For political reasons, I've been asked to use Capistrano to deploy Chef
Solo to a list of servers and interact with an API.
I'm new to Cap and new(ish) to Ruby. I'm looking for some help re: best way
to use server definitions in Cap 3.
>From Cap 3 docs
<http://capistranorb.com/documentation/getting-started/preparing-your-application/>
:
# Extended Server Syntax
# ======================
# This can be used to drop a more detailed server
# definition into the server list. The second argument
# is something that quacks like a hash and can be used
# to set extended properties on the server.
server 'example.com', roles: %w{web app}, my_property: :my_value
vs
# Simple Role Syntax
# ==================
# Supports bulk-adding hosts to roles, the primary
# server in each group is considered to be the first
# unless any hosts have the primary property set.
role :app, %w{example.com}
role :web, %w{example.com}
role :db, %w{example.com}
Which was the de facto standard in Cap 2.
My server list looks like this:
server '10.0.1.1', roles: [:hdp_ambari, :hdp_namenode, :hdp_hbase, :
hdp_monitor_server, :hdp_zookeeper, :hdp_ganglia], internal:
'ip-10-0-1-1.ec2.internal'
server '10.0.1.2', roles: [:hdp_namenode_secondary, :hdp_datanode, :
hdp_zookeeper, :hdp_ganglia], internal:'ip-10-0-1-2.ec2.internal'
I want to set :options to a hash of keys and internals corresponding to a
given role. What's the best way to do this?
So desired output something like:
set :options, {
default_filesystem: "hdfs://ip-10-0-1-1.ec2.internal:8020",
hbase_master: 'ip-10-0-1-2.ec2.internal',
job_history_server: 'ip-10-0-1-1.ec2.internal',
namenode: 'ip-10-0-1-1.ec2.internal',
namenode_secondary: 'ip-10-0-1-2.ec2.internal',
yarn_node: 'ip-10-0-1-3.ec2.internal',
zookeepers: ['ip-10-0-1-1.ec2.internal', 'ip-10-0-1-2.ec2.internal']
}
For which I was thinking...
task :set_configuration_options do
# Fetching custom server.properties via http://tinyurl.com/ldkm7ov
on roles(:hdp_namenode), in: :parallel do |server|
namenode_internal = p server.internal
filesystem_internal = "hdfs://#{namenode_internal}:8020"
end
options = {
default_filesystem: "#{filesystem_internal}",
namenode: "#{namenode_internal}"
}
end
set :options, options
But there's got to be a cleaner, better way, right? I'm having a tough time
finding good examples or docs on this.
*Relevant details:*
Versions:
- Ruby - 2.0.0p481
- Capistrano - 3.0.1
- Rake / Rails / etc - 10.1.0, 0.9.6
Platform:
- Working on.... Mac OSX 10.9.2
- Deploying to... CentOS 6.4
Logs:
- Not relevant. Asking for general advice, best practices re: extended
server syntax.
Files:
- Capfile, deploy.rb, and stage attached.
--
You received this message because you are subscribed to the Google Groups
"Capistrano" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to capistrano+unsubscr...@googlegroups.com.
To view this discussion on the web, visit
https://groups.google.com/d/msgid/capistrano/703ba8c0-e2ca-44fe-bfa3-1161abeeb118%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
set :blueprint_name, fetch(:stage).split(':').last
set :cluster_name, "HDP_#{fetch(:stage).split(':').last}"
# internal: is a custom server property we need to set configuration options in
Ambari.
server '10.0.1.1', roles: [:hdp_ambari, :hdp_namenode, :hdp_hbase,
:hdp_monitor_server, :hdp_zookeeper, :hdp_ganglia],
internal:'ip-10-0-1-47.ec2.internal'
server '10.0.1.2', roles: [:hdp_namenode_secondary, :hdp_datanode,
:hdp_zookeeper, :hdp_ganglia], internal:'ip-10-0-1-48.ec2.internal'
server '10.0.1.3', roles: [:hdp_datanode, :hdp_yarn, :hdp_job_history_server,
:hdp_zookeeper, :hdp_ganglia], internal:'ip-10-0-1-49.ec2.internal'
server '10.0.0.4', roles: [:hdp_client, :hdp_zookeeper_client, :hdp_ganglia],
internal:'ip-10-0-0-30.ec2.internal'
# Ambari API calls ONLY use the internal IP. Everything else can use Amazon's
internal DNS.
set :base_uri, '10.0.1.1:8080'
set :username, 'admin'
set :password, 'admin'
task :set_configuration_options do
# Build array (see staging and prod) that matches up roles with the internal
IP. See ambari_api.rb.
# Fetching custom server.properties via http://tinyurl.com/ldkm7ov
on roles(:hdp_namenode), in: :parallel do |server|
namenode_internal = p server.internal
filesystem_internal = "hdfs://#{namenode_internal}:8020"
end
on roles(:hdp_namenode_secondary), in: :parallel do |server|
namenode_secondary_internal = p server.internal
end
on roles(:hdp_hbase), in: :parallel do |server|
hbase_internal = p server.internal
end
on roles(:hdp_job_history_server), in: :parallel do |server|
job_history_internal = p server.internal
end
on roles(:hdp_yarn), in: :parallel do |server|
yarn_internal = p server.internal
end
on roles(:hdp_zookeeper), in: :parallel do |server|
zookeeper_internal = p server.internal # needs to return array for
multiple internals
end
configuration_options = {
default_filesystem: "#{filesystem_internal}",
hbase_master: "#{hbase_internal}",
job_history_server: "#{job_history_internal}",
namenode: "#{namenode_internal}",
namenode_secondary: "#{namenode_secondary_internal}",
yarn_node: "#{yarn_internal}",
zookeepers: "#{zookeeper_internal}"
}
#Note: Could also throw base_uri in here, then set it below.
end
set :configuration_options, configuration_options
components = {
hdp_namenode: %w[
NAMENODE
HDFS_CLIENT
YARN_CLIENT
TEZ_CLIENT
],
hdp_namenode_secondary: %w[
SECONDARY_NAMENODE
HDFS_CLIENT
YARN_CLIENT
TEZ_CLIENT
],
hdp_yarn: %w[
APP_TIMELINE_SERVER
],
hdp_datanode: %w[
HBASE_REGIONSERVER
NODEMANAGER
DATANODE
HDFS_CLIENT
YARN_CLIENT
TEZ_CLIENT
],
hdp_hbase: %w[
HBASE_MASTER
],
hdp_zookeeper: %w[
ZOOKEEPER_SERVER
ZOOKEEPER_CLIENT
],
hdp_zookeeper_client: %w[
ZOOKEEPER_CLIENT
],
hdp_ambari: %w[
AMBARI_SERVER
],
hdp_client: %w[
HBASE_CLIENT
TEZ_CLIENT
SQOOP
PIG
HDFS_CLIENT
YARN_CLIENT
MAPREDUCE2_CLIENT
TEZ_CLIENT
],
hdp_job_history_server: %w[
HISTORYSERVER
],
hdp_resource_manager: %w[
RESOURCEMANAGER
],
hdp_monitor_server: %w[
GANGLIA_SERVER
NAGIOS_SERVER
],
hdp_ganglia: %w[
GANGLIA_MONITOR
],
}
host_groups = servers.collect do |server|
{
name: server[:roles].first.to_s,
components: server[:roles].collect { |role| components[role] }.flatten.uniq,
hosts: server[:internal]
}
end
set :host_groups, host_groups#
# Put here shared configuration shared among all children
#
# Read more about configurations:
# https://github.com/railsware/capistrano-multiconfig/blob/master/README.md
require 'aws-sdk'
require 'JSON'
require 'open-uri'
require 'active_support/all'
require 'retryable'
require 'heroku-api'
require 'chef_solo'
set :heroku_api_key, '[redacted]'
set :default_stage, 'development'
ask :branch, ENV['BRANCH'] || 'master'
set :application, proc { fetch(:stage).split(':').reverse[1] }
set :env, proc { fetch(:stage).split(':').last }
set :repo_url, proc { "g...@github.com:[redacted]/#{fetch(:application)}.git" }
set :deploy_to, proc { "/var/www/#{fetch(:application)}" }
set :scm, :git
set :format, :pretty
set :log_level, :debug
set :pty, true
set :linked_files, %w{config/database.yml}
set :linked_dirs, %w{bin log tmp/pids tmp/cache tmp/sockets vendor/bundle
public/system}
set :storm_dir, proc { File.expand_path(File.join(File.dirname(__FILE__), '..',
'bin', 'storm-0.9.0.1')) }
set :default_env, { path: "/opt/ruby/bin:#{fetch(:storm_dir)}/bin:$PATH" }
set :keep_releases, 5
set :topology_bundler_group, 'default'
set :cluster_name, fetch(:topology_name)
set :nimbus_security_group, lambda { "[redacted]" }
set :username, ENV['USER'] || ENV['username']
set :chef_project_location, "/root/chef-repo"
def use_rvm(command, gemfile_path = nil)
command_prefix = gemfile_path ? "cd #{gemfile_path};
BUNDLE_GEMFILE=./Gemfile" : ''
"bash --login --rcfile /usr/local/rvm/scripts/rvm -c '#{command_prefix}
#{command}'"
end
def get_newest_code_for(project_location, git_branch=nil)
git_branch ||= branch
try_sudo "sh -c 'cd #{project_location} && git fetch origin && git reset
--hard HEAD && git checkout #{git_branch.gsub('origin/', '')} && git reset
--hard #{git_branch}'", :as => nil
end
def build_runlist(roles)
role_list = StringIO.new({
"run_list" => [ "role[#{roles.join(',')}]" ]
}.to_json).string
end
namespace :chef_solo do
desc 'Get newest code'
task :update_git do
on server(:all) do |server|
get_newest_code_for(:chef_project_location, fetch(:branch))
end
end
desc 'Upload roles and execute Chef Solo'
task :upload_and_run do
on server(:all) do |server|
role_list = build_runlist(server[:roles])
upload! role_list, "/tmp/node.json"
run "#{sudo} cd #{chef_project_location} && chef-solo -c
/root/chef-repo/client-config/solo.rb -j /tmp/node.json"
end
end
end
require 'bundler'
Bundler.require :default
# Load DSL and Setup multiple configurations
# https://github.com/railsware/capistrano-multiconfig
require 'capistrano/multiconfig'
# Includes default deployment tasks
require 'capistrano/framework'
# Includes tasks from other gems included in your Gemfile
#
# For documentation on these, see for example:
#
# https://github.com/capistrano/rvm
# https://github.com/capistrano/rbenv
# https://github.com/capistrano/chruby
# https://github.com/capistrano/bundler
# https://github.com/capistrano/rails/tree/master/assets
# https://github.com/capistrano/rails/tree/master/migrations
# https://github.com/railsware/capistrano-uptodate
# https://github.com/railsware/capistrano-patch
# https://github.com/railsware/capistrano-calendar
#
# require 'capistrano/rvm'
# require 'capistrano/rbenv'
# require 'capistrano/chruby'
# require 'capistrano/bundler'
# require 'capistrano/rails/assets'
# require 'capistrano/rails/migrations'
# require 'capistrano/uptodate'
# require 'capistrano/patch'
# require 'capistrano/calendar'
$LOAD_PATH.unshift File.join(File.dirname(__FILE__), 'lib')
# Loads custom tasks
Dir.glob('tasks/*.cap').each { |r| import r }
# vim syntax=ruby
