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                 return JobQueue::factory( $conf );
 122         }
 123
 124         /**
 125          * Insert jobs into the respective queues of which they belong
 126          *
 127          * This inserts the jobs into the queue specified by $wgJobTypeConf
 128          * and updates the aggregate job queue information cache as needed.
 129          *
 130          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 131          * @throws InvalidArgumentException
 132          * @return void
 133          */
 134         public function push( $jobs ) {
 135                 global $wgJobTypesExcludedFromDefaultQueue;
 136
 137                 if ( $this->invalidDomain ) {
 138                         // Do not enqueue job that cannot be run (T171371)
 139                         $e = new LogicException( "Domain '{$this->domain}' is not recognized." );
 140                         MWExceptionHandler::logException( $e );
 141                         return;
 142                 }
 143
 144                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 145                 if ( $jobs === [] ) {
 146                         return;
 147                 }
 148
 149                 $this->assertValidJobs( $jobs );
 150
 151                 $jobsByType = []; // (job type => list of jobs)
 152                 foreach ( $jobs as $job ) {
 153                         $jobsByType[$job->getType()][] = $job;
 154                 }
 155
 156                 foreach ( $jobsByType as $type => $jobs ) {
 157                         $this->get( $type )->push( $jobs );
 158                 }
 159
 160                 if ( $this->cache->hasField( 'queues-ready', 'list' ) ) {
 161                         $list = $this->cache->getField( 'queues-ready', 'list' );
 162                         if ( count( array_diff( array_keys( $jobsByType ), $list ) ) ) {
 163                                 $this->cache->clear( 'queues-ready' );
 164                         }
 165                 }
 166
 167                 $cache = ObjectCache::getLocalClusterInstance();
 168                 $cache->set(
 169                         $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', self::TYPE_ANY ),
 170                         'true',
 171                         15
 172                 );
 173                 if ( array_diff( array_keys( $jobsByType ), $wgJobTypesExcludedFromDefaultQueue ) ) {
 174                         $cache->set(
 175                                 $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', self::TYPE_DEFAULT ),
 176                                 'true',
 177                                 15
 178                         );
 179                 }
 180         }
 181
 182         /**
 183          * Buffer jobs for insertion via push() or call it now if in CLI mode
 184          *
 185          * Note that pushLazyJobs() is registered as a deferred update just before
 186          * DeferredUpdates::doUpdates() in MediaWiki and JobRunner classes in order
 187          * to be executed as the very last deferred update (T100085, T154425).
 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          * Push all jobs buffered via lazyPush() into their respective queues
 214          *
 215          * @return void
 216          * @since 1.26
 217          * @deprecated Since 1.33 Not needed anymore
 218          */
 219         public static function pushLazyJobs() {
 220                 wfDeprecated( __METHOD__, '1.33' );
 221         }
 222
 223         /**
 224          * Pop a job off one of the job queues
 225          *
 226          * This pops a job off a queue as specified by $wgJobTypeConf and
 227          * updates the aggregate job queue information cache as needed.
 228          *
 229          * @param int|string $qtype JobQueueGroup::TYPE_* constant or job type string
 230          * @param int $flags Bitfield of JobQueueGroup::USE_* constants
 231          * @param array $blacklist List of job types to ignore
 232          * @return Job|bool Returns false on failure
 233          */
 234         public function pop( $qtype = self::TYPE_DEFAULT, $flags = 0, array $blacklist = [] ) {
 235                 $job = false;
 236
 237                 if ( is_string( $qtype ) ) { // specific job type
 238                         if ( !in_array( $qtype, $blacklist ) ) {
 239                                 $job = $this->get( $qtype )->pop();
 240                         }
 241                 } else { // any job in the "default" jobs types
 242                         if ( $flags & self::USE_CACHE ) {
 243                                 if ( !$this->cache->hasField( 'queues-ready', 'list', self::PROC_CACHE_TTL ) ) {
 244                                         $this->cache->setField( 'queues-ready', 'list', $this->getQueuesWithJobs() );
 245                                 }
 246                                 $types = $this->cache->getField( 'queues-ready', 'list' );
 247                         } else {
 248                                 $types = $this->getQueuesWithJobs();
 249                         }
 250
 251                         if ( $qtype == self::TYPE_DEFAULT ) {
 252                                 $types = array_intersect( $types, $this->getDefaultQueueTypes() );
 253                         }
 254
 255                         $types = array_diff( $types, $blacklist ); // avoid selected types
 256                         shuffle( $types ); // avoid starvation
 257
 258                         foreach ( $types as $type ) { // for each queue...
 259                                 $job = $this->get( $type )->pop();
 260                                 if ( $job ) { // found
 261                                         break;
 262                                 } else { // not found
 263                                         $this->cache->clear( 'queues-ready' );
 264                                 }
 265                         }
 266                 }
 267
 268                 return $job;
 269         }
 270
 271         /**
 272          * Acknowledge that a job was completed
 273          *
 274          * @param Job $job
 275          * @return void
 276          */
 277         public function ack( Job $job ) {
 278                 $this->get( $job->getType() )->ack( $job );
 279         }
 280
 281         /**
 282          * Register the "root job" of a given job into the queue for de-duplication.
 283          * This should only be called right *after* all the new jobs have been inserted.
 284          *
 285          * @param Job $job
 286          * @return bool
 287          */
 288         public function deduplicateRootJob( Job $job ) {
 289                 return $this->get( $job->getType() )->deduplicateRootJob( $job );
 290         }
 291
 292         /**
 293          * Wait for any replica DBs or backup queue servers to catch up.
 294          *
 295          * This does nothing for certain queue classes.
 296          *
 297          * @return void
 298          */
 299         public function waitForBackups() {
 300                 global $wgJobTypeConf;
 301
 302                 // Try to avoid doing this more than once per queue storage medium
 303                 foreach ( $wgJobTypeConf as $type => $conf ) {
 304                         $this->get( $type )->waitForBackups();
 305                 }
 306         }
 307
 308         /**
 309          * Get the list of queue types
 310          *
 311          * @return array List of strings
 312          */
 313         public function getQueueTypes() {
 314                 return array_keys( $this->getCachedConfigVar( 'wgJobClasses' ) );
 315         }
 316
 317         /**
 318          * Get the list of default queue types
 319          *
 320          * @return array List of strings
 321          */
 322         public function getDefaultQueueTypes() {
 323                 global $wgJobTypesExcludedFromDefaultQueue;
 324
 325                 return array_diff( $this->getQueueTypes(), $wgJobTypesExcludedFromDefaultQueue );
 326         }
 327
 328         /**
 329          * Check if there are any queues with jobs (this is cached)
 330          *
 331          * @param int $type JobQueueGroup::TYPE_* constant
 332          * @return bool
 333          * @since 1.23
 334          */
 335         public function queuesHaveJobs( $type = self::TYPE_ANY ) {
 336                 $cache = ObjectCache::getLocalClusterInstance();
 337                 $key = $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', $type );
 338
 339                 $value = $cache->get( $key );
 340                 if ( $value === false ) {
 341                         $queues = $this->getQueuesWithJobs();
 342                         if ( $type == self::TYPE_DEFAULT ) {
 343                                 $queues = array_intersect( $queues, $this->getDefaultQueueTypes() );
 344                         }
 345                         $value = count( $queues ) ? 'true' : 'false';
 346                         $cache->add( $key, $value, 15 );
 347                 }
 348
 349                 return ( $value === 'true' );
 350         }
 351
 352         /**
 353          * Get the list of job types that have non-empty queues
 354          *
 355          * @return string[] List of job types that have non-empty queues
 356          */
 357         public function getQueuesWithJobs() {
 358                 $types = [];
 359                 foreach ( $this->getCoalescedQueues() as $info ) {
 360                         /** @var JobQueue $queue */
 361                         $queue = $info['queue'];
 362                         $nonEmpty = $queue->getSiblingQueuesWithJobs( $this->getQueueTypes() );
 363                         if ( is_array( $nonEmpty ) ) { // batching features supported
 364                                 $types = array_merge( $types, $nonEmpty );
 365                         } else { // we have to go through the queues in the bucket one-by-one
 366                                 foreach ( $info['types'] as $type ) {
 367                                         if ( !$this->get( $type )->isEmpty() ) {
 368                                                 $types[] = $type;
 369                                         }
 370                                 }
 371                         }
 372                 }
 373
 374                 return $types;
 375         }
 376
 377         /**
 378          * Get the size of the queus for a list of job types
 379          *
 380          * @return int[] Map of (job type => size)
 381          */
 382         public function getQueueSizes() {
 383                 $sizeMap = [];
 384                 foreach ( $this->getCoalescedQueues() as $info ) {
 385                         /** @var JobQueue $queue */
 386                         $queue = $info['queue'];
 387                         $sizes = $queue->getSiblingQueueSizes( $this->getQueueTypes() );
 388                         if ( is_array( $sizes ) ) { // batching features supported
 389                                 $sizeMap = $sizeMap + $sizes;
 390                         } else { // we have to go through the queues in the bucket one-by-one
 391                                 foreach ( $info['types'] as $type ) {
 392                                         $sizeMap[$type] = $this->get( $type )->getSize();
 393                                 }
 394                         }
 395                 }
 396
 397                 return $sizeMap;
 398         }
 399
 400         /**
 401          * @return JobQueue[]
 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 }