=head1 NAME
-DBIx::Class::Storage::DBI::Replication - Replicated database support
+DBIx::Class::Storage::DBI::Replication - EXPERIMENTAL Replicated database support
=head1 SYNOPSIS
[ "dbi:mysql:database=test;hostname=master", "username", "password", { AutoCommit => 1 } ], # master
[ "dbi:mysql:database=test;hostname=slave1", "username", "password", { priority => 10 } ], # slave1
[ "dbi:mysql:database=test;hostname=slave2", "username", "password", { priority => 10 } ], # slave2
+ [ $dbh, '','', {priority=>10}], # add in a preexisting database handle
+ [ sub { DBI->connect }, '', '', {priority=>10}], # DBD::Multi will call this coderef for connects
<...>,
{ limit_dialect => 'LimitXY' } # If needed, see below
] );
=head1 DESCRIPTION
-This class implements replicated data store for DBI. Currently you can define one master and numerous slave database
-connections. All write-type queries (INSERT, UPDATE, DELETE and even LAST_INSERT_ID) are routed to master database,
-all read-type queries (SELECTs) go to the slave database.
+Warning: This class is marked EXPERIMENTAL. It works for the authors but does
+not currently have automated tests so your mileage may vary.
-For every slave database you can define a priority value, which controls data source usage pattern. It uses
-L<DBD::Multi>, so first the lower priority data sources used (if they have the same priority, the are used
-randomized), than if all low priority data sources fail, higher ones tried in order.
+This class implements replicated data store for DBI. Currently you can define
+one master and numerous slave database connections. All write-type queries
+(INSERT, UPDATE, DELETE and even LAST_INSERT_ID) are routed to master
+database, all read-type queries (SELECTs) go to the slave database.
+
+For every slave database you can define a priority value, which controls data
+source usage pattern. It uses L<DBD::Multi>, so first the lower priority data
+sources used (if they have the same priority, the are used randomized), than
+if all low priority data sources fail, higher ones tried in order.
=head1 CONFIGURATION
=head2 Limit dialect
-If you use LIMIT in your queries (effectively, if you use SQL::Abstract::Limit), do not forget to set up limit_dialect (perldoc SQL::Abstract::Limit) by passing it as an option in the (optional) hash reference to connect_info.
-DBIC can not set it up automatically, since it can not guess DBD::Multi connection types.
+If you use LIMIT in your queries (effectively, if you use
+SQL::Abstract::Limit), do not forget to set up limit_dialect (perldoc
+SQL::Abstract::Limit) by passing it as an option in the (optional) hash
+reference to connect_info. DBIC can not set it up automatically, since it can
+not guess DBD::Multi connection types.
=cut
$global_options = ref $info->[-1] eq 'HASH' ? pop( @$info ) : {};
if( ref( $options = $info->[0]->[-1] ) eq 'HASH' ) {
# Local options present in dsn, merge them with global options
- map { $global_options->{$_} = $options->{$_} } keys %$options;
- pop @{$info->[0]};
+ map { $global_options->{$_} = $options->{$_} } keys %$options;
+ pop @{$info->[0]};
}
- # We need to copy-pass $global_options, since connect_info clears it while processing options
- $self->write_source->connect_info( [ @{$info->[0]}, { %$global_options } ] );
-
- @dsns = map { ($_->[3]->{priority} || 10) => $_ } @{$info}[1..@$info-1];
+ # We need to copy-pass $global_options, since connect_info clears it while
+ # processing options
+ $self->write_source->connect_info( @{$info->[0]}, { %$global_options } );
+
+ ## allow either a DSN string or an already connect $dbh. Just remember if
+ ## you use the $dbh option then DBD::Multi has no idea how to reconnect in
+ ## the event of a failure.
+
+ @dsns = map {
+ ## if the first element in the arrayhash is a ref, make that the value
+ my $db = ref $_->[0] ? $_->[0] : $_;
+ ($_->[3]->{priority} || 10) => $db;
+ } @{$info->[0]}[1..@{$info->[0]}-1];
+
$global_options->{dsns} = \@dsns;
$self->read_source->connect_info( [ 'dbi:Multi:', undef, undef, { %$global_options } ] );