Merge pull request #5162 from kennyj/schema_cache_dump

[Proposal] Schema cache dump

activerecord/CHANGELOG.md

@@ -1,5 +1,28 @@
 ## Rails 4.0.0 (unreleased) ##
 
+*   Added the schema cache dump feature.
+
+    `Schema cache dump` feature was implemetend. This feature can dump/load internal state of `SchemaCache` instance
+    because we want to boot rails more quickly when we have many models.
+
+    Usage notes:
+
+      1) execute rake task.
+      RAILS_ENV=production bundle exec rake db:schema:cache:dump
+      => generate db/schema_cache.dump
+
+      2) add config.use_schema_cache_dump = true in config/production.rb. BTW, true is default.
+
+      3) boot rails.
+      RAILS_ENV=production bundle exec rails server
+      => use db/schema_cache.db
+
+      4) If you remove clear dumped cache, execute rake task.
+      RAILS_ENV=production bundle exec rake db:schema:cache:clear
+      => remove db/schema_cache.dump
+
+    *kennyj*
+
 *   Added support for partial indices to PostgreSQL adapter
 
     The `add_index` method now supports a `where` option that receives a

activerecord/lib/active_record/connection_adapters/abstract_adapter.rb

@@ -86,6 +86,11 @@ module ActiveRecord
         end
       end
 
+      def schema_cache=(cache)
+        cache.connection = self
+        @schema_cache = cache
+      end
+
       def expire
         @in_use = false
       end

activerecord/lib/active_record/connection_adapters/schema_cache.rb

@@ -1,26 +1,17 @@
 module ActiveRecord
   module ConnectionAdapters
     class SchemaCache
-      attr_reader :columns, :columns_hash, :primary_keys, :tables
-      attr_reader :connection
+      attr_reader :columns, :columns_hash, :primary_keys, :tables, :version
+      attr_accessor :connection
 
       def initialize(conn)
         @connection = conn
-        @tables     = {}
 
-        @columns = Hash.new do |h, table_name|
-          h[table_name] = conn.columns(table_name)
-        end
-
-        @columns_hash = Hash.new do |h, table_name|
-          h[table_name] = Hash[columns[table_name].map { |col|
-            [col.name, col]
-          }]
-        end
-
-        @primary_keys = Hash.new do |h, table_name|
-          h[table_name] = table_exists?(table_name) ? conn.primary_key(table_name) : nil
-        end
+        @columns      = {}
+        @columns_hash = {}
+        @primary_keys = {}
+        @tables       = {}
+        prepare_default_proc
       end
 
       # A cached lookup for table existence.
@@ -30,12 +21,22 @@ module ActiveRecord
         @tables[name] = connection.table_exists?(name)
       end
 
+      # Add internal cache for table with +table_name+.
+      def add(table_name)
+        if table_exists?(table_name)
+          @primary_keys[table_name]
+          @columns[table_name]
+          @columns_hash[table_name]
+        end
+      end
+
       # Clears out internal caches
       def clear!
         @columns.clear
         @columns_hash.clear
         @primary_keys.clear
         @tables.clear
+        @version = nil
       end
 
       # Clear out internal caches for table with +table_name+.
@@ -45,6 +46,37 @@ module ActiveRecord
         @primary_keys.delete table_name
         @tables.delete table_name
       end
+
+      def marshal_dump
+        # if we get current version during initialization, it happens stack over flow.
+        @version = ActiveRecord::Migrator.current_version
+        [@version] + [:@columns, :@columns_hash, :@primary_keys, :@tables].map do |val|
+          self.instance_variable_get(val).inject({}) { |h, v| h[v[0]] = v[1]; h }
+        end
+      end
+
+      def marshal_load(array)
+        @version, @columns, @columns_hash, @primary_keys, @tables = array
+        prepare_default_proc
+      end
+
+      private
+
+      def prepare_default_proc
+        @columns.default_proc = Proc.new do |h, table_name|
+          h[table_name] = connection.columns(table_name)
+        end
+
+        @columns_hash.default_proc = Proc.new do |h, table_name|
+          h[table_name] = Hash[columns[table_name].map { |col|
+            [col.name, col]
+          }]
+        end
+
+        @primary_keys.default_proc = Proc.new do |h, table_name|
+          h[table_name] = table_exists?(table_name) ? connection.primary_key(table_name) : nil
+        end
+      end
     end
   end
 end

activerecord/lib/active_record/railtie.rb

@@ -107,7 +107,7 @@ module ActiveRecord
       config.watchable_files.concat ["#{app.root}/db/schema.rb", "#{app.root}/db/structure.sql"]
     end
 
-    config.after_initialize do
+    config.after_initialize do |app|
       ActiveSupport.on_load(:active_record) do
         ActiveRecord::Base.instantiate_observers
 
@@ -115,6 +115,21 @@ module ActiveRecord
           ActiveRecord::Base.instantiate_observers
         end
       end
+
+      ActiveSupport.on_load(:active_record) do
+        if app.config.use_schema_cache_dump
+          filename = File.join(app.config.paths["db"].first, "schema_cache.dump")
+          if File.file?(filename)
+            cache = Marshal.load(open(filename, 'rb') { |f| f.read })
+            if cache.version == ActiveRecord::Migrator.current_version
+              ActiveRecord::Base.connection.schema_cache = cache
+            else
+              warn "schema_cache.dump is expired. Current version is #{ActiveRecord::Migrator.current_version}, but cache version is #{cache.version}."
+            end
+          end
+        end
+      end
+
     end
   end
 end

activerecord/lib/active_record/railties/databases.rake

@@ -372,6 +372,25 @@ db_namespace = namespace :db do
     task :load_if_ruby => 'db:create' do
       db_namespace["schema:load"].invoke if ActiveRecord::Base.schema_format == :ruby
     end
+
+    namespace :cache do
+      desc 'Create a db/schema_cache.dump file.'
+      task :dump => :environment do
+        con = ActiveRecord::Base.connection
+        filename = File.join(Rails.application.config.paths["db"].first, "schema_cache.dump")
+
+        con.schema_cache.clear!
+        con.tables.each { |table| con.schema_cache.add(table) }
+        open(filename, 'wb') { |f| f.write(Marshal.dump(con.schema_cache)) }
+      end
+
+      desc 'Clear a db/schema_cache.dump file.'
+      task :clear => :environment do
+        filename = File.join(Rails.application.config.paths["db"].first, "schema_cache.dump")
+        FileUtils.rm(filename) if File.exists?(filename)
+      end
+    end
+
   end
 
   namespace :structure do

activerecord/test/cases/connection_adapters/schema_cache_test.rb

@@ -39,6 +39,21 @@ module ActiveRecord
         assert_equal 0, @cache.tables.size
         assert_equal 0, @cache.primary_keys.size
       end
+
+      def test_dump_and_load
+        @cache.columns['posts']
+        @cache.columns_hash['posts']
+        @cache.tables['posts']
+        @cache.primary_keys['posts']
+
+        @cache = Marshal.load(Marshal.dump(@cache))
+
+        assert_equal 12, @cache.columns['posts'].size
+        assert_equal 12, @cache.columns_hash['posts'].size
+        assert @cache.tables['posts']
+        assert_equal 'id', @cache.primary_keys['posts']
+      end
+
     end
   end
 end

railties/lib/rails/application/configuration.rb

@@ -11,7 +11,7 @@ module Rails
                     :force_ssl, :helpers_paths, :logger, :log_tags, :preload_frameworks,
                     :railties_order, :relative_url_root, :secret_token,
                     :serve_static_assets, :ssl_options, :static_cache_control, :session_options,
-                    :time_zone, :reload_classes_only_on_change
+                    :time_zone, :reload_classes_only_on_change, :use_schema_cache_dump
 
       attr_writer :log_level
       attr_reader :encoding
@@ -41,6 +41,7 @@ module Rails
         @file_watcher                  = ActiveSupport::FileUpdateChecker
         @exceptions_app                = nil
         @autoflush_log                 = true
+        @use_schema_cache_dump         = true
 
         @assets = ActiveSupport::OrderedOptions.new
         @assets.enabled                  = false

railties/test/application/initializers/frameworks_test.rb

@@ -193,5 +193,31 @@ module ApplicationTests
       require "#{app_path}/config/environment"
       assert_nil defined?(ActiveRecord::Base)
     end
+
+    test "use schema cache dump" do
+      Dir.chdir(app_path) do
+        `rails generate model post title:string`
+        `bundle exec rake db:migrate`
+        `bundle exec rake db:schema:cache:dump`
+      end
+      require "#{app_path}/config/environment"
+      ActiveRecord::Base.connection.drop_table("posts") # force drop posts table for test.
+      assert ActiveRecord::Base.connection.schema_cache.tables["posts"]
+    end
+
+    test "expire schema cache dump" do
+      Dir.chdir(app_path) do
+        `rails generate model post title:string`
+        `bundle exec rake db:migrate`
+        `bundle exec rake db:schema:cache:dump`
+
+        `bundle exec rake db:rollback`
+      end
+      silence_warnings {
+        require "#{app_path}/config/environment"
+        assert !ActiveRecord::Base.connection.schema_cache.tables["posts"]
+      }
+    end
+
   end
 end

railties/test/application/rake_test.rb

@@ -138,5 +138,24 @@ module ApplicationTests
       end
       assert File.exists?(File.join(app_path, 'db', 'my_structure.sql'))
     end
+
+    def test_rake_dump_schema_cache
+      Dir.chdir(app_path) do
+        `rails generate model post title:string`
+        `rails generate model product name:string`
+        `bundle exec rake db:migrate`
+        `bundle exec rake db:schema:cache:dump`
+      end
+      assert File.exists?(File.join(app_path, 'db', 'schema_cache.dump'))
+    end
+
+    def test_rake_clear_schema_cache
+      Dir.chdir(app_path) do
+        `bundle exec rake db:schema:cache:dump`
+        `bundle exec rake db:schema:cache:clear`
+      end
+      assert !File.exists?(File.join(app_path, 'db', 'schema_cache.dump'))
+    end
+
   end
 end