capistrano_multiconfig_parallel

Gem Version Repo Size Gem Downloads Analytics Join the chat at https://gitter.im/bogdanRada/capistrano_multiconfig_parallel

DEMO

capistrano multiconfig parallel

Description

CapistranoMulticonfigParallel is a simple ruby implementation that allows you to run multiple tasks in parallel for multiple applications and uses websockets for inter-process communication and has a interactive menu

IMPORTANT! The whole reason for this gem was for using Caphub in a more easy way and allowing you to run tasks in parallel for multiple aplications . However this can be used for normal applications also, if you want for example to deploy your app to multiple sandboxes on development environment or even deploy in parallel to multiple stages.

CAUTION!! PLEASE READ CAREFULLY!! Capistrano is not thread-safe. However in order to work around this problem, each of the task is executing inside a thread that spawns a new process in order to run capistrano tasks The thread monitors the process. This works well, however if the tasks you are executing is working with files, you might get into deadlocks because multiple proceses try to access same resource. Instead of using files , please consider using StringIO instead.

NEW Improvements in version 2.3.0

  • added more validations and refactored the stages helper in order to properly detect independent deploys

NEW Improvements in version 2.2.0

  • the websocket server will automatically detect an unused port and use that, in case port is not set

NEW Improvements in version 2.1.0

  • some internal refactoring and updates for activesupport

NEW Improvements in version 2.0.0

  • Code for handling websocket events when a task is invoked was moved to a new gem capistrano_sentinel

  • You can now deploy applications from anywhere on your computer without having to add THIS gem to the Gemfiles of those applications, however we recommend you to add the capistrano_sentinel gem to those Gemfiles for better stability. The applications though must contain proper configuration for capistrano as you would normally do!!!

  • If capistrano_sentinel is included in one of the applications you are deploying, but you forget to require it, this gem will automatically patch the Capfile to require the gem. Otherwise the Capfile is not changed.

  • However this will work even if you don't add the capistrano_sentinel to your Gemfile, because the new version will automatically create three files "Gemfile.multi_cap", "Gemfile.multi_cap.lock" and "Capfile.multi_cap" where the dependency to capistrano_sentinel will be added and those new files will be used when deploying. This new Gemfile and the new Capfile will be clones of the original files, but will only add as dependency the capistrano_sentinel gem which is used to produce websocket events before a capistrano task is executed. And this gem will consume them and act depending on configuration.

  • After installing this gem locally on your system, you can just create anywhere on your computer a YAML file as described in section 2.2) Deploying multiple applications from a central location ( we suggest using the home directory) and then you can start deploying applications without adding new dependencies to those applications.

  • When the command will finish , the files created will be removed. ("Gemfile.multi_cap", "Gemfile.multi_cap.lock" and "Capfile.multi_cap") if they exist in the application root directory.

  • If you are using RVM and there is a .rvmrc file in your project root directory and bash is available , the script will use bash emulator in order to properly load RVM gemsets because .rvmrc files need trusting. If you use .rvmrc files please make sure you have the bash executable available in your PATH . Please consider switching to .ruby-version and .ruby-gemset files instead which are faster and don't require trusting. You can read more about it here RVM Workflow

  • Dont use BUNDLE EXEC command when running the executable for this gem unless you are using version 1 or this gem is part of the Gemfile of the application you are using. Otherwise just use the multi_cap executable without any prefix ( Unless this gem is part of your gemfile, in which case the BUNDLE EXEC prefix is needed )

If you are using a older version than 2.0 please refer to this README for version 1.7.2 or lower

Requirements

  1. celluloid >= 0.16
  2. celluloid-pmap >= 0.2.2
  3. celluloid_pubsub >= 0.8.4
  4. composable_state_machine >= 1.0.2
  5. terminal-table >= 1.5.2
  6. colorize >= 0.7
  7. eventmachine >= 1.0.3
  8. right_popen >= 1.1.3
  9. ActiveSuport >= 4.0
  10. configliere >= 0.4
  11. inquirer >= 0.2
  12. devnull >= 0.1
  13. capistrano_sentinel >= 0.0.17
  14. powerbar >= 1.0.17

Compatibility

MRI >= 2.0

Ruby 1.8, 1.9 are not officially supported.

Rubinius and Jruby support dropped.

Installation Instructions

Add the following to your Gemfile after requiring capistrano and capistrano-multiconfig:

  gem "capistrano_sentinel"

Add the following to your Capfile after requiring capistrano and capistrano-multiconfig

  require 'capistrano_sentinel'

Install locally on your system the capistrano_multiconfig_parallel gem using this command :

Please check first what is the last release here: Release Details and use that version instead of the one below

  gem install capistrano_multiconfig_parallel -v 2.1.1

You will need to have MRI >= 2.2 when you install in this way due to activesupport dependency.

Or if you need to use MRI > 2.0 and < 2.2 you can use this command:

  gem install capistrano_multiconfig_parallel # This will fail due to activesupport dependency, but next command will fix it
  gem install capistrano_multiconfig_parallel:2.1.1 activesupport:4.0.0

You can replace the activesupport dependency with any version greater than 4.0 . I would suggest to use Ruby > 2.2.2 though.

For MRI > 2.0 and < 2.2 you can use though this workaround. Create a Gemfile on your sytem and add this line of code:


 gem 'capistrano_multiconfig_parallel'

 if RUBY_VERSION < "2.2.0"
   gem 'activesupport', '>= 4.0', '< 5.0'
 end

And use bundle exec multi_cap from that directory when running the executable.

For Ruby >= 2.2 please use the first scenario and don't use BUNDLE EXEC in that case!!!

Please read Release Details if you are upgrading. We break backward compatibility between large ticks but you can expect it to be specified at release notes.

Default Configuration:

---
multi_debug: true
multi_secvential: false
check_app_bundler_dependencies: false
websocket_server:
  enable_debug: false
  adapter: ''
  log_file_path: './log/multi_cap_websocket.log'
terminal:
  clear_screen: false  

development_stages:
  - development
  - webdev

apply_stage_confirmation:
  - production  
task_confirmations:
  - deploy:symlink:release

application_dependencies: []
---

Available command line options when executing a command

  • --multi-debug

    • if option is present and has value TRUE , will enable debugging of workers
  • --multi-secvential

    • If parallel executing does not work for you, you can use this option so that each process is executed normally and ouputted to the screen. However this means that all other tasks will have to wait for each other to finish before starting
  • --check_app_bundler_dependencies

    • if option is present and has value TRUE , will check the application bundler dependecies are satisfied before delegating the job to the actor that executes the capistrano action
  • --websocket_server.enable_debug

    • if option is present and has value TRUE, will enable debugging of websocket communication between the workers
  • --websocket_server.adapter

    • Can be specified a different adapter for the websocket reactor class used for publish subscribe communication
  • --websocket_server.log_file_path

    • Enables the logging of websocket communication into a different file -
  • --terminal.clear_screen

    • Enables the clear screen to happen before the table status is displayed on screen
  • --config_dir

    • The directory where the config file (multi_cap.yml) is located. The gem will try to detect if file exist in current directory from where the script is runnning
  • --log_dir

    • The path where the log files will be created. By default it depends on the config_dir. It will create a log directory in the same directory as the configuration directory
  • --development_stages

    • if option is present and has value an ARRAY of STRINGS, each of them will be used as a development stage
  • --apply_stage_confirmation

    • If option is present and has value an ARRAY of STRING, each string should be the name of a stage. The confirmations will only be applied if the workers are executing tasks to one of that stage.
  • --task_confirmations:

    • if option is present and has value an ARRAY of Strings (Task Names), then will require a confirmation from user before executing the task. This will syncronize all workers to wait before executing that task, then a confirmation will be displayed, and when user will confirm , all workers will resume their operation.
  • --application_dependencies

    • This should be used only for Caphub-like applications , in order to deploy dependencies of an application in parallel and is described at section 2.) Multiple applications

Usage Instructions

In order to override default configurations, you can either specify them at runtime using command-line , or you can create a file in location config/multi_cap.yml . This settings will be overriden by command line arguments( if any)

1) Single Apps ( normal Rails or rack applications)

CapistranoMulticonfigParallel recognizes only "development" and "webdev" as stages for development if you use other stages for development, you need to configure it like this

---
development_stages:
  - development
  - webdev
  - something_custom
----

1.1) Deploying the application to multiple sandboxes ( works only with development environments)

# <box_name>     - the name of a sandbox
#<development_stage> - the name of one of the stages you previously configured
#<task_name> - the capistrano task that you want to execute ( example: 'deploy' )

multi_cap  <development_stage> <task_name>   BOX=<box_name>,<box_name>  

For Capistrano 2 application, the jobs will receive -S box=,, although you will use BOX when running the command. This so that i can maintain compatibility between versions

The script will ask if there are any other environment variables that user might want to pass to each of the sandboxes separately.

2.1) Deploying the application to multiple stages ( Using the customized command "deploy_multi_stages")


 multi_cap deploy_multi_stages  STAGES=development, staging, production

NOTE: IF you want to execute a different command on all stages, you can specify environment variable ACTION=task_name either when you specify the STAGES, or can be done individually for each task when prompted about additional ENV options

The script will ask if there are any other environment variables that user might want to pass to each of the stages separately.

If you use capistrano-gitflow, the workers will first deploy to all the other stages and only after staging is tagged , will trigger a new worker to start deploying to production

2.) Multiple Apps ( like Caphub )

Configuration for this types of application is more complicated

---
application_dependencies:
    - app: foo'
      priority: 1
      dependencies: []
    - app: bar
      priority: 1
      dependencies:
        - foo
    - app: foo2
      priority: 1
      dependencies:
        - foo
        - bar
---

The "development_stages" options is used so that the gem can know if sandboxes are allowed for those environments.

The dependencies are being kept in the option "application_dependencies" This is an array of hashes. Each hash has only the keys "app" ( app name), "priority" and "dependencies" ( an array of app names that this app is dependent to)

In this example, if we execute this command:

 multi_cap foo2:development deploy

Will ask user if he wants to deploy the apps "foo" and "bar" , since they appear in the dependencies list for the application "foo2"

2.1) Deploying multiple application to multiple stages ( Using the customized command "deploy_multi_stages")


 multi_cap deploy_multi_stages  STAGES=development, staging, production

Demo:

capistrano multiconfig parallel

NOTE: IF you want to execute a different command on all stages, you can specify environment variable ACTION=task_name either when you specify the STAGES, or can be done individually for each task when prompted about additional ENV options

The script will ask for additional ENV options for each stage.

If you use capistrano-gitflow, the workers will first deploy to all the other stages and only after staging is tagged , will trigger a new worker to start deploying to production

2.2) Deploying multiple applications from a central location

This works like described in section 2) Multiple applications ), but now the application_dependencies Array can also have a path key inside each item that should point to the DocumentRoot of that application, and the configuration now accepts two new arguments config_dir ( this should point to the directory where the file multi_cap.yml resides) and log_dir( this can be used if you want your logs created in different directory)

This will only work if all applications listed in the configuration file have the gem capistrano_sentinel as part of their Gemfile.

Example of configuration: - create a directory anywhere (e.g.**mkdir /tmp/app**) - create a Gemfile and add only this two lines:

source 'http://rubygems.org'

gem 'capistrano_multiconfig_parallel'

or you could install it directly from command line:

gem install capistrano_multiconfig_parallel

Then create a file called multi_cap.yml in the root folder and create a configuration like this.

   ---
   application_dependencies:
       - app: foo'
         priority: 1
         path: /some/path/to/foo
         dependencies: []
       - app: bar
         priority: 1
         path: /some/path/to/bar
         dependencies:
           - foo
       - app: foo2
         priority: 1
         path: /some/path/to/foo2
         dependencies:
           - foo
           - bar
   ---

and then you can run like you would normally do: (Apppend bundle exec to the below command only if you have a Gemfile)

multi_cap foo2:development deploy

Known Limitations

  • Currently it works only if Celluloid.cores >= 2
  • RVM users that use .rvmrc files need to have bash terminal emulator available

Testing

To test, do the following:

  1. cd to the gem root.
  2. bundle install
  3. bundle exec rake

Contributions

Please log all feedback/issues via Github Issues. Thanks.

Contributing to capistrano_multiconfig_parallel

  • Check out the latest master to make sure the feature hasn't been implemented or the bug hasn't been fixed yet.
  • Check out the issue tracker to make sure someone already hasn't requested it and/or contributed it.
  • Fork the project.
  • Start a feature/bugfix branch.
  • Commit and push until you are happy with your contribution.
  • Make sure to add tests for it. This is important so I don't break it in a future version unintentionally.
  • Please try not to mess with the Rakefile, version, or history. If you want to have your own version, or is otherwise necessary, that is fine, but please isolate to its own commit so I can cherry-pick around it.
  • You can read more details about contributing in the Contributing document

== Copyright

Copyright (c) 2015 bogdanRada. See LICENSE.txt for further details.