SquareSurf InfluxDB Cookbook

Chef cookbook for installing and configuring
InfluxDB. This cookbook took inspiration from the
influxdb cookbook. The
main difference is that this cookbook tries to take the approach that
InfluxDB should be always running once it is installed.
Whereas my understanding of the influxdb
cookbook is that it tries
to seperate configuration from starting the service into two processes.

Requirements

• Chef version 0.11 or higher
• Ruby 1.9 (preferably from the Chef full-stack installer)

Tested Platforms

• Ubuntu 14.04

Note: I'm open to more platforms, this is just what I started with in order to
fill my current need.

Resources

squaresurf_influxdb_user

Use the squaresurf_influxdb_user resource to create/update/delete an
InfluxDB database user.

Actions

• :create (Default)
• :delete

Attributes

• username - Name for the user. (Defaults to the name of the block)
• password - Password for the user.
• database - The database the user will have access to.
• admin - A bool whether or not the user is a database admin. (default: false)
• read_from - A regex to match the series the user can read from. (default: '.*')
• write_to - A regex to match the series the user can write to. (default: '.*')

Examples

Create a user

squaresurf_influxdb_user 'john' do
  action :create
  password 'super secret'
  database 'metrics'
  admin true
  read_from '/^public$/'
  write_to '/^john_workspace$/'
end

Delete a user

squaresurf_infludb_user 'old_user' do
  action :delete
  database 'sensitive data'
end

squaresurf_influxdb_cluster_admin

Use the squaresurf_influxdb_cluster_admin resource to create/update/delete
an InfluxDB cluster admin.

Actions

• :create (Default)
• :delete

Attributes

• username - Name for the cluster admin. (Defaults to the name of the block)
• password - Password for the cluster admin.

Examples

Create a cluster admin

squaresurf_influxdb_cluster_admin 'jane' do
  action :create
  password 'super secret again'
end

Delete a cluster admin

squaresurf_infludb_cluster_admin 'old_cluster_admin' do
  action :delete
end

squaresurf_influxdb_database

Use the squaresurf_influxdb_database resource to create/update/delete an
InfluxDB database.

Actions

• :create (Default)
• :delete

Attributes

• database - Database name. (Defaults to the name of the block)
• options - An extra hash of options to send to the create api.

Examples

Create a database

squaresurf_influxdb_database 'metrics'

Create a database with a 30 day retention policy

squaresurf_influxdb_database 'metrics' do
  options spaces: [
    {
      name: 'default',
      retentionPolicy: '30d',
      shardDuration: '7d',
      regEx: '/.*/',
      replicationFactor: 1,
      split: 1
    }
  ]
end

Delete a database

squaresurf_infludb_database 'old_project_name' do
  action :delete
end

Recipes

squaresurf_influxdb::default

This recipe includes the squaresurf_influxdb::install recipe then the
squaresurf_influxdb::configure recipe.

squaresurf_influxdb::install

This recipe will download the appropriate package according to the version
attribute and install it.

squaresurf_influxdb::configure

This recipe will:
* Setup the influxdb service to be enabled and start.
* Write out to toml the config and benchmark_config from the node attributes.
* Update the main cluster admin password if it has changed.
* Create an alternate cluster admin user if the main admin username is
different than root
* Delete the root user if the main cluster admin username is different than root

squaresurf_influxdb::test

This is used by our chef spec and kitchen test suites.

squaresurf_influxdb::test_exceptions

This is used by our chef spec test suite.

Attributes

General Attributes

Attributes that will affect general cookbook usage.

default.squaresurf_influxdb.fail_on_error = true

If set to false then failures will not fail the chef run and will log errors
instead.

Install attributes

Attributes that affect the installation of InfluxDB.

default.squaresurf_influxdb.version = :latest

This will decide which version of InfluxDB to download and install. It can
either be the version string or the ruby symbol :latest in order to install
the latest available.

default.squaresurf_influxdb.update_version = false

This will decide whether or not to install a new version of influxdb if it is
already installed and the version attribute above differs from that of the
installed version.

InfluxDB Client Attributes

Attributes that affect the ability for this cookbook to connect to InfluxDB in
order to configure users, cluster admins, and databases.

default.squaresurf_influxdb.client_retries = 10

This is how many times we should try to connect to InfluxDB before giving up.
Set this to nil or '-1' to retry indefinitely.

default.squaresurf_influxdb.client_hosts = ['localhost']

This is the host we should connect to InfluxDB with.

default.squaresurf_influxdb.client_use_ssl = false

Whether or not we should connect via ssl with our client when configuring
users, cluster admins, and databases.

default.squaresurf_influxdb.admin_username = 'root'

The cluster admin username to use when connecting to InfluxDB.

default.squaresurf_influxdb.admin_password = 'root'

The cluster admin password to use when connecting to InfluxDB.

default.squaresurf_influxdb.admin_old_username = 'root'

The previous cluster admin username for use if you ever change the username to
another value. This is so that we can create a new cluster admin and delete
this old cluster admin.

default.squaresurf_influxdb.admin_old_password = 'root'

The previous cluster admin password for use if you ever change the password to
another value. This is so that we can set the new password.

InfluxDB TOML Config Attributes

default.squaresurf_influxdb.config

This should be a ruby hash that will be converted to toml for the InfluxDB
config.toml. The default was generated from the default config.toml that comes
with InfluxDB.

default.squaresurf_influxdb.benchmark_config

This should be a ruby hash that will be converted to toml for the InfluxDB
benchmark_config.toml. The default was generated from the default
benchmark_config.toml that comes with InfluxDB.

Usage

The one main attribute that you should always change is the
node.squaresurf_influxdb.admin_password. That way you won't have a default
cluster admin password that anyone can connect to your server with.

squaresurf_influxdb::default

Either include the recipe in your run_list or include_recipe from within
one of your recipes. For example you could do the following to setup influxdb
and create a database named metrics and a user named sensu.

# Make sure to set the admin password to some secret value. It would probably
# be even better to use an encrypted data bag here as shown in the next example.
node.set.squaresurf_influxdb.admin_password = 'super secret'

include_recipe 'squaresurf_influxdb::default'

squaresurf_influxdb_database 'metrics'

squaresurf_influxdb_user 'sensu' do
    password 'some secret value'
    database 'metrics'
end

squaresurf_influxdb::install and squaresurf_influxdb::configure

Instead of using the default recipe you can seperate your logic into an install
then a configure in case you need to setup some items vefore the service is
started. For example you could set up ssl configuration for the http api like so:

cert_file = '/opt/influxdb/ssl_cert.pem'
node.set.squaresurf_influxdb.config.api['ssl-cert'] = cert_file
node.set.squaresurf_influxdb.config.api['ssl-port'] = 8084

cert = Chef::EncryptedDataBagItem.load('ssl', 'influxdb')

admin = Chef::EncryptedDataBagItem.load('influxdb', 'admin')
node.default.squaresurf_influxdb.admin_password = admin['password']
node.default.squaresurf_influxdb.admin_old_password = admin['old_password']

# install first so that the /opt/influxdb dir and influxdb user will be created
include_recipe 'squaresurf_influxdb::install'

file cert_file do
  content cert['cert'] + "\n" + cert['key']
  user 'influxdb'
  group 'influxdb'
  mode '0700'
end

include_recipe 'squaresurf_influxdb::configure'

Contributing

1. Fork the repository on Github
2. Create a named feature branch (like add_component_x)
3. Write your change
4. Write tests for your change
5. Run the tests, ensuring they all pass
6. Submit a Pull Request using Github

Testing

There is a Rakefile with tasks to make testing easy. You will probably need to
bundle install before running any of them.

• rake test:all # Run all tests
• rake test:chefspec # Run chefspec
• rake test:foodcritic # Run foodcritic linter against cookbook
• rake test:kitchen_test # This is here as a convenience so that the test suite will check kitchen as well as the other tests
• rake test:rubocop # Run rubocop against cookbook ruby files

config and benchmark_config attribute generation

There is a convenience script toml_to_attr.rb to help generate the config and
benchmark_config ruby hashes from a default toml file included with influxdb.

If you haven't run bundle install for the test suite above you'll probably
need to do so in order to run toml_to_attr.rb.

License and Authors

License

The License is the MIT License and can be found in the LICENSE file.

Authors

• Daniel Paul Searles (squaresurf)