Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
A Wukong plugin that allows you to dynamically update database schema using your pre-defined models.
Ruby
Branch: master

Fetching latest commit…

Cannot retrieve the latest commit at this time

Failed to load latest commit information.
bin
lib
spec
.gitignore
Gemfile
README.md
Rakefile
wukong-migrate.gemspec

README.md

Wukong Migrate

A Wukong plugin that allows you to update database schema using predefined models from your deploy pack and a migration DSL.

Commands

The following commands are available

Generate

Creates a new migration file templated to match the chosen database. This command will create a new ruby file in db/migrate named after the argument supplied to generate. The only currently supported database is Elasticsearch, with future plans to support Hbase and others.

$ bundle exec wu-migrate generate change_schema --db=elasticsearch
# INFO 2013-07-29 14:14:51 [MigrateRunner       ] -- Creating migration: db/migrate/change_schema.rb

Perform

Performs the specified migration. This will talk to the database directly and apply the changes found in the migration. The tool will only perform migrations found in db/migrate, and specifying the .rb extension is not necessary. Currently, this is NOT an idempotent operation, and makes no guarantee of data safety on multiple invocations. Configuration is derived from command line parameters and through Wukong::Deploy settings.

$ bundle exec wu-migrate perform change_schema --db=elasticsearch
# INFO 2013-07-29 14:18:20 [MigrateRunner       ] -- Creating index jedi
# INFO 2013-07-29 14:18:21 [MigrateRunner       ] -- Operation complete
# INFO 2013-07-29 14:18:21 [MigrateRunner       ] -- Add alias :light_side for index jedi
# INFO 2013-07-29 14:18:21 [MigrateRunner       ] -- Operation complete

All

Performs all available migrations in db/migrate. This is useful when setting up mirrored or development databases from existing migrations.

Syntax

The following syntax is used to define migrations.

Elasticsearch

All definitions take place inside of a .define block.

EsMigration.define 'name_of_migration' do
...
end

Top-level methods are create_index, update_index, and delete_index. They can be used with or without block syntax.

EsMigration.define 'name_of_migration' do
  create_index(:index_name) do
  ...
  end
  delete_index(:old_index)         
end

Inside of an index block, you have access to mappings, aliases and index-level settings. Aliases are created/deleted one at a time and optionally accept filters. Mapping methods accept blocks for object-level settings.

EsMigration.define 'name_of_migration' do
  create_index(:index_name) do
    number_of_replicas 4
    ...
    alias_to           :other_name, filter: { range: { date: { gt: '2013-05-05' } } }
    remove_alias       :prior_name
    create_mapping(:obj_type) do
      dynamic true
      source  false
      ...
    end               
  end
  delete_index(:old_index)         
end
Something went wrong with that request. Please try again.