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         /** @var string Wiki ID */
  38         protected $wiki;
  39
  40         /** @var array Map of (bucket => (queue => JobQueue, types => list of types) */
  41         protected $coalescedQueues;
  42
  43         const TYPE_DEFAULT = 1; // integer; jobs popped by default
  44         const TYPE_ANY = 2; // integer; any job
  45
  46         const USE_CACHE = 1; // integer; use process or persistent cache
  47         const USE_PRIORITY = 2; // integer; respect deprioritization
  48
  49         const PROC_CACHE_TTL = 15; // integer; seconds
  50
  51         const CACHE_VERSION = 1; // integer; cache version
  52
  53         /**
  54          * @param string $wiki Wiki ID
  55          */
  56         protected function __construct( $wiki ) {
  57                 $this->wiki = $wiki;
  58                 $this->cache = new ProcessCacheLRU( 10 );
  59         }
  60
  61         /**
  62          * @param bool|string $wiki Wiki ID
  63          * @return JobQueueGroup
  64          */
  65         public static function singleton( $wiki = false ) {
  66                 $wiki = ( $wiki === false ) ? wfWikiID() : $wiki;
  67                 if ( !isset( self::$instances[$wiki] ) ) {
  68                         self::$instances[$wiki] = new self( $wiki );
  69                 }
  70
  71                 return self::$instances[$wiki];
  72         }
  73
  74         /**
  75          * Destroy the singleton instances
  76          *
  77          * @return void
  78          */
  79         public static function destroySingletons() {
  80                 self::$instances = array();
  81         }
  82
  83         /**
  84          * Get the job queue object for a given queue type
  85          *
  86          * @param string $type
  87          * @return JobQueue
  88          */
  89         public function get( $type ) {
  90                 global $wgJobTypeConf;
  91
  92                 $conf = array( 'wiki' => $this->wiki, 'type' => $type );
  93                 if ( isset( $wgJobTypeConf[$type] ) ) {
  94                         $conf = $conf + $wgJobTypeConf[$type];
  95                 } else {
  96                         $conf = $conf + $wgJobTypeConf['default'];
  97                 }
  98
  99                 return JobQueue::factory( $conf );
 100         }
 101
 102         /**
 103          * Insert jobs into the respective queues of with the belong.
 104          *
 105          * This inserts the jobs into the queue specified by $wgJobTypeConf
 106          * and updates the aggregate job queue information cache as needed.
 107          *
 108          * @param Job|array $jobs A single Job or a list of Jobs
 109          * @throws MWException
 110          * @return bool
 111          */
 112         public function push( $jobs ) {
 113                 $jobs = is_array( $jobs ) ? $jobs : array( $jobs );
 114                 if ( !count( $jobs ) ) {
 115                         return true;
 116                 }
 117
 118                 $jobsByType = array(); // (job type => list of jobs)
 119                 foreach ( $jobs as $job ) {
 120                         if ( $job instanceof Job ) {
 121                                 $jobsByType[$job->getType()][] = $job;
 122                         } else {
 123                                 throw new MWException( "Attempted to push a non-Job object into a queue." );
 124                         }
 125                 }
 126
 127                 $ok = true;
 128                 foreach ( $jobsByType as $type => $jobs ) {
 129                         if ( $this->get( $type )->push( $jobs ) ) {
 130                                 JobQueueAggregator::singleton()->notifyQueueNonEmpty( $this->wiki, $type );
 131                         } else {
 132                                 $ok = false;
 133                         }
 134                 }
 135
 136                 if ( $this->cache->has( 'queues-ready', 'list' ) ) {
 137                         $list = $this->cache->get( 'queues-ready', 'list' );
 138                         if ( count( array_diff( array_keys( $jobsByType ), $list ) ) ) {
 139                                 $this->cache->clear( 'queues-ready' );
 140                         }
 141                 }
 142
 143                 return $ok;
 144         }
 145
 146         /**
 147          * Pop a job off one of the job queues
 148          *
 149          * This pops a job off a queue as specified by $wgJobTypeConf and
 150          * updates the aggregate job queue information cache as needed.
 151          *
 152          * @param int|string $qtype JobQueueGroup::TYPE_DEFAULT or type string
 153          * @param int $flags Bitfield of JobQueueGroup::USE_* constants
 154          * @return Job|bool Returns false on failure
 155          */
 156         public function pop( $qtype = self::TYPE_DEFAULT, $flags = 0 ) {
 157                 if ( is_string( $qtype ) ) { // specific job type
 158                         if ( ( $flags & self::USE_PRIORITY ) && $this->isQueueDeprioritized( $qtype ) ) {
 159                                 return false; // back off
 160                         }
 161                         $job = $this->get( $qtype )->pop();
 162                         if ( !$job ) {
 163                                 JobQueueAggregator::singleton()->notifyQueueEmpty( $this->wiki, $qtype );
 164                         }
 165
 166                         return $job;
 167                 } else { // any job in the "default" jobs types
 168                         if ( $flags & self::USE_CACHE ) {
 169                                 if ( !$this->cache->has( 'queues-ready', 'list', self::PROC_CACHE_TTL ) ) {
 170                                         $this->cache->set( 'queues-ready', 'list', $this->getQueuesWithJobs() );
 171                                 }
 172                                 $types = $this->cache->get( 'queues-ready', 'list' );
 173                         } else {
 174                                 $types = $this->getQueuesWithJobs();
 175                         }
 176
 177                         if ( $qtype == self::TYPE_DEFAULT ) {
 178                                 $types = array_intersect( $types, $this->getDefaultQueueTypes() );
 179                         }
 180                         shuffle( $types ); // avoid starvation
 181
 182                         foreach ( $types as $type ) { // for each queue...
 183                                 if ( ( $flags & self::USE_PRIORITY ) && $this->isQueueDeprioritized( $type ) ) {
 184                                         continue; // back off
 185                                 }
 186                                 $job = $this->get( $type )->pop();
 187                                 if ( $job ) { // found
 188                                         return $job;
 189                                 } else { // not found
 190                                         JobQueueAggregator::singleton()->notifyQueueEmpty( $this->wiki, $type );
 191                                         $this->cache->clear( 'queues-ready' );
 192                                 }
 193                         }
 194
 195                         return false; // no jobs found
 196                 }
 197         }
 198
 199         /**
 200          * Acknowledge that a job was completed
 201          *
 202          * @param Job $job
 203          * @return bool
 204          */
 205         public function ack( Job $job ) {
 206                 return $this->get( $job->getType() )->ack( $job );
 207         }
 208
 209         /**
 210          * Register the "root job" of a given job into the queue for de-duplication.
 211          * This should only be called right *after* all the new jobs have been inserted.
 212          *
 213          * @param Job $job
 214          * @return bool
 215          */
 216         public function deduplicateRootJob( Job $job ) {
 217                 return $this->get( $job->getType() )->deduplicateRootJob( $job );
 218         }
 219
 220         /**
 221          * Wait for any slaves or backup queue servers to catch up.
 222          *
 223          * This does nothing for certain queue classes.
 224          *
 225          * @return void
 226          * @throws MWException
 227          */
 228         public function waitForBackups() {
 229                 global $wgJobTypeConf;
 230
 231                 wfProfileIn( __METHOD__ );
 232                 // Try to avoid doing this more than once per queue storage medium
 233                 foreach ( $wgJobTypeConf as $type => $conf ) {
 234                         $this->get( $type )->waitForBackups();
 235                 }
 236                 wfProfileOut( __METHOD__ );
 237         }
 238
 239         /**
 240          * Get the list of queue types
 241          *
 242          * @return array List of strings
 243          */
 244         public function getQueueTypes() {
 245                 return array_keys( $this->getCachedConfigVar( 'wgJobClasses' ) );
 246         }
 247
 248         /**
 249          * Get the list of default queue types
 250          *
 251          * @return array List of strings
 252          */
 253         public function getDefaultQueueTypes() {
 254                 global $wgJobTypesExcludedFromDefaultQueue;
 255
 256                 return array_diff( $this->getQueueTypes(), $wgJobTypesExcludedFromDefaultQueue );
 257         }
 258
 259         /**
 260          * Get the list of job types that have non-empty queues
 261          *
 262          * @return array List of job types that have non-empty queues
 263          */
 264         public function getQueuesWithJobs() {
 265                 $types = array();
 266                 foreach ( $this->getCoalescedQueues() as $info ) {
 267                         $nonEmpty = $info['queue']->getSiblingQueuesWithJobs( $this->getQueueTypes() );
 268                         if ( is_array( $nonEmpty ) ) { // batching features supported
 269                                 $types = array_merge( $types, $nonEmpty );
 270                         } else { // we have to go through the queues in the bucket one-by-one
 271                                 foreach ( $info['types'] as $type ) {
 272                                         if ( !$this->get( $type )->isEmpty() ) {
 273                                                 $types[] = $type;
 274                                         }
 275                                 }
 276                         }
 277                 }
 278
 279                 return $types;
 280         }
 281
 282         /**
 283          * Get the size of the queus for a list of job types
 284          *
 285          * @return array Map of (job type => size)
 286          */
 287         public function getQueueSizes() {
 288                 $sizeMap = array();
 289                 foreach ( $this->getCoalescedQueues() as $info ) {
 290                         $sizes = $info['queue']->getSiblingQueueSizes( $this->getQueueTypes() );
 291                         if ( is_array( $sizes ) ) { // batching features supported
 292                                 $sizeMap = $sizeMap + $sizes;
 293                         } else { // we have to go through the queues in the bucket one-by-one
 294                                 foreach ( $info['types'] as $type ) {
 295                                         $sizeMap[$type] = $this->get( $type )->getSize();
 296                                 }
 297                         }
 298                 }
 299
 300                 return $sizeMap;
 301         }
 302
 303         /**
 304          * @return array
 305          */
 306         protected function getCoalescedQueues() {
 307                 global $wgJobTypeConf;
 308
 309                 if ( $this->coalescedQueues === null ) {
 310                         $this->coalescedQueues = array();
 311                         foreach ( $wgJobTypeConf as $type => $conf ) {
 312                                 $queue = JobQueue::factory(
 313                                         array( 'wiki' => $this->wiki, 'type' => 'null' ) + $conf );
 314                                 $loc = $queue->getCoalesceLocationInternal();
 315                                 if ( !isset( $this->coalescedQueues[$loc] ) ) {
 316                                         $this->coalescedQueues[$loc]['queue'] = $queue;
 317                                         $this->coalescedQueues[$loc]['types'] = array();
 318                                 }
 319                                 if ( $type === 'default' ) {
 320                                         $this->coalescedQueues[$loc]['types'] = array_merge(
 321                                                 $this->coalescedQueues[$loc]['types'],
 322                                                 array_diff( $this->getQueueTypes(), array_keys( $wgJobTypeConf ) )
 323                                         );
 324                                 } else {
 325                                         $this->coalescedQueues[$loc]['types'][] = $type;
 326                                 }
 327                         }
 328                 }
 329
 330                 return $this->coalescedQueues;
 331         }
 332
 333         /**
 334          * Check if jobs should not be popped of a queue right now.
 335          * This is only used for performance, such as to avoid spamming
 336          * the queue with many sub-jobs before they actually get run.
 337          *
 338          * @param string $type
 339          * @return bool
 340          */
 341         public function isQueueDeprioritized( $type ) {
 342                 if ( $this->cache->has( 'isDeprioritized', $type, 5 ) ) {
 343                         return $this->cache->get( 'isDeprioritized', $type );
 344                 }
 345                 if ( $type === 'refreshLinks2' ) {
 346                         // Don't keep converting refreshLinksPartition => refreshLinks jobs if the
 347                         // later jobs have not been done yet. This helps throttle queue spam.
 348                         // @TODO: this is mostly a WMF-specific hack and should be removed when
 349                         // refreshLinks2 jobs are drained.
 350                         $deprioritized = $this->get( 'refreshLinks' )->getSize() > 10000;
 351                         $this->cache->set( 'isDeprioritized', $type, $deprioritized );
 352
 353                         return $deprioritized;
 354                 }
 355
 356                 return false;
 357         }
 358
 359         /**
 360          * Execute any due periodic queue maintenance tasks for all queues.
 361          *
 362          * A task is "due" if the time ellapsed since the last run is greater than
 363          * the defined run period. Concurrent calls to this function will cause tasks
 364          * to be attempted twice, so they may need their own methods of mutual exclusion.
 365          *
 366          * @return int Number of tasks run
 367          */
 368         public function executeReadyPeriodicTasks() {
 369                 global $wgMemc;
 370
 371                 list( $db, $prefix ) = wfSplitWikiID( $this->wiki );
 372                 $key = wfForeignMemcKey( $db, $prefix, 'jobqueuegroup', 'taskruns', 'v1' );
 373                 $lastRuns = $wgMemc->get( $key ); // (queue => task => UNIX timestamp)
 374
 375                 $count = 0;
 376                 $tasksRun = array(); // (queue => task => UNIX timestamp)
 377                 foreach ( $this->getQueueTypes() as $type ) {
 378                         $queue = $this->get( $type );
 379                         foreach ( $queue->getPeriodicTasks() as $task => $definition ) {
 380                                 if ( $definition['period'] <= 0 ) {
 381                                         continue; // disabled
 382                                 } elseif ( !isset( $lastRuns[$type][$task] )
 383                                         || $lastRuns[$type][$task] < ( time() - $definition['period'] )
 384                                 ) {
 385                                         try {
 386                                                 if ( call_user_func( $definition['callback'] ) !== null ) {
 387                                                         $tasksRun[$type][$task] = time();
 388                                                         ++$count;
 389                                                 }
 390                                         } catch ( JobQueueError $e ) {
 391                                                 MWExceptionHandler::logException( $e );
 392                                         }
 393                                 }
 394                         }
 395                 }
 396
 397                 $wgMemc->merge( $key, function ( $cache, $key, $lastRuns ) use ( $tasksRun ) {
 398                         if ( is_array( $lastRuns ) ) {
 399                                 foreach ( $tasksRun as $type => $tasks ) {
 400                                         foreach ( $tasks as $task => $timestamp ) {
 401                                                 if ( !isset( $lastRuns[$type][$task] )
 402                                                         || $timestamp > $lastRuns[$type][$task]
 403                                                 ) {
 404                                                         $lastRuns[$type][$task] = $timestamp;
 405                                                 }
 406                                         }
 407                                 }
 408                         } else {
 409                                 $lastRuns = $tasksRun;
 410                         }
 411
 412                         return $lastRuns;
 413                 } );
 414
 415                 return $count;
 416         }
 417
 418         /**
 419          * @param $name string
 420          * @return mixed
 421          */
 422         private function getCachedConfigVar( $name ) {
 423                 global $wgConf, $wgMemc;
 424
 425                 if ( $this->wiki === wfWikiID() ) {
 426                         return $GLOBALS[$name]; // common case
 427                 } else {
 428                         list( $db, $prefix ) = wfSplitWikiID( $this->wiki );
 429                         $key = wfForeignMemcKey( $db, $prefix, 'configvalue', $name );
 430                         $value = $wgMemc->get( $key ); // ('v' => ...) or false
 431                         if ( is_array( $value ) ) {
 432                                 return $value['v'];
 433                         } else {
 434                                 $value = $wgConf->getConfig( $this->wiki, $name );
 435                                 $wgMemc->set( $key, array( 'v' => $value ), 86400 + mt_rand( 0, 86400 ) );
 436
 437                                 return $value;
 438                         }
 439                 }
 440         }
 441 }