1 package DBIx::Class::Storage::DBI::Replicated;
4 use DBIx::Class::Storage::DBI;
5 use DBIx::Class::Storage::DBI::Replicated::Pool;
6 use DBIx::Class::Storage::DBI::Replicated::Balancer;
7 use Scalar::Util qw(blessed);
9 extends 'DBIx::Class::Storage::DBI', 'Moose::Object';
13 DBIx::Class::Storage::DBI::Replicated - ALPHA Replicated database support
17 The Following example shows how to change an existing $schema to a replicated
18 storage type, add some replicated (readonly) databases, and perform reporting
21 ## Change storage_type in your schema class
22 $schema->storage_type( ['::DBI::Replicated', {balancer=>'::Random'}] );
24 ## Add some slaves. Basically this is an array of arrayrefs, where each
25 ## arrayref is database connect information
27 $schema->storage->connect_replicants(
28 [$dsn1, $user, $pass, \%opts],
29 [$dsn1, $user, $pass, \%opts],
30 [$dsn1, $user, $pass, \%opts],
35 Warning: This class is marked ALPHA. We are using this in development and have
36 some basic test coverage but the code hasn't yet been stressed by a variety
37 of databases. Individual DB's may have quirks we are not aware of. Please
38 use this in development and pass along your experiences/bug fixes.
40 This class implements replicated data store for DBI. Currently you can define
41 one master and numerous slave database connections. All write-type queries
42 (INSERT, UPDATE, DELETE and even LAST_INSERT_ID) are routed to master
43 database, all read-type queries (SELECTs) go to the slave database.
45 Basically, any method request that L<DBIx::Class::Storage::DBI> would normally
46 handle gets delegated to one of the two attributes: L</read_handler> or to
47 L</write_handler>. Additionally, some methods need to be distributed
48 to all existing storages. This way our storage class is a drop in replacement
49 for L<DBIx::Class::Storage::DBI>.
51 Read traffic is spread across the replicants (slaves) occuring to a user
52 selected algorithm. The default algorithm is random weighted.
56 The consistancy betweeen master and replicants is database specific. The Pool
57 gives you a method to validate it's replicants, removing and replacing them
58 when they fail/pass predefined criteria. It is recommened that your application
59 define two schemas, one using the replicated storage and another that just
60 connects to the master.
64 This class defines the following attributes.
68 Contains the classname which will instantiate the L</pool> object. Defaults
69 to: L<DBIx::Class::Storage::DBI::Replicated::Pool>.
78 'create_pool' => 'new',
84 Contains a hashref of initialized information to pass to the Balancer object.
85 See L<DBIx::Class::Storage::Replicated::Pool> for available arguments.
100 The replication pool requires a balance class to provider the methods for
101 choose how to spread the query load across each replicant in the pool.
105 has 'balancer_type' => (
110 'create_balancer' => 'new',
116 Contains a hashref of initialized information to pass to the Balancer object.
117 See L<DBIx::Class::Storage::Replicated::Balancer> for available arguments.
121 has 'balancer_args' => (
131 Is a <DBIx::Class::Storage::DBI::Replicated::Pool> or derived class. This is a
132 container class for one or more replicated databases.
138 isa=>'DBIx::Class::Storage::DBI::Replicated::Pool',
149 Is a <DBIx::Class::Storage::DBI::Replicated::Balancer> or derived class. This
150 is a class that takes a pool (<DBIx::Class::Storage::DBI::Replicated::Pool>)
156 isa=>'DBIx::Class::Storage::DBI::Replicated::Balancer',
158 handles=>[qw/auto_validate_every/],
163 The master defines the canonical state for a pool of connected databases. All
164 the replicants are expected to match this databases state. Thus, in a classic
165 Master / Slaves distributed system, all the slaves are expected to replicate
166 the Master's state as quick as possible. This is the only database in the
167 pool of databases that is allowed to handle write traffic.
173 isa=>'DBIx::Class::Storage::DBI',
177 =head1 ATTRIBUTES IMPLEMENTING THE DBIx::Storage::DBI INTERFACE
179 The following methods are delegated all the methods required for the
180 L<DBIx::Class::Storage::DBI> interface.
184 Defines an object that implements the read side of L<BIx::Class::Storage::DBI>.
188 has 'read_handler' => (
201 Defines an object that implements the write side of L<BIx::Class::Storage::DBI>.
205 has 'write_handler' => (
218 deployment_statements
238 This class defines the following methods.
242 L<DBIx::Class::Schema> when instantiating it's storage passed itself as the
243 first argument. We need to invoke L</new> on the underlying parent class, make
244 sure we properly give it a L<Moose> meta class, and then correctly instantiate
245 our attributes. Basically we pass on whatever the schema has in it's class
246 data for 'storage_type_args' to our replicated storage type.
251 my $class = shift @_;
252 my $schema = shift @_;
253 my $storage_type_args = shift @_;
254 my $obj = $class->SUPER::new($schema, $storage_type_args, @_);
256 ## Hate to do it this way, but can't seem to get advice on the attribute working right
257 ## maybe we can do a type and coercion for it.
258 if( $storage_type_args->{balancer_type} && $storage_type_args->{balancer_type}=~m/^::/) {
259 $storage_type_args->{balancer_type} = 'DBIx::Class::Storage::DBI::Replicated::Balancer'.$storage_type_args->{balancer_type};
260 eval "require $storage_type_args->{balancer_type}";
263 return $class->meta->new_object(
264 __INSTANCE__ => $obj,
272 Lazy builder for the L</master> attribute.
277 DBIx::Class::Storage::DBI->new;
280 =head2 _build_pool_type
282 Lazy builder for the L</pool_type> attribute.
286 sub _build_pool_type {
287 return 'DBIx::Class::Storage::DBI::Replicated::Pool';
292 Lazy builder for the L</pool> attribute.
298 $self->create_pool(%{$self->pool_args});
301 =head2 _build_balancer_type
303 Lazy builder for the L</balancer_type> attribute.
307 sub _build_balancer_type {
308 return 'DBIx::Class::Storage::DBI::Replicated::Balancer::First';
311 =head2 _build_balancer
313 Lazy builder for the L</balancer> attribute. This takes a Pool object so that
314 the balancer knows which pool it's balancing.
318 sub _build_balancer {
320 $self->create_balancer(
322 master=>$self->master,
323 %{$self->balancer_args},);
326 =head2 _build_write_handler
328 Lazy builder for the L</write_handler> attribute. The default is to set this to
333 sub _build_write_handler {
334 return shift->master;
337 =head2 _build_read_handler
339 Lazy builder for the L</read_handler> attribute. The default is to set this to
344 sub _build_read_handler {
345 return shift->balancer;
348 =head2 around: connect_replicants
350 All calls to connect_replicants needs to have an existing $schema tacked onto
351 top of the args, since L<DBIx::Storage::DBI> needs it.
355 around 'connect_replicants' => sub {
356 my ($method, $self, @args) = @_;
357 $self->$method($self->schema, @args);
362 Returns an array of of all the connected storage backends. The first element
363 in the returned array is the master, and the remainings are each of the
371 return grep {defined $_ && blessed $_} (
377 =head2 set_reliable_storage
379 Sets the current $schema to be 'reliable', that is all queries, both read and
380 write are sent to the master
384 sub set_reliable_storage {
386 my $schema = $self->schema;
387 my $write_handler = $self->schema->storage->write_handler;
389 $schema->storage->read_handler($write_handler);
392 =head2 set_balanced_storage
394 Sets the current $schema to be use the </balancer> for all reads, while all
395 writea are sent to the master only
399 sub set_balanced_storage {
401 my $schema = $self->schema;
402 my $write_handler = $self->schema->storage->balancer;
404 $schema->storage->read_handler($write_handler);
409 Check that the master and at least one of the replicants is connected.
417 $self->master->connected &&
418 $self->pool->connected_replicants;
421 =head2 ensure_connected
423 Make sure all the storages are connected.
427 sub ensure_connected {
429 foreach my $source ($self->all_storages) {
430 $source->ensure_connected(@_);
436 Set the limit_dialect for all existing storages
442 foreach my $source ($self->all_storages) {
443 $source->limit_dialect(@_);
449 Set the quote_char for all existing storages
455 foreach my $source ($self->all_storages) {
456 $source->quote_char(@_);
462 Set the name_sep for all existing storages
468 foreach my $source ($self->all_storages) {
469 $source->name_sep(@_);
475 Set the schema object for all existing storages
481 foreach my $source ($self->all_storages) {
482 $source->set_schema(@_);
488 set a debug flag across all storages
494 foreach my $source ($self->all_storages) {
501 set a debug object across all storages
507 foreach my $source ($self->all_storages) {
508 $source->debugobj(@_);
514 set a debugfh object across all storages
520 foreach my $source ($self->all_storages) {
521 $source->debugfh(@_);
527 set a debug callback across all storages
533 foreach my $source ($self->all_storages) {
534 $source->debugcb(@_);
540 disconnect everything
546 foreach my $source ($self->all_storages) {
547 $source->disconnect(@_);
553 Norbert Csongrádi <bert@cpan.org>
555 Peter Siklósi <einon@einon.hu>
557 John Napiorkowski <john.napiorkowski@takkle.com>
561 You may distribute this code under the same terms as Perl itself.