1 package Catalyst::Plugin::Scheduler;
5 use base qw/Class::Accessor::Fast Class::Data::Inheritable/;
7 use DateTime::Event::Cron;
8 use DateTime::TimeZone;
12 use Storable qw/lock_store lock_retrieve/;
15 our $VERSION = '0.06';
17 __PACKAGE__->mk_classdata( '_events' => [] );
18 __PACKAGE__->mk_accessors('_event_state');
21 my ( $class, %args ) = @_;
23 unless ( $args{event} ) {
24 Catalyst::Exception->throw(
25 message => 'The schedule method requires an event parameter' );
28 my $conf = $class->config->{scheduler};
31 trigger => $args{trigger},
32 event => $args{event},
33 auto_run => ( defined $args{auto_run} ) ? $args{auto_run} : 1,
38 # replace keywords that Set::Crontab doesn't support
39 $args{at} = _prepare_cron( $args{at} );
41 # parse the cron entry into a DateTime::Set
43 eval { $set = DateTime::Event::Cron->from_cron( $args{at} ) };
45 Catalyst::Exception->throw(
46 "Scheduler: Unable to parse 'at' value "
51 $event->{at} = $args{at};
56 push @{ $class->_events }, $event;
62 $c->NEXT::dispatch(@_);
64 $c->_get_event_state();
68 # check if a minute has passed since our last check
69 # This check is not run if the user is manually triggering an event
70 if ( time - $c->_event_state->{last_check} < 60 ) {
71 return unless $c->req->params->{schedule_trigger};
73 my $last_check = $c->_event_state->{last_check};
74 $c->_event_state->{last_check} = time;
75 $c->_save_event_state();
77 my $conf = $c->config->{scheduler};
78 my $last_check_dt = DateTime->from_epoch(
80 time_zone => $conf->{time_zone}
82 my $now = DateTime->now( time_zone => $conf->{time_zone} );
85 for my $event ( @{ $c->_events } ) {
88 if ( $event->{trigger}
89 && $event->{trigger} eq $c->req->params->{schedule_trigger} )
92 # manual trigger, run it now
93 next EVENT unless $c->_event_authorized;
97 next EVENT unless $event->{set};
98 $next_run = $event->{set}->next($last_check_dt);
101 if ( $next_run <= $now ) {
103 # do some security checking for non-auto-run events
104 if ( !$event->{auto_run} ) {
105 next EVENT unless $c->_event_authorized;
108 # make sure we're the only process running this event
109 next EVENT unless $c->_mark_running($event);
111 my $event_name = $event->{trigger} || $event->{event};
112 $c->log->debug("Scheduler: Executing $event_name")
113 if $c->config->{scheduler}->{logging};
116 local $c->{error} = [];
118 # return value/output from the event, if any
124 # do not allow the event to modify the response
125 local $c->res->{body};
126 local $c->res->{cookies};
127 local $c->res->{headers};
128 local $c->res->{location};
129 local $c->res->{status};
131 if ( ref $event->{event} eq 'CODE' ) {
132 $output = $event->{event}->($c);
135 $output = $c->forward( $event->{event} );
138 my @errors = @{ $c->{error} };
139 push @errors, $@ if $@;
142 'Scheduler: Error executing ' . "$event_name: $_" )
144 $output = join '; ', @errors;
147 $c->_mark_finished( $event, $output );
155 # initial configuration
156 $c->config->{scheduler}->{logging} ||= ( $c->debug ) ? 1 : 0;
157 $c->config->{scheduler}->{time_zone} ||= $c->_detect_timezone();
158 $c->config->{scheduler}->{state_file} ||= $c->path_to('scheduler.state');
159 $c->config->{scheduler}->{hosts_allow} ||= '127.0.0.1';
160 $c->config->{scheduler}->{yaml_file} ||= $c->path_to('scheduler.yml');
168 return ( $c->NEXT::dump_these(@_) ) unless @{ $c->_events };
170 # for debugging, we dump out a list of all events with their next
173 $c->NEXT::dump_these(@_),
174 [ 'Scheduled Events', $c->scheduler_state ],
178 sub scheduler_state {
181 $c->_get_event_state();
183 my $conf = $c->config->{scheduler};
184 my $now = DateTime->now( time_zone => $conf->{time_zone} );
186 my $last_check = $c->_event_state->{last_check};
187 my $last_check_dt = DateTime->from_epoch(
188 epoch => $last_check,
189 time_zone => $conf->{time_zone},
193 for my $event ( @{ $c->_events } ) {
195 for my $key ( qw/at trigger event auto_run/ ) {
196 $dump->{$key} = $event->{$key} if $event->{$key};
199 # display the next run time
200 if ( $event->{set} ) {
201 my $next_run = $event->{set}->next($last_check_dt);
204 . q{ } . $next_run->hms
205 . q{ } . $next_run->time_zone_short_name;
208 # display the last run time
210 = $c->_event_state->{events}->{ $event->{event} }->{last_run};
212 $last_run = DateTime->from_epoch(
214 time_zone => $conf->{time_zone},
218 . q{ } . $last_run->hms
219 . q{ } . $last_run->time_zone_short_name;
222 # display the result of the last run
224 = $c->_event_state->{events}->{ $event->{event} }->{last_output};
226 $dump->{last_output} = $output;
229 push @{$event_dump}, $dump;
235 # check and reload the YAML file with schedule data
239 # each process needs to load the YAML file independently
240 if ( $c->_event_state->{yaml_mtime}->{$$} ||= 0 ) {
241 return if ( time - $c->_event_state->{last_check} < 60 );
244 return unless -e $c->config->{scheduler}->{yaml_file};
247 my $mtime = ( stat $c->config->{scheduler}->{yaml_file} )->mtime;
248 if ( $mtime > $c->_event_state->{yaml_mtime}->{$$} ) {
249 $c->_event_state->{yaml_mtime}->{$$} = $mtime;
251 # clean up old PID files listed in yaml_mtime
252 foreach my $pid ( keys %{ $c->_event_state->{yaml_mtime} } ) {
253 if ( $c->_event_state->{yaml_mtime}->{$pid} < $mtime ) {
254 delete $c->_event_state->{yaml_mtime}->{$pid};
257 $c->_save_event_state();
259 # wipe out all current events and reload from YAML
262 my $yaml = YAML::LoadFile( $c->config->{scheduler}->{yaml_file} );
264 foreach my $event ( @{$yaml} ) {
265 $c->schedule( %{$event} );
268 $c->log->info( "Scheduler: PID $$ loaded "
270 . ' events from YAML file' )
271 if $c->config->{scheduler}->{logging};
275 $c->log->error("Scheduler: Error reading YAML file: $@");
279 # Detect the current time zone
280 sub _detect_timezone {
284 eval { $tz = DateTime::TimeZone->new( name => 'local' ) };
287 'Scheduler: Unable to autodetect local time zone, using UTC')
288 if $c->config->{scheduler}->{logging};
293 'Scheduler: Using autodetected time zone: ' . $tz->name )
294 if $c->config->{scheduler}->{logging};
299 # Check for authorized users on non-auto events
300 sub _event_authorized {
303 # this should never happen, but just in case...
304 return unless $c->req->address;
306 my $hosts_allow = $c->config->{scheduler}->{hosts_allow};
307 $hosts_allow = [$hosts_allow] unless ref($hosts_allow) eq 'ARRAY';
308 my $allowed = Set::Scalar->new( @{$hosts_allow} );
309 return $allowed->contains( $c->req->address );
312 # get the state from the state file
313 sub _get_event_state {
316 if ( -e $c->config->{scheduler}->{state_file} ) {
318 lock_retrieve $c->config->{scheduler}->{state_file} );
322 # initialize the state file
324 { last_check => time,
329 $c->_save_event_state();
333 # Check the state file to ensure we are the only process running an event
335 my ( $c, $event ) = @_;
337 $c->_get_event_state();
340 $c->_event_state->{events}->{ $event->{event} }->{running};
342 # this is a 2-step process to prevent race conditions
343 # 1. write the state file with our PID
344 $c->_event_state->{events}->{ $event->{event} }->{running} = $$;
345 $c->_save_event_state();
347 # 2. re-read the state file and make sure it's got the same PID
348 $c->_get_event_state();
349 if ( $c->_event_state->{events}->{ $event->{event} }->{running} == $$ ) {
356 # Mark an event as finished
358 my ( $c, $event, $output ) = @_;
360 $c->_event_state->{events}->{ $event->{event} }->{running} = 0;
361 $c->_event_state->{events}->{ $event->{event} }->{last_run} = time;
362 $c->_event_state->{events}->{ $event->{event} }->{last_output} = $output;
363 $c->_save_event_state();
366 # update the state file on disk
367 sub _save_event_state {
370 lock_store $c->_event_state, $c->config->{scheduler}->{state_file};
373 # Set::Crontab does not support day names, or '@' shortcuts
377 return $cron unless $cron =~ /\w/;
403 'yearly' => '0 0 1 1 *',
404 'annually' => '0 0 1 1 *',
405 'monthly' => '0 0 1 * *',
406 'weekly' => '0 0 * * 0',
407 'daily' => '0 0 * * *',
408 'midnight' => '0 0 * * *',
409 'hourly' => '0 * * * *',
412 if ( $cron =~ /^\@/ ) {
414 return $replace_at{ $cron };
417 for my $name ( keys %replace ) {
418 my $value = $replace{$name};
419 $cron =~ s/$name/$value/i;
420 last unless $cron =~ /\w/;
432 Catalyst::Plugin::Scheduler - Schedule events to run in a cron-like fashion
436 use Catalyst qw/Scheduler/;
438 # run remove_sessions in the Cron controller every hour
439 __PACKAGE__->schedule(
441 event => '/cron/remove_sessions'
444 # Run a subroutine at 4:05am every Sunday
445 __PACKAGE__->schedule(
450 # A long-running scheduled event that must be triggered
451 # manually by an authorized user
452 __PACKAGE__->schedule(
453 trigger => 'rebuild_search_index',
454 event => '/cron/rebuild_search_index',
456 $ wget -q http://www.myapp.com/?schedule_trigger=rebuild_search_index
460 This plugin allows you to schedule events to run at recurring intervals.
461 Events will run during the first request which meets or exceeds the specified
462 time. Depending on the level of traffic to the application, events may or may
463 not run at exactly the correct time, but it should be enough to satisfy many
464 basic scheduling needs.
468 Configuration is optional and is specified in MyApp->config->{scheduler}.
472 Set to 1 to enable logging of events as they are executed. This option is
473 enabled by default when running under -Debug mode. Errors are always logged
474 regardless of the value of this option.
478 The time zone of your system. This will be autodetected where possible, or
479 will default to UTC (GMT). You can override the detection by providing a
480 valid L<DateTime> time zone string, such as 'America/New_York'.
484 The current state of every event is stored in a file. By default this is
485 $APP_HOME/scheduler.state. This file is created on the first request if it
486 does not already exist.
490 The location of the optional YAML event configuration file. By default this
491 is $APP_HOME/scheduler.yml.
495 This option specifies IP addresses for trusted users. This option defaults
496 to 127.0.0.1. Multiple addresses can be specified by using an array
497 reference. This option is used for both events where auto_run is set to 0
498 and for manually-triggered events.
500 __PACKAGE__->config->{scheduler}->{hosts_allow} = '192.168.1.1';
501 __PACKAGE__->config->{scheduler}->{hosts_allow} = [
508 =head2 AUTOMATED EVENTS
510 Events are scheduled by calling the class method C<schedule>.
514 event => '/cron/remove_sessions',
517 package MyApp::Controller::Cron;
519 sub remove_sessions : Private {
520 my ( $self, $c ) = @_;
522 $c->delete_expired_sessions;
527 The time to run an event is specified using L<crontab(5)>-style syntax.
529 5 0 * * * # 5 minutes after midnight, every day
530 15 14 1 * * # run at 2:15pm on the first of every month
531 0 22 * * 1-5 # run at 10 pm on weekdays
532 5 4 * * sun # run at 4:05am every Sunday
541 month 0-12 (or names, see below)
542 day of week 0-7 (0 or 7 is Sun, or use names)
544 Instead of the first five fields, one of seven special strings may appear:
548 @yearly Run once a year, "0 0 1 1 *".
549 @annually (same as @yearly)
550 @monthly Run once a month, "0 0 1 * *".
551 @weekly Run once a week, "0 0 * * 0".
552 @daily Run once a day, "0 0 * * *".
553 @midnight (same as @daily)
554 @hourly Run once an hour, "0 * * * *".
558 The event to run at the specified time can be either a Catalyst private
559 action path or a coderef. Both types of event methods will receive the $c
560 object from the current request, but you must not rely on any request-specific
561 information present in $c as it will be from a random user request at or near
562 the event's specified run time.
564 Important: Methods used for events should be marked C<Private> so that
565 they can not be executed via the browser.
569 The auto_run parameter specifies when the event is allowed to be executed.
570 By default this option is set to 1, so the event will be executed during the
571 first request that matches the specified time in C<at>.
573 If set to 0, the event will only run when a request is made by a user from
574 an authorized address. The purpose of this option is to allow long-running
575 tasks to execute only for certain users.
579 event => '/cron/rebuild_search_index',
583 package MyApp::Controller::Cron;
585 sub rebuild_search_index : Private {
586 my ( $self, $c ) = @_;
588 # rebuild the search index, this may take a long time
591 Now, the search index will only be rebuilt when a request is made from a user
592 whose IP address matches the list in the C<hosts_allow> config option. To
593 run this event, you probably want to ping the app from a cron job.
595 0 0 * * * wget -q http://www.myapp.com/
599 To create an event that does not run on a set schedule and must be manually
600 triggered, you can specify the C<trigger> option instead of C<at>.
602 __PACKAGE__->schedule(
603 trigger => 'send_email',
604 event => '/events/send_email',
607 The event may then be triggered by a standard web request from an authorized
608 user. The trigger to run is specified by using a special GET parameter,
609 'schedule_trigger'; the path requested does not matter.
611 http://www.myapp.com/?schedule_trigger=send_email
613 By default, manual events may only be triggered by requests made from
614 localhost (127.0.0.1). To allow other addresses to run events, use the
615 configuration option L</hosts_allow>.
617 =head1 SCHEDULING USING A YAML FILE
619 As an alternative to using the schedule() method, you may define scheduled
620 events in an external YAML file. By default, the plugin looks for the
621 existence of a file called C<schedule.yml> in your application's home
622 directory. You can change the filename using the configuration option
625 Modifications to this file will be re-read once per minute during the normal
626 event checking process.
628 Here's an example YAML configuration file with 4 events. Each event is
629 denoted with a '-' character, followed by the same parameters used by the
630 C<schedule> method. Note that coderef events are not supported by the YAML
635 event: /cron/delete_sessions
636 - event: /cron/send_email
642 event: /cron/rebuild_search_index
646 All events are run inside of an eval container. This protects the user from
647 receiving any error messages or page crashes if an event fails to run
648 properly. All event errors are logged, even if logging is disabled.
650 =head1 PLUGIN SUPPORT
652 Other plugins may register scheduled events if they need to perform periodic
653 maintenance. Plugin authors, B<be sure to inform your users> if you do this!
654 Events should be registered from a plugin's C<setup> method.
660 if ( $c->can('schedule') ) {
670 The time at which an event will run is determined completely by the requests
671 made to the application. Apps with heavy traffic may have events run at very
672 close to the correct time, whereas apps with low levels of traffic may see
673 events running much later than scheduled. If this is a problem, you can use
674 a real cron entry that simply hits your application at the desired time.
676 0 * * * * wget -q http://www.myapp.com/
678 Events which consume a lot of time will slow the request processing for the
679 user who triggers the event. For these types of events, you should use
680 auto_run => 0 or manual event triggering.
684 The plugin only checks once per minute if any events need to be run, so the
685 overhead on each request is minimal. On my test server, the difference
686 between running with Scheduler and without was only around 0.02% (0.004
689 Of course, when a scheduled event runs, performance will depend on what's
690 being run in the event.
696 Schedule is a class method for adding scheduled events. See the
697 L<"/SCHEDULING"> section for more information.
699 =head2 scheduler_state
701 The current state of all scheduled events is available in an easy-to-use
702 format by calling $c->scheduler_state. You can use this data to build an
703 admin view into the scheduling engine, for example. This same data is also
704 displayed on the Catalyst debug screen.
706 This method returns an array reference containing a hash reference for each
711 'last_run' => '2005-12-29 16:29:33 EST',
715 'next_run' => '2005-12-30 00:00:00 EST',
716 'event' => '/cron/session_cleanup'
721 'next_run' => '2005-12-30 00:00:00 EST',
722 'event' => '/cron/build_rss'
726 =head1 INTERNAL METHODS
728 The following methods are extended by this plugin.
734 The main scheduling logic takes place during the dispatch phase.
738 On the Catalyst debug screen, all scheduled events are displayed along with
739 the next time they will be executed.
751 Andy Grundman, <andy@hybridized.org>
755 This program is free software, you can redistribute it and/or modify it
756 under the same terms as Perl itself.