The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.

NAME

DBIx::Class::Schema::Slave - DBIx::Class::Schema for slave (EXPERIMENTAL)

CAUTION

DIBx::Class::Schema::Slave is EXPERIMENTAL and DO NOT use. Please check DBIx::Class::Storage::DBI::Replicated or DBIx::Class::Storage::DBI::Replication. DBIx::Class::Schema::Slave will be deleted.

SYNOPSIS

  # In your MyApp::Schema (DBIx::Class::Schema based)
  package MyApp::Schema;

  use base 'DBIx::Class::Schema';

  __PACKAGE__->load_components( qw/ Schema::Slave / );
  __PACKAGE__->slave_moniker('::Slave');
  __PACKAGE__->slave_connect_info( [
      [ 'dbi:mysql:database:hostname=host', 'username', 'passsword', { ... } ],
      [ 'dbi:mysql:database:hostname=host', 'username', 'passsword', { ... } ],
      [ 'dbi:mysql:database:hostname=host', 'username', 'passsword', { ... } ],
      ...,
  ] );

  # As it is now, DBIx::Class::Schema::Slave works out with DBIx::Class::Schema::Loader.
  # If you use DBIx::Class::Schema::Loader based MyApp::Schema, maybe it is just like below.

  # In your MyApp::Schema (DBIx::Class::Schema::Loader based)
  package MyApp::Schema;

  use base 'DBIx::Class::Schema::Loader';

  __PACKAGE__->load_components( qw/ Schema::Slave / );
  __PACKAGE__->slave_moniker('::Slave');
  __PACKAGE__->slave_connect_info( [
      [ 'dbi:mysql:database:hostname=host', 'username', 'passsword', { ... } ],
      [ 'dbi:mysql:database:hostname=host', 'username', 'passsword', { ... } ],
      [ 'dbi:mysql:database:hostname=host', 'username', 'passsword', { ... } ],
      ...,
  ] );
  __PACKAGE__->loader_options(
      relationships => 1,
      components    => [ qw/
          ...
          ...
          ...
          Row::Slave # DO NOT forget to specify
          Core
      / ],
  );

  # Somewhere in your code
  use MyApp::Schema;

  # First, connect to master
  my $schema = MyApp::Schema->connect( @master_connect_info );

  # Retrieving from master
  my $master = $schema->resultset('Track')->find( $id );

  # Retrieving from slave
  my $slave = $schema->resultset('Track::Slave')->find( $id );

See DBIx::Class::Schema.

DESCRIPTION

DBIx::Class::Schema::Slave is DBIx::Class::Schema for slave. DBIx::Class::Schema::Slave creates result_source classes for slave automatically, and connects slave datasources as you like (or at rondom). You can retrieve rows from either master or slave in the same way DBIx::Class::Schema provies but you can neither add nor remove rows from slave.

SETTIN UP DBIx::Class::Schema::Slave

Setting it up manually

First, load DBIx::Class::Schema::Slave as component in your MyApp::Schema.

  # In your MyApp::Schema
  package MyApp::Schema;

  use base 'DBIx::Class::Schema';

  __PACKAGE__->load_components( qw/ Schema::Slave / );

Set "slave_moniker" as you like. If you do not specify, ::Slave is set.

  __PACKAGE__->slave_moniker('::Slave');

Set "slave_connect_info" as ARRAYREF of ARRAYREF.

  __PACKAGE__->slave_connect_info( [
      [ 'dbi:mysql:database:hostname=host', 'user', 'passsword', { ... } ],
      [ 'dbi:mysql:database:hostname=host', 'user', 'passsword', { ... } ],
      [ 'dbi:mysql:database:hostname=host', 'user', 'passsword', { ... } ],
      ...,
  ] );

Next, you have MyApp::Schema::Artist, MyApp::Schema::Album, MyApp::Schema::Track, load these result_source classes.

  __PACKAGE__->load_classes( qw/ Artist Album Track / );

In running "register_source", DBIx::Class::Schema::Slave creates slave result_source classes MyApp::Schema::Artist::Slave, MyApp::Schema::Album::Slave and MyApp::Schema::Track::Slave automatically. If you set ::MySlave to "slave_moniker", it creates MyApp::Schema::Artist::MySlave, MyApp::Schema::Album::MySlave and MyApp::Schema::Track::MySlave.

  # MyApp::Schema::Artist wouldn't be loaded
  # MyApp::Schema::Artist::Slave wouldn't be created
  __PACKAGE__->load_classes( qw/ Album Track / );

I recommend every result_source classes to be loaded.

  # Every result_source classes are loaded
  __PACKAGE__->load_classes;

Next, load DBIx::Class::Row::Slave as component in your result_source classes.

  # In your MyApp::Schema::Artist;
  package MyApp::Schema::Artist;

  use base 'DBIx::Class';

  __PACKEAGE__->load_components( qw/ ... Row::Slave Core / );

Using DBIx::Class::Schema::Loader

As it is now, DBIx::Class::Schema::Slave WORKS OUT with DBIx::Class::Schema::Loader. First, load DBIx::Class::Schema::Slave as component in your MyApp::Schema.

  # In your MyApp::Schema
  package MyApp::Schema;

  use base 'DBIx::Class::Schema::Loader';

  __PACKAGE__->load_components( qw/ Schema::Slave / );

Set "slave_moniker" as you like. If you do not specify, ::Slave is set.

  __PACKAGE__->slave_moniker('::Slave');

Set "slave_connect_info" as ARRAYREF of ARRAYREF.

  __PACKAGE__->slave_connect_info( [
      [ 'dbi:mysql:database:hostname=host', 'user', 'passsword', { ... } ],
      [ 'dbi:mysql:database:hostname=host', 'user', 'passsword', { ... } ],
      [ 'dbi:mysql:database:hostname=host', 'user', 'passsword', { ... } ],
      ...,
  ] );

Call "loader_options" in DBIx::Class::Schema::Loader. DO NOT forget to specify DBIx::Class::Row::Slave as component.

  __PACKAGE__->loader_options(
      relationships => 1,
      components    => [ qw/
          ...
          Row::Slave # DO NOT forget to load
          Core
      / ],
  );

Connecting (Create Schema instance)

To connect your Schema, provive connect_info not for slave but for master.

  my $schema = MyApp::Schema->connect( @master_connect_info );

Retrieving

Retrieving from master, you don't have to care about anything.

  my $album_master     = $schema->resultset('Album')->find( $id );
  my $itr_album_master = $schema->resultset('Album')->search( { ... }, { ... } );

Retrieving from slave, set slave moniker to "resultset".

  my $track_slave     = $schema->resultset('Album::Slave')->find( $id );
  my $itr_track_slave = $schema->resultset('Album::Slave')->search( { ... }, { ... } );

Adding and removing rows

You can either create a new row or remove some rows from master. But you can neither create a new row nor remove some rows from slave.

  # These complete normally
  my $track = $schema->resultset('Track')->create( {
      created_on  => $dt->now || undef,
      modified_on => $dt->now || undef,
      album_id    => $album->id || undef,
      title       => $title || undef,
      time        => $time || undef,
  } );
  $track->title('WORLD\'S END SUPERNOVA');
  $track->update;
  $track->delete;

  # You got an error!
  # DBIx::Class::ResultSet::create(): Can't insert via result source "Track::Slave". This is slave connection.
  my $track = $schema->resultset('Track::Slave')->create( {
      created_on  => $dt->now || undef,
      modified_on => $dt->now || undef,
      album_id    => $album->id || undef,
      title       => $title || undef,
      time        => $time || undef,
  } );

  $track->title('TEAM ROCK');
  # You got an error!
  # DBIx::Class::Row::Slave::update(): Can't update via result source "Track::Slave". This is slave connection.
  $track->update;

  # And, you got an error!
  # DBIx::Class::Row::Slave::delete(): Can't delete via result source "Track::Slave". This is slave connection.
  $track->delete;

DO NOT call "update_all" in DBIx::Class::ResultSet, "delete_all" in DBIx::Class::ResultSet, "populate" in DBIx::Class::ResultSet and "update_or_create" in DBIx::Class::ResultSet via slave result_sources. Also you SHOULD NOT call "find_or_new" in DBIx::Class::ResultSet, "find_or_create" in DBIx::Class::ResultSet via slave result_sources.

CLASS DATA

slave_moniker

Moniker suffix for slave. ::Slave default.

  # In your MyApp::Schema
  __PACKAGE__->slave_moniker('::Slave');

IMPORTANT: If you have already MyApp::Schema::Artist::Slave, DO NOT set ::Slave to slave_moniker. Set ::SlaveFor or something else.

slave_connect_info

connect_infos ARRAYREF of ARRAYREF for slave.

  # In your MyApp::Schema
  __PACKAGE__->slave_connect_info( [
      [ 'dbi:mysql:database:hostname=host', 'username', 'passsword', { ... } ],
      [ 'dbi:mysql:database:hostname=host', 'username', 'passsword', { ... } ],
      [ 'dbi:mysql:database:hostname=host', 'username', 'passsword', { ... } ],
      ...,
  ] );

slave_schema

Schema for slave. You can get this by "slave".

METHODS

register_source

Arguments: $moniker, $result_source
Return Value: none

Registers the DBIx::Class::ResultSource in the schema with the given moniker and re-maps class_mappings and source_registrations.

  # Re-mapped class_mappings
  class_mappings => {
      MyApp::Schema::Artist        => 'Artist',
      MyApp::Schema::Artist::Slave => 'Artist::Slave',
      MyApp::Schema::Album         => 'Album',
      MyApp::Schema::Album::Slave  => 'Album::Slave',
      MyApp::Schema::Track         => 'Track',
      MyApp::Schema::Track::Slave  => 'Track::Slave',
  }

  # Re-mapped source_registrations
  source_registrations => {
      MyApp::Schema::Artist => {
          bless( {
              ...,
              ...,
              ...,
          }, DBIx::Class::ResultSource::Table )
      },
      MyApp::Schema::Artist::Slave => {
          bless( {
              ...,
              ...,
              ...,
          }, DBIx::Class::ResultSource::Table )
      },
      ...,
      ...,
      ...,
      MyApp::Schema::Track::Slave => {
          bless( {
              ...,
              ...,
              ...,
          }, DBIx::Class::ResultSource::Table )
      },
  }

See "register_source" in DBIx::Class::Schema.

resultset

Arguments: $moniker
Return Value: $result_set

If $moniker is slave moniker, this method returns $result_set for slave. See "resultset" in DBIx::Class::Schema.

  my $master_rs = $schema->resultset('Artist');
  my $slave_rs  = $schema->resultset('Artist::Slave');

sources

Argunemts: none
Return Value: @sources

This method returns the sorted alphabetically source monikers of all source registrations on this schema. See "sources" in DBIx::Class::Schema.

  # Returns all sources including slave sources
  my @all_sources = $schema->sources;

master_sources

Argunemts: none
Return Value: @sources

This method returns the sorted alphabetically master source monikers of all source registrations on this schema.

  my @master_sources = $schema->master_sources;

slave_sources

Argunemts: none
Return Value: @sources

This method returns the sorted alphabetically slave source monikers of all source registrations on this schema.

  my @slave_sources = $schema->slave_sources;

slave_connect

Arguments: @info
Return Value: $slave_schema

This method creates slave connection, and store it in slave_schema. You can get this by "slave". Usualy, you don't have to call it directry.

slave

Getter for "slave_schema". You can get schema for slave if it stored in "slave_schema".

  my $slave_schema = $schema->slave;

select_connect_info

Argunemts: none
Return Value: $connect_info

You can define this method in your schema class as you like. This method has to return $connect_info as ARRAYREF. If "select_connect_info" returns undef, undef value or not ARRAYREF, "_select_connect_info" will be called, and return $connect_info at random from "slave_connect_info".

  # In your MyApp::Schame
  sub select_connect_info {
      my $self = shift;

      my @connect_info = @{$self->slave_connect_info};
      my $connect_info;
      # Some algorithm to select connect_info here

      return $connect_info;
  }

is_slave

Arguments: $string
Return Value: 1 or 0

This method returns 1 if $string (moniker, class name and so on) is slave stuff, otherwise returns 0.

  __PACKAGE__->slave_moniker('::Slave');

  # Returns 0
  $self->is_slave('Artist');

  # Returns 1
  $self->is_slave('Artist::Slave');

  # Returns 1
  $self->is_slave('MyApp::Model::DBIC::MyApp::Artist::Slave');

  __PACKAGE__->slave_moniker('::SlaveFor');

  # Returns 0
  $self->is_slave('Artist');

  # Returns 1
  $self->is_slave('Artist::SlaveFor');

  # Returns 1
  $self->is_slave('MyApp::Model::DBIC::MyApp::Artist::SlaveFor');

INTERNAL METHOD

_select_connect_info

Return Value: $connect_info

Internal method. This method returns $connect_info for slave as ARRAYREF. Usually, you don't have to call it directry. If you select $connect_info as you like, define "select_connect_info" in your schema class. See "select_connect_info" for more information.

AUTHOR

travail travail@cabane.no-ip.org

COPYRIGHT

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

The full text of the license can be found in the LICENSE file included with this module.