1 package DBIx::Class::Storage::DBI::Replicated;
5 die('The following modules are required for Replication ' . DBIx::Class::Optional::Dependencies->req_missing_for ('replicated') . "\n" )
6 unless DBIx::Class::Optional::Dependencies->req_ok_for ('replicated');
10 use DBIx::Class::Storage::DBI;
11 use DBIx::Class::Storage::DBI::Replicated::Pool;
12 use DBIx::Class::Storage::DBI::Replicated::Balancer;
13 use DBIx::Class::Storage::DBI::Replicated::Types qw/BalancerClassNamePart DBICSchema DBICStorageDBI/;
14 use MooseX::Types::Moose qw/ClassName HashRef Object/;
15 use Scalar::Util 'reftype';
17 use List::Util qw/min max reduce/;
21 use namespace::clean -except => 'meta';
25 DBIx::Class::Storage::DBI::Replicated - BETA Replicated database support
29 The Following example shows how to change an existing $schema to a replicated
30 storage type, add some replicated (read-only) databases, and perform reporting
33 You should set the 'storage_type attribute to a replicated type. You should
34 also define your arguments, such as which balancer you want and any arguments
35 that the Pool object should get.
37 my $schema = Schema::Class->clone;
38 $schema->storage_type( ['::DBI::Replicated', {balancer=>'::Random'}] );
39 $schema->connection(...);
41 Next, you need to add in the Replicants. Basically this is an array of
42 arrayrefs, where each arrayref is database connect information. Think of these
43 arguments as what you'd pass to the 'normal' $schema->connect method.
45 $schema->storage->connect_replicants(
46 [$dsn1, $user, $pass, \%opts],
47 [$dsn2, $user, $pass, \%opts],
48 [$dsn3, $user, $pass, \%opts],
51 Now, just use the $schema as you normally would. Automatically all reads will
52 be delegated to the replicants, while writes to the master.
54 $schema->resultset('Source')->search({name=>'etc'});
56 You can force a given query to use a particular storage using the search
57 attribute 'force_pool'. For example:
59 my $RS = $schema->resultset('Source')->search(undef, {force_pool=>'master'});
61 Now $RS will force everything (both reads and writes) to use whatever was setup
62 as the master storage. 'master' is hardcoded to always point to the Master,
63 but you can also use any Replicant name. Please see:
64 L<DBIx::Class::Storage::DBI::Replicated::Pool> and the replicants attribute for more.
66 Also see transactions and L</execute_reliably> for alternative ways to
67 force read traffic to the master. In general, you should wrap your statements
68 in a transaction when you are reading and writing to the same tables at the
69 same time, since your replicants will often lag a bit behind the master.
71 See L<DBIx::Class::Storage::DBI::Replicated::Instructions> for more help and
76 Warning: This class is marked BETA. This has been running a production
77 website using MySQL native replication as its backend and we have some decent
78 test coverage but the code hasn't yet been stressed by a variety of databases.
79 Individual DBs may have quirks we are not aware of. Please use this in first
80 development and pass along your experiences/bug fixes.
82 This class implements replicated data store for DBI. Currently you can define
83 one master and numerous slave database connections. All write-type queries
84 (INSERT, UPDATE, DELETE and even LAST_INSERT_ID) are routed to master
85 database, all read-type queries (SELECTs) go to the slave database.
87 Basically, any method request that L<DBIx::Class::Storage::DBI> would normally
88 handle gets delegated to one of the two attributes: L</read_handler> or to
89 L</write_handler>. Additionally, some methods need to be distributed
90 to all existing storages. This way our storage class is a drop in replacement
91 for L<DBIx::Class::Storage::DBI>.
93 Read traffic is spread across the replicants (slaves) occurring to a user
94 selected algorithm. The default algorithm is random weighted.
98 The consistency between master and replicants is database specific. The Pool
99 gives you a method to validate its replicants, removing and replacing them
100 when they fail/pass predefined criteria. Please make careful use of the ways
101 to force a query to run against Master when needed.
105 Replicated Storage has additional requirements not currently part of
106 L<DBIx::Class>. See L<DBIx::Class::Optional::Dependencies> for more details.
110 This class defines the following attributes.
114 The underlying L<DBIx::Class::Schema> object this storage is attaching
127 Contains the classname which will instantiate the L</pool> object. Defaults
128 to: L<DBIx::Class::Storage::DBI::Replicated::Pool>.
135 default=>'DBIx::Class::Storage::DBI::Replicated::Pool',
137 'create_pool' => 'new',
143 Contains a hashref of initialized information to pass to the Balancer object.
144 See L<DBIx::Class::Storage::DBI::Replicated::Pool> for available arguments.
158 The replication pool requires a balance class to provider the methods for
159 choose how to spread the query load across each replicant in the pool.
163 has 'balancer_type' => (
165 isa=>BalancerClassNamePart,
168 default=> 'DBIx::Class::Storage::DBI::Replicated::Balancer::First',
170 'create_balancer' => 'new',
176 Contains a hashref of initialized information to pass to the Balancer object.
177 See L<DBIx::Class::Storage::DBI::Replicated::Balancer> for available arguments.
181 has 'balancer_args' => (
191 Is a L<DBIx::Class::Storage::DBI::Replicated::Pool> or derived class. This is a
192 container class for one or more replicated databases.
198 isa=>'DBIx::Class::Storage::DBI::Replicated::Pool',
209 Is a L<DBIx::Class::Storage::DBI::Replicated::Balancer> or derived class. This
210 is a class that takes a pool (L<DBIx::Class::Storage::DBI::Replicated::Pool>)
216 isa=>'DBIx::Class::Storage::DBI::Replicated::Balancer',
218 handles=>[qw/auto_validate_every/],
223 The master defines the canonical state for a pool of connected databases. All
224 the replicants are expected to match this databases state. Thus, in a classic
225 Master / Slaves distributed system, all the slaves are expected to replicate
226 the Master's state as quick as possible. This is the only database in the
227 pool of databases that is allowed to handle write traffic.
237 =head1 ATTRIBUTES IMPLEMENTING THE DBIx::Storage::DBI INTERFACE
239 The following methods are delegated all the methods required for the
240 L<DBIx::Class::Storage::DBI> interface.
244 my $method_dispatch = {
256 deployment_statements
259 build_datetime_parser
275 with_deferred_fk_checks
284 relname_to_table_alias
286 _default_dbi_connect_attributes
288 _dbic_connect_attributes
294 bind_attribute_by_data_type
300 _per_row_update_delete
301 _dbh_execute_inserts_with_no_binds
302 _select_args_to_query
305 _multipk_update_delete
306 _normalize_connect_info
321 /, Class::MOP::Class->initialize('DBIx::Class::Storage::DBIHacks')->get_method_list ],
326 _dbh_columns_info_for
329 unimplemented => [qw/
330 _arm_global_destructor
333 source_bind_attributes
335 get_use_dbms_capability
336 set_use_dbms_capability
355 # the capability framework
356 # not sure if CMOP->initialize does evil things to DBIC::S::DBI, fix if a problem
358 { $_ =~ /^ _ (?: use | supports | determine_supports ) _ /x }
359 ( Class::MOP::Class->initialize('DBIx::Class::Storage::DBI')->get_all_method_names )
363 if (DBIx::Class::_ENV_::DBICTEST) {
366 for my $type (keys %$method_dispatch) {
367 for (@{$method_dispatch->{$type}}) {
368 push @{$seen->{$_}}, $type;
372 if (my @dupes = grep { @{$seen->{$_}} > 1 } keys %$seen) {
374 'The following methods show up multiple times in ::Storage::DBI::Replicated handlers:',
375 (map { "$_: " . (join ', ', @{$seen->{$_}}) } sort @dupes),
380 if (my @cant = grep { ! DBIx::Class::Storage::DBI->can($_) } keys %$seen) {
382 '::Storage::DBI::Replicated specifies handling of the following *NON EXISTING* ::Storage::DBI methods:',
389 for my $method (@{$method_dispatch->{unimplemented}}) {
390 __PACKAGE__->meta->add_method($method, sub {
392 $self->throw_exception("$method must not be called on ".(blessed $self).' objects');
398 Defines an object that implements the read side of L<BIx::Class::Storage::DBI>.
402 has 'read_handler' => (
406 handles=>$method_dispatch->{reader},
411 Defines an object that implements the write side of L<BIx::Class::Storage::DBI>,
412 as well as methods that don't write or read that can be called on only one
413 storage, methods that return a C<$dbh>, and any methods that don't make sense to
418 has 'write_handler' => (
422 handles=>$method_dispatch->{writer},
427 has _master_connect_info_opts =>
428 (is => 'rw', isa => HashRef, default => sub { {} });
430 =head2 around: connect_info
432 Preserves master's C<connect_info> options (for merging with replicants.)
433 Also sets any Replicated-related options from connect_info, such as
434 C<pool_type>, C<pool_args>, C<balancer_type> and C<balancer_args>.
438 around connect_info => sub {
439 my ($next, $self, $info, @extra) = @_;
441 my $merge = Hash::Merge->new('LEFT_PRECEDENT');
444 for my $arg (@$info) {
445 next unless (reftype($arg)||'') eq 'HASH';
446 %opts = %{ $merge->merge($arg, \%opts) };
450 if (@opts{qw/pool_type pool_args/}) {
451 $self->pool_type(delete $opts{pool_type})
455 $merge->merge((delete $opts{pool_args} || {}), $self->pool_args)
458 ## Since we possibly changed the pool_args, we need to clear the current
459 ## pool object so that next time it is used it will be rebuilt.
463 if (@opts{qw/balancer_type balancer_args/}) {
464 $self->balancer_type(delete $opts{balancer_type})
465 if $opts{balancer_type};
467 $self->balancer_args(
468 $merge->merge((delete $opts{balancer_args} || {}), $self->balancer_args)
471 $self->balancer($self->_build_balancer)
475 $self->_master_connect_info_opts(\%opts);
479 @res = $self->$next($info, @extra);
481 $res[0] = $self->$next($info, @extra);
484 # Make sure master is blessed into the correct class and apply role to it.
485 my $master = $self->master;
486 $master->_determine_driver;
487 Moose::Meta::Class->initialize(ref $master);
489 DBIx::Class::Storage::DBI::Replicated::WithDSN->meta->apply($master);
491 # link pool back to master
492 $self->pool->master($master);
494 wantarray ? @res : $res[0];
499 This class defines the following methods.
503 L<DBIx::Class::Schema> when instantiating its storage passed itself as the
504 first argument. So we need to massage the arguments a bit so that all the
505 bits get put into the correct places.
510 my ($class, $schema, $storage_type_args, @args) = @_;
521 Lazy builder for the L</master> attribute.
527 my $master = DBIx::Class::Storage::DBI->new($self->schema);
533 Lazy builder for the L</pool> attribute.
539 $self->create_pool(%{$self->pool_args});
542 =head2 _build_balancer
544 Lazy builder for the L</balancer> attribute. This takes a Pool object so that
545 the balancer knows which pool it's balancing.
549 sub _build_balancer {
551 $self->create_balancer(
553 master=>$self->master,
554 %{$self->balancer_args},
558 =head2 _build_write_handler
560 Lazy builder for the L</write_handler> attribute. The default is to set this to
565 sub _build_write_handler {
566 return shift->master;
569 =head2 _build_read_handler
571 Lazy builder for the L</read_handler> attribute. The default is to set this to
576 sub _build_read_handler {
577 return shift->balancer;
580 =head2 around: connect_replicants
582 All calls to connect_replicants needs to have an existing $schema tacked onto
583 top of the args, since L<DBIx::Storage::DBI> needs it, and any C<connect_info>
584 options merged with the master, with replicant opts having higher priority.
588 around connect_replicants => sub {
589 my ($next, $self, @args) = @_;
592 $r = [ $r ] unless reftype $r eq 'ARRAY';
594 $self->throw_exception('coderef replicant connect_info not supported')
595 if ref $r->[0] && reftype $r->[0] eq 'CODE';
597 # any connect_info options?
599 $i++ while $i < @$r && (reftype($r->[$i])||'') ne 'HASH';
602 $r->[$i] = {} unless $r->[$i];
604 # merge if two hashes
605 my @hashes = @$r[$i .. $#{$r}];
607 $self->throw_exception('invalid connect_info options')
608 if (grep { reftype($_) eq 'HASH' } @hashes) != @hashes;
610 $self->throw_exception('too many hashrefs in connect_info')
613 my $merge = Hash::Merge->new('LEFT_PRECEDENT');
614 my %opts = %{ $merge->merge(reverse @hashes) };
617 splice @$r, $i+1, ($#{$r} - $i), ();
619 # make sure master/replicants opts don't clash
620 my %master_opts = %{ $self->_master_connect_info_opts };
621 if (exists $opts{dbh_maker}) {
622 delete @master_opts{qw/dsn user password/};
624 delete $master_opts{dbh_maker};
627 %opts = %{ $merge->merge(\%opts, \%master_opts) };
633 $self->$next($self->schema, @args);
638 Returns an array of of all the connected storage backends. The first element
639 in the returned array is the master, and the remainings are each of the
646 return grep {defined $_ && blessed $_} (
648 values %{ $self->replicants },
652 =head2 execute_reliably ($coderef, ?@args)
654 Given a coderef, saves the current state of the L</read_handler>, forces it to
655 use reliable storage (e.g. sets it to the master), executes a coderef and then
656 restores the original state.
662 $schema->resultset('User')->create({name=>$name});
663 my $user_rs = $schema->resultset('User')->find({name=>$name});
667 my $user_rs = $schema->storage->execute_reliably($reliably, 'John');
669 Use this when you must be certain of your database state, such as when you just
670 inserted something and need to get a resultset including it, etc.
674 sub execute_reliably {
675 my ($self, $coderef, @args) = @_;
677 unless( ref $coderef eq 'CODE') {
678 $self->throw_exception('Second argument must be a coderef');
681 ##Get copy of master storage
682 my $master = $self->master;
684 ##Get whatever the current read hander is
685 my $current = $self->read_handler;
687 ##Set the read handler to master
688 $self->read_handler($master);
690 ## do whatever the caller needs
692 my $want_array = wantarray;
696 @result = $coderef->(@args);
697 } elsif(defined $want_array) {
698 ($result[0]) = ($coderef->(@args));
703 $self->throw_exception("coderef returned an error: $_");
705 ##Reset to the original state
706 $self->read_handler($current);
709 return wantarray ? @result : $result[0];
712 =head2 set_reliable_storage
714 Sets the current $schema to be 'reliable', that is all queries, both read and
715 write are sent to the master
719 sub set_reliable_storage {
721 my $schema = $self->schema;
722 my $write_handler = $self->schema->storage->write_handler;
724 $schema->storage->read_handler($write_handler);
727 =head2 set_balanced_storage
729 Sets the current $schema to be use the </balancer> for all reads, while all
730 writes are sent to the master only
734 sub set_balanced_storage {
736 my $schema = $self->schema;
737 my $balanced_handler = $self->schema->storage->balancer;
739 $schema->storage->read_handler($balanced_handler);
744 Check that the master and at least one of the replicants is connected.
751 $self->master->connected &&
752 $self->pool->connected_replicants;
755 =head2 ensure_connected
757 Make sure all the storages are connected.
761 sub ensure_connected {
763 foreach my $source ($self->all_storages) {
764 $source->ensure_connected(@_);
770 Set the limit_dialect for all existing storages
776 foreach my $source ($self->all_storages) {
777 $source->limit_dialect(@_);
779 return $self->master->limit_dialect;
784 Set the quote_char for all existing storages
790 foreach my $source ($self->all_storages) {
791 $source->quote_char(@_);
793 return $self->master->quote_char;
798 Set the name_sep for all existing storages
804 foreach my $source ($self->all_storages) {
805 $source->name_sep(@_);
807 return $self->master->name_sep;
812 Set the schema object for all existing storages
818 foreach my $source ($self->all_storages) {
819 $source->set_schema(@_);
825 set a debug flag across all storages
832 foreach my $source ($self->all_storages) {
836 return $self->master->debug;
847 return $self->master->debugobj(@_);
858 return $self->master->debugfh(@_);
869 return $self->master->debugcb(@_);
874 disconnect everything
880 foreach my $source ($self->all_storages) {
881 $source->disconnect(@_);
887 set cursor class on all storages, or return master's
892 my ($self, $cursor_class) = @_;
895 $_->cursor_class($cursor_class) for $self->all_storages;
897 $self->master->cursor_class;
902 set cursor class on all storages, or return master's, alias for L</cursor_class>
908 my ($self, $cursor_class) = @_;
911 $_->cursor($cursor_class) for $self->all_storages;
913 $self->master->cursor;
918 sets the L<DBIx::Class::Storage::DBI/unsafe> option on all storages or returns
919 master's current setting
927 $_->unsafe(@_) for $self->all_storages;
930 return $self->master->unsafe;
933 =head2 disable_sth_caching
935 sets the L<DBIx::Class::Storage::DBI/disable_sth_caching> option on all storages
936 or returns master's current setting
940 sub disable_sth_caching {
944 $_->disable_sth_caching(@_) for $self->all_storages;
947 return $self->master->disable_sth_caching;
950 =head2 lag_behind_master
952 returns the highest Replicant L<DBIx::Class::Storage::DBI/lag_behind_master>
957 sub lag_behind_master {
960 return max map $_->lag_behind_master, $self->replicants;
963 =head2 is_replicating
965 returns true if all replicants return true for
966 L<DBIx::Class::Storage::DBI/is_replicating>
973 return (grep $_->is_replicating, $self->replicants) == ($self->replicants);
976 =head2 connect_call_datetime_setup
978 calls L<DBIx::Class::Storage::DBI/connect_call_datetime_setup> for all storages
982 sub connect_call_datetime_setup {
984 $_->connect_call_datetime_setup for $self->all_storages;
989 $_->_populate_dbh for $self->all_storages;
994 $_->_connect for $self->all_storages;
999 $_->_rebless for $self->all_storages;
1002 sub _determine_driver {
1004 $_->_determine_driver for $self->all_storages;
1007 sub _driver_determined {
1011 $_->_driver_determined(@_) for $self->all_storages;
1014 return $self->master->_driver_determined;
1020 $_->_init for $self->all_storages;
1023 sub _run_connection_actions {
1026 $_->_run_connection_actions for $self->all_storages;
1029 sub _do_connection_actions {
1033 $_->_do_connection_actions(@_) for $self->all_storages;
1037 sub connect_call_do_sql {
1039 $_->connect_call_do_sql(@_) for $self->all_storages;
1042 sub disconnect_call_do_sql {
1044 $_->disconnect_call_do_sql(@_) for $self->all_storages;
1047 sub _seems_connected {
1050 return min map $_->_seems_connected, $self->all_storages;
1056 return min map $_->_ping, $self->all_storages;
1059 # not using the normalized_version, because we want to preserve
1060 # version numbers much longer than the conventional xxx.yyyzzz
1061 my $numify_ver = sub {
1063 my @numparts = split /\D+/, $ver;
1064 my $format = '%d.' . (join '', ('%06d') x (@numparts - 1));
1066 return sprintf $format, @numparts;
1071 if (not $self->_dbh_details->{info}) {
1072 $self->_dbh_details->{info} = (
1073 reduce { $a->[0] < $b->[0] ? $a : $b }
1074 map [ $numify_ver->($_->{dbms_version}), $_ ],
1075 map $_->_server_info, $self->all_storages
1079 return $self->next::method;
1082 sub _get_server_version {
1085 return $self->_server_info->{dbms_version};
1090 Due to the fact that replicants can lag behind a master, you must take care to
1091 make sure you use one of the methods to force read queries to a master should
1092 you need realtime data integrity. For example, if you insert a row, and then
1093 immediately re-read it from the database (say, by doing $row->discard_changes)
1094 or you insert a row and then immediately build a query that expects that row
1095 to be an item, you should force the master to handle reads. Otherwise, due to
1096 the lag, there is no certainty your data will be in the expected state.
1098 For data integrity, all transactions automatically use the master storage for
1099 all read and write queries. Using a transaction is the preferred and recommended
1100 method to force the master to handle all read queries.
1102 Otherwise, you can force a single query to use the master with the 'force_pool'
1105 my $row = $resultset->search(undef, {force_pool=>'master'})->find($pk);
1107 This attribute will safely be ignore by non replicated storages, so you can use
1108 the same code for both types of systems.
1110 Lastly, you can use the L</execute_reliably> method, which works very much like
1113 For debugging, you can turn replication on/off with the methods L</set_reliable_storage>
1114 and L</set_balanced_storage>, however this operates at a global level and is not
1115 suitable if you have a shared Schema object being used by multiple processes,
1116 such as on a web application server. You can get around this limitation by
1117 using the Schema clone method.
1119 my $new_schema = $schema->clone;
1120 $new_schema->set_reliable_storage;
1122 ## $new_schema will use only the Master storage for all reads/writes while
1123 ## the $schema object will use replicated storage.
1127 John Napiorkowski <john.napiorkowski@takkle.com>
1129 Based on code originated by:
1131 Norbert Csongrádi <bert@cpan.org>
1132 Peter Siklósi <einon@einon.hu>
1136 You may distribute this code under the same terms as Perl itself.
1140 __PACKAGE__->meta->make_immutable;