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