includes/jobqueue/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  */
  22 use MediaWiki\MediaWikiServices;
  23
  24 /**
  25  * Class to handle enqueueing of background jobs
  26  *
  27  * @ingroup JobQueue
  28  * @since 1.21
  29  */
  30 class JobQueueGroup {
  31         /** @var JobQueueGroup[] */
  32         protected static $instances = [];
  33
  34         /** @var ProcessCacheLRU */
  35         protected $cache;
  36
  37         /** @var string Wiki domain ID */
  38         protected $domain;
  39         /** @var string|bool Read only rationale (or false if r/w) */
  40         protected $readOnlyReason;
  41         /** @var bool Whether the wiki is not recognized in configuration */
  42         protected $invalidDomain = false;
  43
  44         /** @var array Map of (bucket => (queue => JobQueue, types => list of types) */
  45         protected $coalescedQueues;
  46
  47         const TYPE_DEFAULT = 1; // integer; jobs popped by default
  48         const TYPE_ANY = 2; // integer; any job
  49
  50         const USE_CACHE = 1; // integer; use process or persistent cache
  51
  52         const PROC_CACHE_TTL = 15; // integer; seconds
  53
  54         const CACHE_VERSION = 1; // integer; cache version
  55
  56         /**
  57          * @param string $domain Wiki domain ID
  58          * @param string|bool $readOnlyReason Read-only reason or false
  59          */
  60         protected function __construct( $domain, $readOnlyReason ) {
  61                 $this->domain = $domain;
  62                 $this->readOnlyReason = $readOnlyReason;
  63                 $this->cache = new MapCacheLRU( 10 );
  64         }
  65
  66         /**
  67          * @param bool|string $domain Wiki domain ID
  68          * @return JobQueueGroup
  69          */
  70         public static function singleton( $domain = false ) {
  71                 global $wgLocalDatabases;
  72
  73                 if ( $domain === false ) {
  74                         $domain = WikiMap::getCurrentWikiDbDomain()->getId();
  75                 }
  76
  77                 if ( !isset( self::$instances[$domain] ) ) {
  78                         self::$instances[$domain] = new self( $domain, wfConfiguredReadOnlyReason() );
  79                         // Make sure jobs are not getting pushed to bogus wikis. This can confuse
  80                         // the job runner system into spawning endless RPC requests that fail (T171371).
  81                         $wikiId = WikiMap::getWikiIdFromDbDomain( $domain );
  82                         if (
  83                                 !WikiMap::isCurrentWikiDbDomain( $domain ) &&
  84                                 !in_array( $wikiId, $wgLocalDatabases )
  85                         ) {
  86                                 self::$instances[$domain]->invalidDomain = true;
  87                         }
  88                 }
  89
  90                 return self::$instances[$domain];
  91         }
  92
  93         /**
  94          * Destroy the singleton instances
  95          *
  96          * @return void
  97          */
  98         public static function destroySingletons() {
  99                 self::$instances = [];
 100         }
 101
 102         /**
 103          * Get the job queue object for a given queue type
 104          *
 105          * @param string $type
 106          * @return JobQueue
 107          */
 108         public function get( $type ) {
 109                 global $wgJobTypeConf;
 110
 111                 $conf = [ 'domain' => $this->domain, 'type' => $type ];
 112                 if ( isset( $wgJobTypeConf[$type] ) ) {
 113                         $conf = $conf + $wgJobTypeConf[$type];
 114                 } else {
 115                         $conf = $conf + $wgJobTypeConf['default'];
 116                 }
 117                 if ( !isset( $conf['readOnlyReason'] ) ) {
 118                         $conf['readOnlyReason'] = $this->readOnlyReason;
 119                 }
 120
 121                 $services = MediaWikiServices::getInstance();
 122                 $conf['stats'] = $services->getStatsdDataFactory();
 123                 $conf['wanCache'] = $services->getMainWANObjectCache();
 124                 $conf['stash'] = $services->getMainObjectStash();
 125
 126                 return JobQueue::factory( $conf );
 127         }
 128
 129         /**
 130          * Insert jobs into the respective queues of which they belong
 131          *
 132          * This inserts the jobs into the queue specified by $wgJobTypeConf
 133          * and updates the aggregate job queue information cache as needed.
 134          *
 135          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 136          * @throws InvalidArgumentException
 137          * @return void
 138          */
 139         public function push( $jobs ) {
 140                 global $wgJobTypesExcludedFromDefaultQueue;
 141
 142                 if ( $this->invalidDomain ) {
 143                         // Do not enqueue job that cannot be run (T171371)
 144                         $e = new LogicException( "Domain '{$this->domain}' is not recognized." );
 145                         MWExceptionHandler::logException( $e );
 146                         return;
 147                 }
 148
 149                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 150                 if ( $jobs === [] ) {
 151                         return;
 152                 }
 153
 154                 $this->assertValidJobs( $jobs );
 155
 156                 $jobsByType = []; // (job type => list of jobs)
 157                 foreach ( $jobs as $job ) {
 158                         $jobsByType[$job->getType()][] = $job;
 159                 }
 160
 161                 foreach ( $jobsByType as $type => $jobs ) {
 162                         $this->get( $type )->push( $jobs );
 163                 }
 164
 165                 if ( $this->cache->hasField( 'queues-ready', 'list' ) ) {
 166                         $list = $this->cache->getField( 'queues-ready', 'list' );
 167                         if ( count( array_diff( array_keys( $jobsByType ), $list ) ) ) {
 168                                 $this->cache->clear( 'queues-ready' );
 169                         }
 170                 }
 171
 172                 $cache = ObjectCache::getLocalClusterInstance();
 173                 $cache->set(
 174                         $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', self::TYPE_ANY ),
 175                         'true',
 176                         15
 177                 );
 178                 if ( array_diff( array_keys( $jobsByType ), $wgJobTypesExcludedFromDefaultQueue ) ) {
 179                         $cache->set(
 180                                 $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', self::TYPE_DEFAULT ),
 181                                 'true',
 182                                 15
 183                         );
 184                 }
 185         }
 186
 187         /**
 188          * Buffer jobs for insertion via push() or call it now if in CLI mode
 189          *
 190          * Note that pushLazyJobs() is registered as a deferred update just before
 191          * DeferredUpdates::doUpdates() in MediaWiki and JobRunner classes in order
 192          * to be executed as the very last deferred update (T100085, T154425).
 193          *
 194          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 195          * @return void
 196          * @since 1.26
 197          */
 198         public function lazyPush( $jobs ) {
 199                 if ( $this->invalidDomain ) {
 200                         // Do not enqueue job that cannot be run (T171371)
 201                         throw new LogicException( "Domain '{$this->domain}' is not recognized." );
 202                 }
 203
 204                 if ( PHP_SAPI === 'cli' || PHP_SAPI === 'phpdbg' ) {
 205                         $this->push( $jobs );
 206                         return;
 207                 }
 208
 209                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 210
 211                 // Throw errors now instead of on push(), when other jobs may be buffered
 212                 $this->assertValidJobs( $jobs );
 213
 214                 DeferredUpdates::addUpdate( new JobQueueEnqueueUpdate( $this->domain, $jobs ) );
 215         }
 216
 217         /**
 218          * Push all jobs buffered via lazyPush() into their respective queues
 219          *
 220          * @return void
 221          * @since 1.26
 222          * @deprecated Since 1.33 Not needed anymore
 223          */
 224         public static function pushLazyJobs() {
 225                 wfDeprecated( __METHOD__, '1.33' );
 226         }
 227
 228         /**
 229          * Pop a job off one of the job queues
 230          *
 231          * This pops a job off a queue as specified by $wgJobTypeConf and
 232          * updates the aggregate job queue information cache as needed.
 233          *
 234          * @param int|string $qtype JobQueueGroup::TYPE_* constant or job type string
 235          * @param int $flags Bitfield of JobQueueGroup::USE_* constants
 236          * @param array $blacklist List of job types to ignore
 237          * @return Job|bool Returns false on failure
 238          */
 239         public function pop( $qtype = self::TYPE_DEFAULT, $flags = 0, array $blacklist = [] ) {
 240                 global $wgJobClasses;
 241
 242                 $job = false;
 243
 244                 if ( !WikiMap::isCurrentWikiDbDomain( $this->domain ) ) {
 245                         throw new JobQueueError(
 246                                 "Cannot pop '{$qtype}' job off foreign '{$this->domain}' wiki queue." );
 247                 } elseif ( is_string( $qtype ) && !isset( $wgJobClasses[$qtype] ) ) {
 248                         // Do not pop jobs if there is no class for the queue type
 249                         throw new JobQueueError( "Unrecognized job type '$qtype'." );
 250                 }
 251
 252                 if ( is_string( $qtype ) ) { // specific job type
 253                         if ( !in_array( $qtype, $blacklist ) ) {
 254                                 $job = $this->get( $qtype )->pop();
 255                         }
 256                 } else { // any job in the "default" jobs types
 257                         if ( $flags & self::USE_CACHE ) {
 258                                 if ( !$this->cache->hasField( 'queues-ready', 'list', self::PROC_CACHE_TTL ) ) {
 259                                         $this->cache->setField( 'queues-ready', 'list', $this->getQueuesWithJobs() );
 260                                 }
 261                                 $types = $this->cache->getField( 'queues-ready', 'list' );
 262                         } else {
 263                                 $types = $this->getQueuesWithJobs();
 264                         }
 265
 266                         if ( $qtype == self::TYPE_DEFAULT ) {
 267                                 $types = array_intersect( $types, $this->getDefaultQueueTypes() );
 268                         }
 269
 270                         $types = array_diff( $types, $blacklist ); // avoid selected types
 271                         shuffle( $types ); // avoid starvation
 272
 273                         foreach ( $types as $type ) { // for each queue...
 274                                 $job = $this->get( $type )->pop();
 275                                 if ( $job ) { // found
 276                                         break;
 277                                 } else { // not found
 278                                         $this->cache->clear( 'queues-ready' );
 279                                 }
 280                         }
 281                 }
 282
 283                 return $job;
 284         }
 285
 286         /**
 287          * Acknowledge that a job was completed
 288          *
 289          * @param Job $job
 290          * @return void
 291          */
 292         public function ack( Job $job ) {
 293                 $this->get( $job->getType() )->ack( $job );
 294         }
 295
 296         /**
 297          * Register the "root job" of a given job into the queue for de-duplication.
 298          * This should only be called right *after* all the new jobs have been inserted.
 299          *
 300          * @param Job $job
 301          * @return bool
 302          */
 303         public function deduplicateRootJob( Job $job ) {
 304                 return $this->get( $job->getType() )->deduplicateRootJob( $job );
 305         }
 306
 307         /**
 308          * Wait for any replica DBs or backup queue servers to catch up.
 309          *
 310          * This does nothing for certain queue classes.
 311          *
 312          * @return void
 313          */
 314         public function waitForBackups() {
 315                 global $wgJobTypeConf;
 316
 317                 // Try to avoid doing this more than once per queue storage medium
 318                 foreach ( $wgJobTypeConf as $type => $conf ) {
 319                         $this->get( $type )->waitForBackups();
 320                 }
 321         }
 322
 323         /**
 324          * Get the list of queue types
 325          *
 326          * @return array List of strings
 327          */
 328         public function getQueueTypes() {
 329                 return array_keys( $this->getCachedConfigVar( 'wgJobClasses' ) );
 330         }
 331
 332         /**
 333          * Get the list of default queue types
 334          *
 335          * @return array List of strings
 336          */
 337         public function getDefaultQueueTypes() {
 338                 global $wgJobTypesExcludedFromDefaultQueue;
 339
 340                 return array_diff( $this->getQueueTypes(), $wgJobTypesExcludedFromDefaultQueue );
 341         }
 342
 343         /**
 344          * Check if there are any queues with jobs (this is cached)
 345          *
 346          * @param int $type JobQueueGroup::TYPE_* constant
 347          * @return bool
 348          * @since 1.23
 349          */
 350         public function queuesHaveJobs( $type = self::TYPE_ANY ) {
 351                 $cache = ObjectCache::getLocalClusterInstance();
 352                 $key = $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', $type );
 353
 354                 $value = $cache->get( $key );
 355                 if ( $value === false ) {
 356                         $queues = $this->getQueuesWithJobs();
 357                         if ( $type == self::TYPE_DEFAULT ) {
 358                                 $queues = array_intersect( $queues, $this->getDefaultQueueTypes() );
 359                         }
 360                         $value = count( $queues ) ? 'true' : 'false';
 361                         $cache->add( $key, $value, 15 );
 362                 }
 363
 364                 return ( $value === 'true' );
 365         }
 366
 367         /**
 368          * Get the list of job types that have non-empty queues
 369          *
 370          * @return string[] List of job types that have non-empty queues
 371          */
 372         public function getQueuesWithJobs() {
 373                 $types = [];
 374                 foreach ( $this->getCoalescedQueues() as $info ) {
 375                         /** @var JobQueue $queue */
 376                         $queue = $info['queue'];
 377                         $nonEmpty = $queue->getSiblingQueuesWithJobs( $this->getQueueTypes() );
 378                         if ( is_array( $nonEmpty ) ) { // batching features supported
 379                                 $types = array_merge( $types, $nonEmpty );
 380                         } else { // we have to go through the queues in the bucket one-by-one
 381                                 foreach ( $info['types'] as $type ) {
 382                                         if ( !$this->get( $type )->isEmpty() ) {
 383                                                 $types[] = $type;
 384                                         }
 385                                 }
 386                         }
 387                 }
 388
 389                 return $types;
 390         }
 391
 392         /**
 393          * Get the size of the queus for a list of job types
 394          *
 395          * @return int[] Map of (job type => size)
 396          */
 397         public function getQueueSizes() {
 398                 $sizeMap = [];
 399                 foreach ( $this->getCoalescedQueues() as $info ) {
 400                         /** @var JobQueue $queue */
 401                         $queue = $info['queue'];
 402                         $sizes = $queue->getSiblingQueueSizes( $this->getQueueTypes() );
 403                         if ( is_array( $sizes ) ) { // batching features supported
 404                                 $sizeMap = $sizeMap + $sizes;
 405                         } else { // we have to go through the queues in the bucket one-by-one
 406                                 foreach ( $info['types'] as $type ) {
 407                                         $sizeMap[$type] = $this->get( $type )->getSize();
 408                                 }
 409                         }
 410                 }
 411
 412                 return $sizeMap;
 413         }
 414
 415         /**
 416          * @return JobQueue[]
 417          */
 418         protected function getCoalescedQueues() {
 419                 global $wgJobTypeConf;
 420
 421                 if ( $this->coalescedQueues === null ) {
 422                         $this->coalescedQueues = [];
 423                         foreach ( $wgJobTypeConf as $type => $conf ) {
 424                                 $queue = JobQueue::factory(
 425                                         [ 'domain' => $this->domain, 'type' => 'null' ] + $conf );
 426                                 $loc = $queue->getCoalesceLocationInternal();
 427                                 if ( !isset( $this->coalescedQueues[$loc] ) ) {
 428                                         $this->coalescedQueues[$loc]['queue'] = $queue;
 429                                         $this->coalescedQueues[$loc]['types'] = [];
 430                                 }
 431                                 if ( $type === 'default' ) {
 432                                         $this->coalescedQueues[$loc]['types'] = array_merge(
 433                                                 $this->coalescedQueues[$loc]['types'],
 434                                                 array_diff( $this->getQueueTypes(), array_keys( $wgJobTypeConf ) )
 435                                         );
 436                                 } else {
 437                                         $this->coalescedQueues[$loc]['types'][] = $type;
 438                                 }
 439                         }
 440                 }
 441
 442                 return $this->coalescedQueues;
 443         }
 444
 445         /**
 446          * @param string $name
 447          * @return mixed
 448          */
 449         private function getCachedConfigVar( $name ) {
 450                 // @TODO: cleanup this whole method with a proper config system
 451                 if ( WikiMap::isCurrentWikiDbDomain( $this->domain ) ) {
 452                         return $GLOBALS[$name]; // common case
 453                 } else {
 454                         $wiki = WikiMap::getWikiIdFromDbDomain( $this->domain );
 455                         $cache = MediaWikiServices::getInstance()->getMainWANObjectCache();
 456                         $value = $cache->getWithSetCallback(
 457                                 $cache->makeGlobalKey( 'jobqueue', 'configvalue', $this->domain, $name ),
 458                                 $cache::TTL_DAY + mt_rand( 0, $cache::TTL_DAY ),
 459                                 function () use ( $wiki, $name ) {
 460                                         global $wgConf;
 461                                         // @TODO: use the full domain ID here
 462                                         return [ 'v' => $wgConf->getConfig( $wiki, $name ) ];
 463                                 },
 464                                 [ 'pcTTL' => WANObjectCache::TTL_PROC_LONG ]
 465                         );
 466
 467                         return $value['v'];
 468                 }
 469         }
 470
 471         /**
 472          * @param array $jobs
 473          * @throws InvalidArgumentException
 474          */
 475         private function assertValidJobs( array $jobs ) {
 476                 foreach ( $jobs as $job ) { // sanity checks
 477                         if ( !( $job instanceof IJobSpecification ) ) {
 478                                 throw new InvalidArgumentException( "Expected IJobSpecification objects" );
 479                         }
 480                 }
 481         }
 482 }