includes/job/JobQueueGroup.php

   1 <?php
   2 /**
   3  * Job queue base code.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @author Aaron Schulz
  22  */
  23
  24 /**
  25  * Class to handle enqueueing of background jobs
  26  *
  27  * @ingroup JobQueue
  28  * @since 1.21
  29  */
  30 class JobQueueGroup {
  31         /** @var Array */
  32         protected static $instances = array();
  33
  34         /** @var ProcessCacheLRU */
  35         protected $cache;
  36
  37         protected $wiki; // string; wiki ID
  38
  39         /** @var array Map of (bucket => (queue => JobQueue, types => list of types) */
  40         protected $coalescedQueues;
  41
  42         const TYPE_DEFAULT = 1; // integer; jobs popped by default
  43         const TYPE_ANY = 2; // integer; any job
  44
  45         const USE_CACHE = 1; // integer; use process or persistent cache
  46         const USE_PRIORITY = 2; // integer; respect deprioritization
  47
  48         const PROC_CACHE_TTL = 15; // integer; seconds
  49
  50         const CACHE_VERSION = 1; // integer; cache version
  51
  52         /**
  53          * @param string $wiki Wiki ID
  54          */
  55         protected function __construct( $wiki ) {
  56                 $this->wiki = $wiki;
  57                 $this->cache = new ProcessCacheLRU( 10 );
  58         }
  59
  60         /**
  61          * @param string $wiki Wiki ID
  62          * @return JobQueueGroup
  63          */
  64         public static function singleton( $wiki = false ) {
  65                 $wiki = ( $wiki === false ) ? wfWikiID() : $wiki;
  66                 if ( !isset( self::$instances[$wiki] ) ) {
  67                         self::$instances[$wiki] = new self( $wiki );
  68                 }
  69
  70                 return self::$instances[$wiki];
  71         }
  72
  73         /**
  74          * Destroy the singleton instances
  75          *
  76          * @return void
  77          */
  78         public static function destroySingletons() {
  79                 self::$instances = array();
  80         }
  81
  82         /**
  83          * Get the job queue object for a given queue type
  84          *
  85          * @param $type string
  86          * @return JobQueue
  87          */
  88         public function get( $type ) {
  89                 global $wgJobTypeConf;
  90
  91                 $conf = array( 'wiki' => $this->wiki, 'type' => $type );
  92                 if ( isset( $wgJobTypeConf[$type] ) ) {
  93                         $conf = $conf + $wgJobTypeConf[$type];
  94                 } else {
  95                         $conf = $conf + $wgJobTypeConf['default'];
  96                 }
  97
  98                 return JobQueue::factory( $conf );
  99         }
 100
 101         /**
 102          * Insert jobs into the respective queues of with the belong.
 103          *
 104          * This inserts the jobs into the queue specified by $wgJobTypeConf
 105          * and updates the aggregate job queue information cache as needed.
 106          *
 107          * @param $jobs Job|array A single Job or a list of Jobs
 108          * @throws MWException
 109          * @return bool
 110          */
 111         public function push( $jobs ) {
 112                 $jobs = is_array( $jobs ) ? $jobs : array( $jobs );
 113
 114                 $jobsByType = array(); // (job type => list of jobs)
 115                 foreach ( $jobs as $job ) {
 116                         if ( $job instanceof Job ) {
 117                                 $jobsByType[$job->getType()][] = $job;
 118                         } else {
 119                                 throw new MWException( "Attempted to push a non-Job object into a queue." );
 120                         }
 121                 }
 122
 123                 $ok = true;
 124                 foreach ( $jobsByType as $type => $jobs ) {
 125                         if ( $this->get( $type )->push( $jobs ) ) {
 126                                 JobQueueAggregator::singleton()->notifyQueueNonEmpty( $this->wiki, $type );
 127                         } else {
 128                                 $ok = false;
 129                         }
 130                 }
 131
 132                 if ( $this->cache->has( 'queues-ready', 'list' ) ) {
 133                         $list = $this->cache->get( 'queues-ready', 'list' );
 134                         if ( count( array_diff( array_keys( $jobsByType ), $list ) ) ) {
 135                                 $this->cache->clear( 'queues-ready' );
 136                         }
 137                 }
 138
 139                 return $ok;
 140         }
 141
 142         /**
 143          * Pop a job off one of the job queues
 144          *
 145          * This pops a job off a queue as specified by $wgJobTypeConf and
 146          * updates the aggregate job queue information cache as needed.
 147          *
 148          * @param $qtype integer|string JobQueueGroup::TYPE_DEFAULT or type string
 149          * @param $flags integer Bitfield of JobQueueGroup::USE_* constants
 150          * @return Job|bool Returns false on failure
 151          */
 152         public function pop( $qtype = self::TYPE_DEFAULT, $flags = 0 ) {
 153                 if ( is_string( $qtype ) ) { // specific job type
 154                         if ( ( $flags & self::USE_PRIORITY ) && $this->isQueueDeprioritized( $qtype ) ) {
 155                                 return false; // back off
 156                         }
 157                         $job = $this->get( $qtype )->pop();
 158                         if ( !$job ) {
 159                                 JobQueueAggregator::singleton()->notifyQueueEmpty( $this->wiki, $qtype );
 160                         }
 161
 162                         return $job;
 163                 } else { // any job in the "default" jobs types
 164                         if ( $flags & self::USE_CACHE ) {
 165                                 if ( !$this->cache->has( 'queues-ready', 'list', self::PROC_CACHE_TTL ) ) {
 166                                         $this->cache->set( 'queues-ready', 'list', $this->getQueuesWithJobs() );
 167                                 }
 168                                 $types = $this->cache->get( 'queues-ready', 'list' );
 169                         } else {
 170                                 $types = $this->getQueuesWithJobs();
 171                         }
 172
 173                         if ( $qtype == self::TYPE_DEFAULT ) {
 174                                 $types = array_intersect( $types, $this->getDefaultQueueTypes() );
 175                         }
 176                         shuffle( $types ); // avoid starvation
 177
 178                         foreach ( $types as $type ) { // for each queue...
 179                                 if ( ( $flags & self::USE_PRIORITY ) && $this->isQueueDeprioritized( $type ) ) {
 180                                         continue; // back off
 181                                 }
 182                                 $job = $this->get( $type )->pop();
 183                                 if ( $job ) { // found
 184                                         return $job;
 185                                 } else { // not found
 186                                         JobQueueAggregator::singleton()->notifyQueueEmpty( $this->wiki, $type );
 187                                         $this->cache->clear( 'queues-ready' );
 188                                 }
 189                         }
 190
 191                         return false; // no jobs found
 192                 }
 193         }
 194
 195         /**
 196          * Acknowledge that a job was completed
 197          *
 198          * @param $job Job
 199          * @return bool
 200          */
 201         public function ack( Job $job ) {
 202                 return $this->get( $job->getType() )->ack( $job );
 203         }
 204
 205         /**
 206          * Register the "root job" of a given job into the queue for de-duplication.
 207          * This should only be called right *after* all the new jobs have been inserted.
 208          *
 209          * @param $job Job
 210          * @return bool
 211          */
 212         public function deduplicateRootJob( Job $job ) {
 213                 return $this->get( $job->getType() )->deduplicateRootJob( $job );
 214         }
 215
 216         /**
 217          * Wait for any slaves or backup queue servers to catch up.
 218          *
 219          * This does nothing for certain queue classes.
 220          *
 221          * @return void
 222          * @throws MWException
 223          */
 224         public function waitForBackups() {
 225                 global $wgJobTypeConf;
 226
 227                 wfProfileIn( __METHOD__ );
 228                 // Try to avoid doing this more than once per queue storage medium
 229                 foreach ( $wgJobTypeConf as $type => $conf ) {
 230                         $this->get( $type )->waitForBackups();
 231                 }
 232                 wfProfileOut( __METHOD__ );
 233         }
 234
 235         /**
 236          * Get the list of queue types
 237          *
 238          * @return array List of strings
 239          */
 240         public function getQueueTypes() {
 241                 return array_keys( $this->getCachedConfigVar( 'wgJobClasses' ) );
 242         }
 243
 244         /**
 245          * Get the list of default queue types
 246          *
 247          * @return array List of strings
 248          */
 249         public function getDefaultQueueTypes() {
 250                 global $wgJobTypesExcludedFromDefaultQueue;
 251
 252                 return array_diff( $this->getQueueTypes(), $wgJobTypesExcludedFromDefaultQueue );
 253         }
 254
 255         /**
 256          * Get the list of job types that have non-empty queues
 257          *
 258          * @return Array List of job types that have non-empty queues
 259          */
 260         public function getQueuesWithJobs() {
 261                 $types = array();
 262                 foreach ( $this->getCoalescedQueues() as $info ) {
 263                         $nonEmpty = $info['queue']->getSiblingQueuesWithJobs( $this->getQueueTypes() );
 264                         if ( is_array( $nonEmpty ) ) { // batching features supported
 265                                 $types = array_merge( $types, $nonEmpty );
 266                         } else { // we have to go through the queues in the bucket one-by-one
 267                                 foreach ( $info['types'] as $type ) {
 268                                         if ( !$this->get( $type )->isEmpty() ) {
 269                                                 $types[] = $type;
 270                                         }
 271                                 }
 272                         }
 273                 }
 274
 275                 return $types;
 276         }
 277
 278         /**
 279          * Get the size of the queus for a list of job types
 280          *
 281          * @return Array Map of (job type => size)
 282          */
 283         public function getQueueSizes() {
 284                 $sizeMap = array();
 285                 foreach ( $this->getCoalescedQueues() as $info ) {
 286                         $sizes = $info['queue']->getSiblingQueueSizes( $this->getQueueTypes() );
 287                         if ( is_array( $sizes ) ) { // batching features supported
 288                                 $sizeMap = $sizeMap + $sizes;
 289                         } else { // we have to go through the queues in the bucket one-by-one
 290                                 foreach ( $info['types'] as $type ) {
 291                                         $sizeMap[$type] = $this->get( $type )->getSize();
 292                                 }
 293                         }
 294                 }
 295
 296                 return $sizeMap;
 297         }
 298
 299         /**
 300          * @return array
 301          */
 302         protected function getCoalescedQueues() {
 303                 global $wgJobTypeConf;
 304
 305                 if ( $this->coalescedQueues === null ) {
 306                         $this->coalescedQueues = array();
 307                         foreach ( $wgJobTypeConf as $type => $conf ) {
 308                                 $queue = JobQueue::factory(
 309                                         array( 'wiki' => $this->wiki, 'type' => 'null' ) + $conf );
 310                                 $loc = $queue->getCoalesceLocationInternal();
 311                                 if ( !isset( $this->coalescedQueues[$loc] ) ) {
 312                                         $this->coalescedQueues[$loc]['queue'] = $queue;
 313                                         $this->coalescedQueues[$loc]['types'] = array();
 314                                 }
 315                                 if ( $type === 'default' ) {
 316                                         $this->coalescedQueues[$loc]['types'] = array_merge(
 317                                                 $this->coalescedQueues[$loc]['types'],
 318                                                 array_diff( $this->getQueueTypes(), array_keys( $wgJobTypeConf ) )
 319                                         );
 320                                 } else {
 321                                         $this->coalescedQueues[$loc]['types'][] = $type;
 322                                 }
 323                         }
 324                 }
 325
 326                 return $this->coalescedQueues;
 327         }
 328
 329         /**
 330          * Check if jobs should not be popped of a queue right now.
 331          * This is only used for performance, such as to avoid spamming
 332          * the queue with many sub-jobs before they actually get run.
 333          *
 334          * @param $type string
 335          * @return bool
 336          */
 337         public function isQueueDeprioritized( $type ) {
 338                 if ( $this->cache->has( 'isDeprioritized', $type, 5 ) ) {
 339                         return $this->cache->get( 'isDeprioritized', $type );
 340                 }
 341                 if ( $type === 'refreshLinks2' ) {
 342                         // Don't keep converting refreshLinks2 => refreshLinks jobs if the
 343                         // later jobs have not been done yet. This helps throttle queue spam.
 344                         $deprioritized = !$this->get( 'refreshLinks' )->isEmpty();
 345                         $this->cache->set( 'isDeprioritized', $type, $deprioritized );
 346
 347                         return $deprioritized;
 348                 }
 349
 350                 return false;
 351         }
 352
 353         /**
 354          * Execute any due periodic queue maintenance tasks for all queues.
 355          *
 356          * A task is "due" if the time ellapsed since the last run is greater than
 357          * the defined run period. Concurrent calls to this function will cause tasks
 358          * to be attempted twice, so they may need their own methods of mutual exclusion.
 359          *
 360          * @return integer Number of tasks run
 361          */
 362         public function executeReadyPeriodicTasks() {
 363                 global $wgMemc;
 364
 365                 list( $db, $prefix ) = wfSplitWikiID( $this->wiki );
 366                 $key = wfForeignMemcKey( $db, $prefix, 'jobqueuegroup', 'taskruns', 'v1' );
 367                 $lastRuns = $wgMemc->get( $key ); // (queue => task => UNIX timestamp)
 368
 369                 $count = 0;
 370                 $tasksRun = array(); // (queue => task => UNIX timestamp)
 371                 foreach ( $this->getQueueTypes() as $type ) {
 372                         $queue = $this->get( $type );
 373                         foreach ( $queue->getPeriodicTasks() as $task => $definition ) {
 374                                 if ( $definition['period'] <= 0 ) {
 375                                         continue; // disabled
 376                                 } elseif ( !isset( $lastRuns[$type][$task] )
 377                                         || $lastRuns[$type][$task] < ( time() - $definition['period'] )
 378                                 ) {
 379                                         try {
 380                                                 if ( call_user_func( $definition['callback'] ) !== null ) {
 381                                                         $tasksRun[$type][$task] = time();
 382                                                         ++$count;
 383                                                 }
 384                                         } catch ( JobQueueError $e ) {
 385                                                 MWExceptionHandler::logException( $e );
 386                                         }
 387                                 }
 388                         }
 389                 }
 390
 391                 $wgMemc->merge( $key, function ( $cache, $key, $lastRuns ) use ( $tasksRun ) {
 392                         if ( is_array( $lastRuns ) ) {
 393                                 foreach ( $tasksRun as $type => $tasks ) {
 394                                         foreach ( $tasks as $task => $timestamp ) {
 395                                                 if ( !isset( $lastRuns[$type][$task] )
 396                                                         || $timestamp > $lastRuns[$type][$task]
 397                                                 ) {
 398                                                         $lastRuns[$type][$task] = $timestamp;
 399                                                 }
 400                                         }
 401                                 }
 402                         } else {
 403                                 $lastRuns = $tasksRun;
 404                         }
 405
 406                         return $lastRuns;
 407                 } );
 408
 409                 return $count;
 410         }
 411
 412         /**
 413          * @param $name string
 414          * @return mixed
 415          */
 416         private function getCachedConfigVar( $name ) {
 417                 global $wgConf, $wgMemc;
 418
 419                 if ( $this->wiki === wfWikiID() ) {
 420                         return $GLOBALS[$name]; // common case
 421                 } else {
 422                         list( $db, $prefix ) = wfSplitWikiID( $this->wiki );
 423                         $key = wfForeignMemcKey( $db, $prefix, 'configvalue', $name );
 424                         $value = $wgMemc->get( $key ); // ('v' => ...) or false
 425                         if ( is_array( $value ) ) {
 426                                 return $value['v'];
 427                         } else {
 428                                 $value = $wgConf->getConfig( $this->wiki, $name );
 429                                 $wgMemc->set( $key, array( 'v' => $value ), 86400 + mt_rand( 0, 86400 ) );
 430
 431                                 return $value;
 432                         }
 433                 }
 434         }
 435 }