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