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 MapCacheLRU */
  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
 125                 return JobQueue::factory( $conf );
 126         }
 127
 128         /**
 129          * Insert jobs into the respective queues of which they belong
 130          *
 131          * This inserts the jobs into the queue specified by $wgJobTypeConf
 132          * and updates the aggregate job queue information cache as needed.
 133          *
 134          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 135          * @throws InvalidArgumentException
 136          * @return void
 137          */
 138         public function push( $jobs ) {
 139                 global $wgJobTypesExcludedFromDefaultQueue;
 140
 141                 if ( $this->invalidDomain ) {
 142                         // Do not enqueue job that cannot be run (T171371)
 143                         $e = new LogicException( "Domain '{$this->domain}' is not recognized." );
 144                         MWExceptionHandler::logException( $e );
 145                         return;
 146                 }
 147
 148                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 149                 if ( $jobs === [] ) {
 150                         return;
 151                 }
 152
 153                 $this->assertValidJobs( $jobs );
 154
 155                 $jobsByType = []; // (job type => list of jobs)
 156                 foreach ( $jobs as $job ) {
 157                         $jobsByType[$job->getType()][] = $job;
 158                 }
 159
 160                 foreach ( $jobsByType as $type => $jobs ) {
 161                         $this->get( $type )->push( $jobs );
 162                 }
 163
 164                 if ( $this->cache->hasField( 'queues-ready', 'list' ) ) {
 165                         $list = $this->cache->getField( 'queues-ready', 'list' );
 166                         if ( count( array_diff( array_keys( $jobsByType ), $list ) ) ) {
 167                                 $this->cache->clear( 'queues-ready' );
 168                         }
 169                 }
 170
 171                 $cache = ObjectCache::getLocalClusterInstance();
 172                 $cache->set(
 173                         $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', self::TYPE_ANY ),
 174                         'true',
 175                         15
 176                 );
 177                 if ( array_diff( array_keys( $jobsByType ), $wgJobTypesExcludedFromDefaultQueue ) ) {
 178                         $cache->set(
 179                                 $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', self::TYPE_DEFAULT ),
 180                                 'true',
 181                                 15
 182                         );
 183                 }
 184         }
 185
 186         /**
 187          * Buffer jobs for insertion via push() or call it now if in CLI mode
 188          *
 189          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 190          * @return void
 191          * @since 1.26
 192          */
 193         public function lazyPush( $jobs ) {
 194                 if ( $this->invalidDomain ) {
 195                         // Do not enqueue job that cannot be run (T171371)
 196                         throw new LogicException( "Domain '{$this->domain}' is not recognized." );
 197                 }
 198
 199                 if ( PHP_SAPI === 'cli' || PHP_SAPI === 'phpdbg' ) {
 200                         $this->push( $jobs );
 201                         return;
 202                 }
 203
 204                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 205
 206                 // Throw errors now instead of on push(), when other jobs may be buffered
 207                 $this->assertValidJobs( $jobs );
 208
 209                 DeferredUpdates::addUpdate( new JobQueueEnqueueUpdate( $this->domain, $jobs ) );
 210         }
 211
 212         /**
 213          * Pop a job off one of the job queues
 214          *
 215          * This pops a job off a queue as specified by $wgJobTypeConf and
 216          * updates the aggregate job queue information cache as needed.
 217          *
 218          * @param int|string $qtype JobQueueGroup::TYPE_* constant or job type string
 219          * @param int $flags Bitfield of JobQueueGroup::USE_* constants
 220          * @param array $blacklist List of job types to ignore
 221          * @return RunnableJob|bool Returns false on failure
 222          */
 223         public function pop( $qtype = self::TYPE_DEFAULT, $flags = 0, array $blacklist = [] ) {
 224                 global $wgJobClasses;
 225
 226                 $job = false;
 227
 228                 if ( !WikiMap::isCurrentWikiDbDomain( $this->domain ) ) {
 229                         throw new JobQueueError(
 230                                 "Cannot pop '{$qtype}' job off foreign '{$this->domain}' wiki queue." );
 231                 } elseif ( is_string( $qtype ) && !isset( $wgJobClasses[$qtype] ) ) {
 232                         // Do not pop jobs if there is no class for the queue type
 233                         throw new JobQueueError( "Unrecognized job type '$qtype'." );
 234                 }
 235
 236                 if ( is_string( $qtype ) ) { // specific job type
 237                         if ( !in_array( $qtype, $blacklist ) ) {
 238                                 $job = $this->get( $qtype )->pop();
 239                         }
 240                 } else { // any job in the "default" jobs types
 241                         if ( $flags & self::USE_CACHE ) {
 242                                 if ( !$this->cache->hasField( 'queues-ready', 'list', self::PROC_CACHE_TTL ) ) {
 243                                         $this->cache->setField( 'queues-ready', 'list', $this->getQueuesWithJobs() );
 244                                 }
 245                                 $types = $this->cache->getField( 'queues-ready', 'list' );
 246                         } else {
 247                                 $types = $this->getQueuesWithJobs();
 248                         }
 249
 250                         if ( $qtype == self::TYPE_DEFAULT ) {
 251                                 $types = array_intersect( $types, $this->getDefaultQueueTypes() );
 252                         }
 253
 254                         $types = array_diff( $types, $blacklist ); // avoid selected types
 255                         shuffle( $types ); // avoid starvation
 256
 257                         foreach ( $types as $type ) { // for each queue...
 258                                 $job = $this->get( $type )->pop();
 259                                 if ( $job ) { // found
 260                                         break;
 261                                 } else { // not found
 262                                         $this->cache->clear( 'queues-ready' );
 263                                 }
 264                         }
 265                 }
 266
 267                 return $job;
 268         }
 269
 270         /**
 271          * Acknowledge that a job was completed
 272          *
 273          * @param RunnableJob $job
 274          * @return void
 275          */
 276         public function ack( RunnableJob $job ) {
 277                 $this->get( $job->getType() )->ack( $job );
 278         }
 279
 280         /**
 281          * Register the "root job" of a given job into the queue for de-duplication.
 282          * This should only be called right *after* all the new jobs have been inserted.
 283          *
 284          * @param RunnableJob $job
 285          * @return bool
 286          */
 287         public function deduplicateRootJob( RunnableJob $job ) {
 288                 return $this->get( $job->getType() )->deduplicateRootJob( $job );
 289         }
 290
 291         /**
 292          * Wait for any replica DBs or backup queue servers to catch up.
 293          *
 294          * This does nothing for certain queue classes.
 295          *
 296          * @return void
 297          */
 298         public function waitForBackups() {
 299                 global $wgJobTypeConf;
 300
 301                 // Try to avoid doing this more than once per queue storage medium
 302                 foreach ( $wgJobTypeConf as $type => $conf ) {
 303                         $this->get( $type )->waitForBackups();
 304                 }
 305         }
 306
 307         /**
 308          * Get the list of queue types
 309          *
 310          * @return array List of strings
 311          */
 312         public function getQueueTypes() {
 313                 return array_keys( $this->getCachedConfigVar( 'wgJobClasses' ) );
 314         }
 315
 316         /**
 317          * Get the list of default queue types
 318          *
 319          * @return array List of strings
 320          */
 321         public function getDefaultQueueTypes() {
 322                 global $wgJobTypesExcludedFromDefaultQueue;
 323
 324                 return array_diff( $this->getQueueTypes(), $wgJobTypesExcludedFromDefaultQueue );
 325         }
 326
 327         /**
 328          * Check if there are any queues with jobs (this is cached)
 329          *
 330          * @param int $type JobQueueGroup::TYPE_* constant
 331          * @return bool
 332          * @since 1.23
 333          */
 334         public function queuesHaveJobs( $type = self::TYPE_ANY ) {
 335                 $cache = ObjectCache::getLocalClusterInstance();
 336                 $key = $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', $type );
 337
 338                 $value = $cache->get( $key );
 339                 if ( $value === false ) {
 340                         $queues = $this->getQueuesWithJobs();
 341                         if ( $type == self::TYPE_DEFAULT ) {
 342                                 $queues = array_intersect( $queues, $this->getDefaultQueueTypes() );
 343                         }
 344                         $value = count( $queues ) ? 'true' : 'false';
 345                         $cache->add( $key, $value, 15 );
 346                 }
 347
 348                 return ( $value === 'true' );
 349         }
 350
 351         /**
 352          * Get the list of job types that have non-empty queues
 353          *
 354          * @return string[] List of job types that have non-empty queues
 355          */
 356         public function getQueuesWithJobs() {
 357                 $types = [];
 358                 foreach ( $this->getCoalescedQueues() as $info ) {
 359                         /** @var JobQueue $queue */
 360                         $queue = $info['queue'];
 361                         $nonEmpty = $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 int[] Map of (job type => size)
 380          */
 381         public function getQueueSizes() {
 382                 $sizeMap = [];
 383                 foreach ( $this->getCoalescedQueues() as $info ) {
 384                         /** @var JobQueue $queue */
 385                         $queue = $info['queue'];
 386                         $sizes = $queue->getSiblingQueueSizes( $this->getQueueTypes() );
 387                         if ( is_array( $sizes ) ) { // batching features supported
 388                                 $sizeMap = $sizeMap + $sizes;
 389                         } else { // we have to go through the queues in the bucket one-by-one
 390                                 foreach ( $info['types'] as $type ) {
 391                                         $sizeMap[$type] = $this->get( $type )->getSize();
 392                                 }
 393                         }
 394                 }
 395
 396                 return $sizeMap;
 397         }
 398
 399         /**
 400          * @return array[]
 401          * @phan-return array<string,array{queue:JobQueue,types:array<string,class-string>}>
 402          */
 403         protected function getCoalescedQueues() {
 404                 global $wgJobTypeConf;
 405
 406                 if ( $this->coalescedQueues === null ) {
 407                         $this->coalescedQueues = [];
 408                         foreach ( $wgJobTypeConf as $type => $conf ) {
 409                                 $queue = JobQueue::factory(
 410                                         [ 'domain' => $this->domain, 'type' => 'null' ] + $conf );
 411                                 $loc = $queue->getCoalesceLocationInternal();
 412                                 if ( !isset( $this->coalescedQueues[$loc] ) ) {
 413                                         $this->coalescedQueues[$loc]['queue'] = $queue;
 414                                         $this->coalescedQueues[$loc]['types'] = [];
 415                                 }
 416                                 if ( $type === 'default' ) {
 417                                         $this->coalescedQueues[$loc]['types'] = array_merge(
 418                                                 $this->coalescedQueues[$loc]['types'],
 419                                                 array_diff( $this->getQueueTypes(), array_keys( $wgJobTypeConf ) )
 420                                         );
 421                                 } else {
 422                                         $this->coalescedQueues[$loc]['types'][] = $type;
 423                                 }
 424                         }
 425                 }
 426
 427                 return $this->coalescedQueues;
 428         }
 429
 430         /**
 431          * @param string $name
 432          * @return mixed
 433          */
 434         private function getCachedConfigVar( $name ) {
 435                 // @TODO: cleanup this whole method with a proper config system
 436                 if ( WikiMap::isCurrentWikiDbDomain( $this->domain ) ) {
 437                         return $GLOBALS[$name]; // common case
 438                 } else {
 439                         $wiki = WikiMap::getWikiIdFromDbDomain( $this->domain );
 440                         $cache = MediaWikiServices::getInstance()->getMainWANObjectCache();
 441                         $value = $cache->getWithSetCallback(
 442                                 $cache->makeGlobalKey( 'jobqueue', 'configvalue', $this->domain, $name ),
 443                                 $cache::TTL_DAY + mt_rand( 0, $cache::TTL_DAY ),
 444                                 function () use ( $wiki, $name ) {
 445                                         global $wgConf;
 446                                         // @TODO: use the full domain ID here
 447                                         return [ 'v' => $wgConf->getConfig( $wiki, $name ) ];
 448                                 },
 449                                 [ 'pcTTL' => WANObjectCache::TTL_PROC_LONG ]
 450                         );
 451
 452                         return $value['v'];
 453                 }
 454         }
 455
 456         /**
 457          * @param array $jobs
 458          * @throws InvalidArgumentException
 459          */
 460         private function assertValidJobs( array $jobs ) {
 461                 foreach ( $jobs as $job ) { // sanity checks
 462                         if ( !( $job instanceof IJobSpecification ) ) {
 463                                 throw new InvalidArgumentException( "Expected IJobSpecification objects" );
 464                         }
 465                 }
 466         }
 467 }