/.mcollective.d/mcollective-public.pem
/home/gitolite/.mcollective.d/peadmin-cacert.pem
/home/gitolite/.mcollective.d/peadmin-cert.pem
/home/gitolite/.mcollective.d/peadmin-private.pem
/home/gitolite/.mcollective.d/peadmin-public.pem
Note: PE2 only requires the .mcollective file as the default auth was psk
class { 'r10k::mcollective':
ensure => false,
}
This will remove the mcollective agent/application and ddl files from disk. This likely would be if you are migrating to Code manager in Puppet Enterprise.
For version control systems that use web driven post-receive processes you can use the example webhook included in this module.
When the webhook receives the post-receive event, it will synchronize environments on your puppet masters.
The webhook uses mcollective for multi-master synchronization and the peadmin
user from Puppet Enterprise by default.
These settings are all configurable for your specific use case, as shown below in these configuration examples.
This is an example of using the webhook without authentication.
The git_webhook
type will use the api token to add the webhook to the "control" repo that contains your puppetfile. This is typically useful when you want to automate the addition of the webhook to the repo.
# Required unless you disable mcollective
include r10k::mcollective
# Internal webhooks often don't need authentication and ssl
# Change the url below if this is changed
class {'r10k::webhook::config':
enable_ssl => false,
protected => false,
}
class {'r10k::webhook':
require => Class['r10k::webhook::config'],
}
# Add webhook to control repository ( the repo where the Puppetfile lives )
#
# Resource git_webhook is provided by https://forge.puppet.com/abrader/gms
git_webhook { 'web_post_receive_webhook' :
ensure => present,
webhook_url => 'http://master.of.masters:8088/payload',
token => hiera('github_api_token'),
project_name => 'organization/control',
server_url => 'https://your.github.enterprise.com',
provider => 'github',
}
# Add webhook to module repo if we are tracking branch in Puppetfile i.e.
# mod 'module_name',
# :git => 'http://github.com/organization/puppet-module_name',
# :branch => 'master'
# The module name is determined from the repo name , i.e. <puppet-><module_name>
# All characters with left and including any hyphen are removed i.e. <puppet->
#
# Resource git_webhook is provided by https://forge.puppet.com/abrader/gms
git_webhook { 'web_post_receive_webhook_for_module' :
ensure => present,
webhook_url => 'http://master.of.masters:8088/module',
token => hiera('github_api_token'),
project_name => 'organization/puppet-module_name',
server_url => 'https://your.github.enterprise.com',
provider => 'github',
}
This is an example of using the webhook with authentication.
The git_webhook
type will use the api token to add the webhook to the "control" repo that contains your puppetfile. This is typically useful when you want to automate the addition of the webhook to the repo.
# Required unless you disable mcollective
include r10k::mcollective
# External webhooks often need authentication and ssl and authentication
# Change the url below if this is changed
class {'r10k::webhook::config':
enable_ssl => true,
protected => true,
notify => Service['webhook'],
}
class {'r10k::webhook':
require => Class['r10k::webhook::config'],
}
# Add webhook to control repository ( the repo where the Puppetfile lives )
#
# Resource git_webhook is provided by https://forge.puppet.com/abrader/gms
git_webhook { 'web_post_receive_webhook' :
ensure => present,
webhook_url => 'https://puppet:[email protected]:8088/payload',
token => hiera('github_api_token'),
project_name => 'organization/control',
server_url => 'https://api.github.com',
disable_ssl_verify => true,
provider => 'github',
}
# Add webhook to module repo if we are tracking branch in Puppetfile i.e.
# mod 'module_name',
# :git => 'http://github.com/organization/puppet-module_name',
# :branch => 'master'
# The module name is determined from the repo name , i.e. <puppet-><module_name>
# All characters with left and including any hyphen are removed i.e. <puppet->
#
# Resource git_webhook is provided by https://forge.puppet.com/abrader/gms
git_webhook { 'web_post_receive_webhook_for_module' :
ensure => present,
webhook_url => 'https://puppet:[email protected]:8088/module',
token => hiera('github_api_token'),
project_name => 'organization/puppet-module_name',
server_url => 'https://api.github.com',
disable_ssl_verify => true,
provider => 'github',
}
GitHub webhooks allow the use of a secret value that gets hashed against the payload to pass a signature in the request X-Hub-Signature header. To support the secret with the webhook do the following type of configuration.
class { 'r10k::webhook::config':
protected => false,
github_secret => 'THISISTHEGITHUBWEBHOOKSECRET',
}
class { 'r10k::webhook':
require => Class['r10k::webhook::config'],
}
For use when moving to Code Manager, or other solutions, and the webhook should be removed.
class {'r10k::webhook::config':
ensure => false,
}
class {'r10k::webhook':
ensure => false,
}
If you have only a single master, you may want to have the webhook run r10k directly rather then as peadmin via mcollective.
This requires you to run as the user that can perform r10k
commands which is typically root.
The peadmin certificate no longer is managed or required.
# Instead of running via mco, run r10k directly
class {'r10k::webhook::config':
use_mcollective => false,
}
# The hook needs to run as root when not running using mcollective
# It will issue r10k deploy environment <branch_from_gitlab_payload> -p
# When git pushes happen.
class {'r10k::webhook':
use_mcollective => false,
user => 'root',
group => '0',
require => Class['r10k::webhook::config'],
}
The following is an example of declaring the webhook when r10k prefixing are enabled.
This script is fed the github/gitlab payload in via stdin. This script is meant to return the prefix as its output. This is needed as the payload does not contain the r10k prefix. The simplest way to determine the prefix is to use the remote url in the payload and find the respective key in r10k.yaml. An example prefix command is located in this repo here. Note that you may need to modify this script depending on your remote configuration to use one of the various remote styles.
file {'/usr/local/bin/prefix_command.rb':
ensure => file,
mode => '0755',
owner => 'root',
group => '0',
source => 'puppet:///modules/r10k/prefix_command.rb',
}
# Required unless you disable mcollective
include r10k::mcollective
class {'r10k::webhook::config':
prefix => true,
prefix_command => '/usr/local/bin/prefix_command.rb',
require => File['/usr/local/bin/prefix_command.rb'],
}
class {'r10k::webhook':
require => Class['r10k::webhook::config'],
}
# Deploy this webhook to your local gitlab server for the puppet/control repo.
#
# Resource git_webhook is provided by https://forge.puppet.com/abrader/gms
git_webhook { 'web_post_receive_webhook' :
ensure => present,
webhook_url => 'https://puppet:[email protected]:8088/payload',
token => hiera('gitlab_api_token'),
project_name => 'puppet/control',
server_url => 'http://your.internal.gitlab.com',
provider => 'gitlab',
}
Currently the webhook relies on existing certificates for its ssl configuration. See this following ticket for more information. Until then, its possible to re-use you existing FOSS mcollective certificates.
Here is an working example on Ubuntu 14.04.2 LTS
class { 'r10k::webhook::config':
public_key_path => '/etc/mcollective/server_public.pem', # Mandatory for FOSS
private_key_path => '/etc/mcollective/server_private.pem', # Mandatory for FOSS
}
class { 'r10k::webhook':
user => 'puppet', # Mandatory for FOSS
group => 'puppet', # Mandatory for FOSS
}
Some instances may require the user/group root/root
(Linux) or root/wheel
(BSD).
Verify that the specified user has the permissions to run r10k
commands.
class { 'r10k::webhook::config':
use_mcollective => false,
public_key_path => '/etc/mcollective/server_public.pem', # Mandatory even when use_mcollective is false
private_key_path => '/etc/mcollective/server_private.pem', # Mandatory even when use_mcollective is false
}
class { 'r10k::webhook':
user => 'root', # Mandatory for FOSS
group => 'root', # Mandatory for FOSS
}
By default, the r10k::webhook::package
class uses the puppet_gem
provider to install the latest ruby 2.1 compatible version of sinatra ('~> 1.0').
If you are overriding r10k::webhook::package::provider
, you will also need to override r10k::webhook::package::sinatra_version
.
You can enable Slack notifications for the webhook. You will need a
Slack webhook URL and the slack-notifier
gem installed.
To get the Slack webhook URL you need to:
- Go to https://slack.com/apps/A0F7XDUAZ-incoming-webhooks.
- Choose your team, press
Configure
. - In configurations press
Add configuration
. - Choose channel, press
Add Incoming WebHooks integration
.
Then configure the webhook to add your Slack Webhook URL.
class { 'r10k::webhook::config':
. . .
slack_webhook => 'http://slack.webhook/webhook', # mandatory for usage
slack_channel => '#channel', # defaults to #default
slack_username => 'r10k', # the username to use
slack_proxy_url => 'http://proxy.example.com:3128' # Optional. Defaults to undef.
}
The default branch of the controlrepo is commonly called production
. This value can be overriden if you use another default branch name, such as master
.
class { 'r10k::webhook::config':
default_branch => 'master', # Optional. Defaults to 'production'
}
To aid in debugging, or to give you some hints as to how to trigger the webhook by unsupported systems, here's a curl command to trigger the webhook to deploy the 'production' environment:
curl -d '
{
"repository": {"name": "foo", "owner": {"login": "foo"}},
"ref": "production"
}' http://puppet-master.example:8088/payload
If you're not sure whether your webhook setup works:
- Try to make a GET request to the
heartbeat
endpoint (e.g. http://puppet-master.example:8088/heartbeat). You should see a short JSON answer similar to{"status":"success","message":"running"}
. - Watch the webhook logfile at
/var/log/webhook/access.log
, and send requests (e.g. using curl). Example output if successful:
$ tail -f /var/log/webhook/access.log
...
[2015-05-03 13:01:45] DEBUG Rack::Handler::WEBrick is mounted on /.
[2015-05-03 13:01:45] INFO WEBrick::HTTPServer#start: pid=7185 port=8088
[2015-05-03 13:02:46] DEBUG accept: 192.30.252.46:39379
[2015-05-03 13:02:46] DEBUG Rack::Handler::WEBrick is invoked.
[2015-05-03 13:02:46] INFO authenticated: puppet
[2015-05-03 13:02:46] INFO message: triggered: r10k deploy environment test_webhook -pv ...
...
If you are building your image with the puppet, you need to prevent the webhook process from starting as a daemon.
The following is an example of declaring the webhook without a background mode
class { 'r10k::webhook':
. . .
background => false
}
If you need to configure webhook to not trigger r10k when changes pushed in some branch or repository, you can list them in
r10k::webhook::config::ignore_environments
parameter as array.
There is an ability to specify it as a regular expression by enclosing it in forward slashes.
Here is an example where the test branch in dev repository and all branches in all repositories that includes the word 'feature' in names will be skipped:
class { 'r10k::webhook::config':
. . .
prefix => ':repo',
ignore_environments => ['dev_test', '/.*feature.*/']
}
You can pass some additional arguments to mco command like --no-progress
or --timeout 60
or any others by specifying them
in the r10k::webhook::config::mco_arguments
parameter as string:
class { 'r10k::webhook::config':
. . .
mco_arguments => '--no-progress'
}
This is the main public class to be declared , handingly installation and configuration declarations
Parameters within r10k
:
A string to be passed in as the source with a hardcode prefix of puppet
A hash of all sources, this gets read out into the file as yaml. Must not be declared with remote
A single string setting the r10k.yaml
configuration value of the same name
A path to the configuration file to manage. Be aware Puppet Enterprise 4.0 and higher may conflict if you manage /etc/puppetlabs/puppet/r10k.yaml
A value passed to the package resource for managing the gem version
Deprecated: for older configfile environments configuration of modulepath in puppet.conf
Deprecated: declare a resource for managing modulepath
in Puppet.conf
When using system ruby , options on how to declare
This module requires the puppetlabs-ruby module. In the event that your environment already includes
the module with some customization, you can use the manage_ruby_dependency
parameter to adjust how this module expresses that requirement.
The supported values are include
,declare
, or ignore
. The values' behavior
is outlined below:
-
declare default This will explicitly declare the ruby module. Additional declarations of the ruby module will result in an inability to compile a catalog.
-
include This will simply include the ruby module. When combined with class ordering, this will permit the user to manage the instantiation of the ruby module elsewhere, potentially with non-standard parameter values.
-
ignore This will assume that ruby is handled via some other mechanism than a puppet module named
ruby
. It is left to the user to insure the requirement be met.
The name of the package to be installed via the provider
The supported installation modes for this module
- portage
- yum
- bundle
- puppet_gem
- gem
- zypper
See r10k::install::portage
class for more information
Options to pass to the provider
declaration
Install mcollective application and agents. This does NOT configure mcollective automatically
Manage a symlink to the configuration file, for systems installed in weird file system configurations
This is the git:
key in r10k, it accepts a hash that can be used to configure
rugged support.
$git_settings = {
'provider' => 'rugged',
'private_key' => '/root/.ssh/id_rsa',
}
class {'r10k':
remote => '[email protected]:acidprime/puppet.git',
git_settings => $git_settings,
}
This is the forge:
key in r10k, it accepts a hash that contains settings for downloading modules from the Puppet Forge.
$forge_settings = {
'proxy' => 'https://proxy.example.com:3128',
'baseurl' => 'https://forgeapi.puppetlabs.com',
}
class {'r10k':
remote => '[email protected]:acidprime/puppet.git',
forge_settings => $forge_settings,
}
This is the deploy:
key in r10k, it accepts a hash that contains setting that control how r10k code deployments behave. Documentation for the settings can be found here.
$deploy_settings = {
'purge_levels' => ['puppetfile'],
}
class {'r10k':
remote => '[email protected]:voxpupuli/puppet.git',
deploy_settings => $deploy_settings,
}
boolean if to manage symlink
Deprecated: Add prerun_command to puppet.conf to run r10k when the agent on the master runs.
Suggest instead declaring r10k::postrun_command
as that will run after the agent runs which prevents r10k from stopping configuration management of masters from occuring as it does with prerun_command
s
r10k::include_postrun_command: true
The concept here is that this is declared on the puppet master(s) that have been configured with r10k. This will cause r10k to synchronize after each puppet run. Any errors synchronizing will be logged to the standard puppet run.
The 4.1.x release deprecates support for:
- Puppet 3
- Ruby 1.9.3
These items are are planned for removal in v5.0.0.
Please log tickets and issues at our Projects site
Modules on the Puppet Forge are open projects, and community contributions are essential for keeping them great. We can’t access the huge number of platforms and myriad of hardware, software, and deployment configurations that Puppet is intended to serve.
We want to keep it as easy as possible to contribute changes so that our modules work in your environment. There are a few guidelines that we need contributors to follow so that we can have a chance of keeping on top of things.
Please see CONTRIBUTING for more details.
This project contains tests for rspec-puppet to verify functionality. For in-depth information please see their respective documentation, as well as CONTRIBUTING.
Quickstart:
gem install bundler
bundle install --without system_tests
bundle exec rake test
bundle exec rake lint
Check the .travis.yml for supported Operating System Versions