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                 $conf['aggregator'] = JobQueueAggregator::singleton();
 118                 if ( !isset( $conf['readOnlyReason'] ) ) {
 119                         $conf['readOnlyReason'] = $this->readOnlyReason;
 120                 }
 121
 122                 return JobQueue::factory( $conf );
 123         }
 124
 125         /**
 126          * Insert jobs into the respective queues of which they belong
 127          *
 128          * This inserts the jobs into the queue specified by $wgJobTypeConf
 129          * and updates the aggregate job queue information cache as needed.
 130          *
 131          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 132          * @throws InvalidArgumentException
 133          * @return void
 134          */
 135         public function push( $jobs ) {
 136                 global $wgJobTypesExcludedFromDefaultQueue;
 137
 138                 if ( $this->invalidDomain ) {
 139                         // Do not enqueue job that cannot be run (T171371)
 140                         $e = new LogicException( "Domain '{$this->domain}' is not recognized." );
 141                         MWExceptionHandler::logException( $e );
 142                         return;
 143                 }
 144
 145                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 146                 if ( $jobs === [] ) {
 147                         return;
 148                 }
 149
 150                 $this->assertValidJobs( $jobs );
 151
 152                 $jobsByType = []; // (job type => list of jobs)
 153                 foreach ( $jobs as $job ) {
 154                         $jobsByType[$job->getType()][] = $job;
 155                 }
 156
 157                 foreach ( $jobsByType as $type => $jobs ) {
 158                         $this->get( $type )->push( $jobs );
 159                 }
 160
 161                 if ( $this->cache->hasField( 'queues-ready', 'list' ) ) {
 162                         $list = $this->cache->getField( 'queues-ready', 'list' );
 163                         if ( count( array_diff( array_keys( $jobsByType ), $list ) ) ) {
 164                                 $this->cache->clear( 'queues-ready' );
 165                         }
 166                 }
 167
 168                 $cache = ObjectCache::getLocalClusterInstance();
 169                 $cache->set(
 170                         $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', self::TYPE_ANY ),
 171                         'true',
 172                         15
 173                 );
 174                 if ( array_diff( array_keys( $jobsByType ), $wgJobTypesExcludedFromDefaultQueue ) ) {
 175                         $cache->set(
 176                                 $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', self::TYPE_DEFAULT ),
 177                                 'true',
 178                                 15
 179                         );
 180                 }
 181         }
 182
 183         /**
 184          * Buffer jobs for insertion via push() or call it now if in CLI mode
 185          *
 186          * Note that pushLazyJobs() is registered as a deferred update just before
 187          * DeferredUpdates::doUpdates() in MediaWiki and JobRunner classes in order
 188          * to be executed as the very last deferred update (T100085, T154425).
 189          *
 190          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 191          * @return void
 192          * @since 1.26
 193          */
 194         public function lazyPush( $jobs ) {
 195                 if ( $this->invalidDomain ) {
 196                         // Do not enqueue job that cannot be run (T171371)
 197                         throw new LogicException( "Domain '{$this->domain}' is not recognized." );
 198                 }
 199
 200                 if ( PHP_SAPI === 'cli' || PHP_SAPI === 'phpdbg' ) {
 201                         $this->push( $jobs );
 202                         return;
 203                 }
 204
 205                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 206
 207                 // Throw errors now instead of on push(), when other jobs may be buffered
 208                 $this->assertValidJobs( $jobs );
 209
 210                 DeferredUpdates::addUpdate( new JobQueueEnqueueUpdate( $this->domain, $jobs ) );
 211         }
 212
 213         /**
 214          * Push all jobs buffered via lazyPush() into their respective queues
 215          *
 216          * @return void
 217          * @since 1.26
 218          * @deprecated Since 1.33 Not needed anymore
 219          */
 220         public static function pushLazyJobs() {
 221                 wfDeprecated( __METHOD__, '1.33' );
 222         }
 223
 224         /**
 225          * Pop a job off one of the job queues
 226          *
 227          * This pops a job off a queue as specified by $wgJobTypeConf and
 228          * updates the aggregate job queue information cache as needed.
 229          *
 230          * @param int|string $qtype JobQueueGroup::TYPE_* constant or job type string
 231          * @param int $flags Bitfield of JobQueueGroup::USE_* constants
 232          * @param array $blacklist List of job types to ignore
 233          * @return Job|bool Returns false on failure
 234          */
 235         public function pop( $qtype = self::TYPE_DEFAULT, $flags = 0, array $blacklist = [] ) {
 236                 $job = false;
 237
 238                 if ( is_string( $qtype ) ) { // specific job type
 239                         if ( !in_array( $qtype, $blacklist ) ) {
 240                                 $job = $this->get( $qtype )->pop();
 241                         }
 242                 } else { // any job in the "default" jobs types
 243                         if ( $flags & self::USE_CACHE ) {
 244                                 if ( !$this->cache->hasField( 'queues-ready', 'list', self::PROC_CACHE_TTL ) ) {
 245                                         $this->cache->setField( 'queues-ready', 'list', $this->getQueuesWithJobs() );
 246                                 }
 247                                 $types = $this->cache->getField( 'queues-ready', 'list' );
 248                         } else {
 249                                 $types = $this->getQueuesWithJobs();
 250                         }
 251
 252                         if ( $qtype == self::TYPE_DEFAULT ) {
 253                                 $types = array_intersect( $types, $this->getDefaultQueueTypes() );
 254                         }
 255
 256                         $types = array_diff( $types, $blacklist ); // avoid selected types
 257                         shuffle( $types ); // avoid starvation
 258
 259                         foreach ( $types as $type ) { // for each queue...
 260                                 $job = $this->get( $type )->pop();
 261                                 if ( $job ) { // found
 262                                         break;
 263                                 } else { // not found
 264                                         $this->cache->clear( 'queues-ready' );
 265                                 }
 266                         }
 267                 }
 268
 269                 return $job;
 270         }
 271
 272         /**
 273          * Acknowledge that a job was completed
 274          *
 275          * @param Job $job
 276          * @return void
 277          */
 278         public function ack( Job $job ) {
 279                 $this->get( $job->getType() )->ack( $job );
 280         }
 281
 282         /**
 283          * Register the "root job" of a given job into the queue for de-duplication.
 284          * This should only be called right *after* all the new jobs have been inserted.
 285          *
 286          * @param Job $job
 287          * @return bool
 288          */
 289         public function deduplicateRootJob( Job $job ) {
 290                 return $this->get( $job->getType() )->deduplicateRootJob( $job );
 291         }
 292
 293         /**
 294          * Wait for any replica DBs or backup queue servers to catch up.
 295          *
 296          * This does nothing for certain queue classes.
 297          *
 298          * @return void
 299          */
 300         public function waitForBackups() {
 301                 global $wgJobTypeConf;
 302
 303                 // Try to avoid doing this more than once per queue storage medium
 304                 foreach ( $wgJobTypeConf as $type => $conf ) {
 305                         $this->get( $type )->waitForBackups();
 306                 }
 307         }
 308
 309         /**
 310          * Get the list of queue types
 311          *
 312          * @return array List of strings
 313          */
 314         public function getQueueTypes() {
 315                 return array_keys( $this->getCachedConfigVar( 'wgJobClasses' ) );
 316         }
 317
 318         /**
 319          * Get the list of default queue types
 320          *
 321          * @return array List of strings
 322          */
 323         public function getDefaultQueueTypes() {
 324                 global $wgJobTypesExcludedFromDefaultQueue;
 325
 326                 return array_diff( $this->getQueueTypes(), $wgJobTypesExcludedFromDefaultQueue );
 327         }
 328
 329         /**
 330          * Check if there are any queues with jobs (this is cached)
 331          *
 332          * @param int $type JobQueueGroup::TYPE_* constant
 333          * @return bool
 334          * @since 1.23
 335          */
 336         public function queuesHaveJobs( $type = self::TYPE_ANY ) {
 337                 $cache = ObjectCache::getLocalClusterInstance();
 338                 $key = $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', $type );
 339
 340                 $value = $cache->get( $key );
 341                 if ( $value === false ) {
 342                         $queues = $this->getQueuesWithJobs();
 343                         if ( $type == self::TYPE_DEFAULT ) {
 344                                 $queues = array_intersect( $queues, $this->getDefaultQueueTypes() );
 345                         }
 346                         $value = count( $queues ) ? 'true' : 'false';
 347                         $cache->add( $key, $value, 15 );
 348                 }
 349
 350                 return ( $value === 'true' );
 351         }
 352
 353         /**
 354          * Get the list of job types that have non-empty queues
 355          *
 356          * @return array List of job types that have non-empty queues
 357          */
 358         public function getQueuesWithJobs() {
 359                 $types = [];
 360                 foreach ( $this->getCoalescedQueues() as $info ) {
 361                         $nonEmpty = $info['queue']->getSiblingQueuesWithJobs( $this->getQueueTypes() );
 362                         if ( is_array( $nonEmpty ) ) { // batching features supported
 363                                 $types = array_merge( $types, $nonEmpty );
 364                         } else { // we have to go through the queues in the bucket one-by-one
 365                                 foreach ( $info['types'] as $type ) {
 366                                         if ( !$this->get( $type )->isEmpty() ) {
 367                                                 $types[] = $type;
 368                                         }
 369                                 }
 370                         }
 371                 }
 372
 373                 return $types;
 374         }
 375
 376         /**
 377          * Get the size of the queus for a list of job types
 378          *
 379          * @return array Map of (job type => size)
 380          */
 381         public function getQueueSizes() {
 382                 $sizeMap = [];
 383                 foreach ( $this->getCoalescedQueues() as $info ) {
 384                         $sizes = $info['queue']->getSiblingQueueSizes( $this->getQueueTypes() );
 385                         if ( is_array( $sizes ) ) { // batching features supported
 386                                 $sizeMap = $sizeMap + $sizes;
 387                         } else { // we have to go through the queues in the bucket one-by-one
 388                                 foreach ( $info['types'] as $type ) {
 389                                         $sizeMap[$type] = $this->get( $type )->getSize();
 390                                 }
 391                         }
 392                 }
 393
 394                 return $sizeMap;
 395         }
 396
 397         /**
 398          * @return array
 399          */
 400         protected function getCoalescedQueues() {
 401                 global $wgJobTypeConf;
 402
 403                 if ( $this->coalescedQueues === null ) {
 404                         $this->coalescedQueues = [];
 405                         foreach ( $wgJobTypeConf as $type => $conf ) {
 406                                 $queue = JobQueue::factory(
 407                                         [ 'wiki' => $this->domain, 'type' => 'null' ] + $conf );
 408                                 $loc = $queue->getCoalesceLocationInternal();
 409                                 if ( !isset( $this->coalescedQueues[$loc] ) ) {
 410                                         $this->coalescedQueues[$loc]['queue'] = $queue;
 411                                         $this->coalescedQueues[$loc]['types'] = [];
 412                                 }
 413                                 if ( $type === 'default' ) {
 414                                         $this->coalescedQueues[$loc]['types'] = array_merge(
 415                                                 $this->coalescedQueues[$loc]['types'],
 416                                                 array_diff( $this->getQueueTypes(), array_keys( $wgJobTypeConf ) )
 417                                         );
 418                                 } else {
 419                                         $this->coalescedQueues[$loc]['types'][] = $type;
 420                                 }
 421                         }
 422                 }
 423
 424                 return $this->coalescedQueues;
 425         }
 426
 427         /**
 428          * @param string $name
 429          * @return mixed
 430          */
 431         private function getCachedConfigVar( $name ) {
 432                 // @TODO: cleanup this whole method with a proper config system
 433                 if ( WikiMap::isCurrentWikiDbDomain( $this->domain ) ) {
 434                         return $GLOBALS[$name]; // common case
 435                 } else {
 436                         $wiki = WikiMap::getWikiIdFromDbDomain( $this->domain );
 437                         $cache = MediaWikiServices::getInstance()->getMainWANObjectCache();
 438                         $value = $cache->getWithSetCallback(
 439                                 $cache->makeGlobalKey( 'jobqueue', 'configvalue', $this->domain, $name ),
 440                                 $cache::TTL_DAY + mt_rand( 0, $cache::TTL_DAY ),
 441                                 function () use ( $wiki, $name ) {
 442                                         global $wgConf;
 443                                         // @TODO: use the full domain ID here
 444                                         return [ 'v' => $wgConf->getConfig( $wiki, $name ) ];
 445                                 },
 446                                 [ 'pcTTL' => WANObjectCache::TTL_PROC_LONG ]
 447                         );
 448
 449                         return $value['v'];
 450                 }
 451         }
 452
 453         /**
 454          * @param array $jobs
 455          * @throws InvalidArgumentException
 456          */
 457         private function assertValidJobs( array $jobs ) {
 458                 foreach ( $jobs as $job ) { // sanity checks
 459                         if ( !( $job instanceof IJobSpecification ) ) {
 460                                 throw new InvalidArgumentException( "Expected IJobSpecification objects" );
 461                         }
 462                 }
 463         }
 464 }