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         /** @var Job[] */
  47         protected $bufferedJobs = [];
  48
  49         const TYPE_DEFAULT = 1; // integer; jobs popped by default
  50         const TYPE_ANY = 2; // integer; any job
  51
  52         const USE_CACHE = 1; // integer; use process or persistent cache
  53
  54         const PROC_CACHE_TTL = 15; // integer; seconds
  55
  56         const CACHE_VERSION = 1; // integer; cache version
  57
  58         /**
  59          * @param string $wiki Wiki ID
  60          * @param string|bool $readOnlyReason Read-only reason or false
  61          */
  62         protected function __construct( $wiki, $readOnlyReason ) {
  63                 $this->wiki = $wiki;
  64                 $this->readOnlyReason = $readOnlyReason;
  65                 $this->cache = new MapCacheLRU( 10 );
  66         }
  67
  68         /**
  69          * @param bool|string $wiki Wiki ID
  70          * @return JobQueueGroup
  71          */
  72         public static function singleton( $wiki = false ) {
  73                 global $wgLocalDatabases;
  74
  75                 $wiki = ( $wiki === false ) ? wfWikiID() : $wiki;
  76
  77                 if ( !isset( self::$instances[$wiki] ) ) {
  78                         self::$instances[$wiki] = new self( $wiki, 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                         if ( $wiki !== wfWikiID() && !in_array( $wiki, $wgLocalDatabases ) ) {
  82                                 self::$instances[$wiki]->invalidWiki = true;
  83                         }
  84                 }
  85
  86                 return self::$instances[$wiki];
  87         }
  88
  89         /**
  90          * Destroy the singleton instances
  91          *
  92          * @return void
  93          */
  94         public static function destroySingletons() {
  95                 self::$instances = [];
  96         }
  97
  98         /**
  99          * Get the job queue object for a given queue type
 100          *
 101          * @param string $type
 102          * @return JobQueue
 103          */
 104         public function get( $type ) {
 105                 global $wgJobTypeConf;
 106
 107                 $conf = [ 'wiki' => $this->wiki, 'type' => $type ];
 108                 $conf += $wgJobTypeConf[$type] ?? $wgJobTypeConf['default'];
 109                 $conf['aggregator'] = JobQueueAggregator::singleton();
 110                 if ( !isset( $conf['readOnlyReason'] ) ) {
 111                         $conf['readOnlyReason'] = $this->readOnlyReason;
 112                 }
 113
 114                 return JobQueue::factory( $conf );
 115         }
 116
 117         /**
 118          * Insert jobs into the respective queues of which they belong
 119          *
 120          * This inserts the jobs into the queue specified by $wgJobTypeConf
 121          * and updates the aggregate job queue information cache as needed.
 122          *
 123          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 124          * @throws InvalidArgumentException
 125          * @return void
 126          */
 127         public function push( $jobs ) {
 128                 global $wgJobTypesExcludedFromDefaultQueue;
 129
 130                 if ( $this->invalidWiki ) {
 131                         // Do not enqueue job that cannot be run (T171371)
 132                         $e = new LogicException( "Domain '{$this->wiki}' is not recognized." );
 133                         MWExceptionHandler::logException( $e );
 134                         return;
 135                 }
 136
 137                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 138                 if ( !count( $jobs ) ) {
 139                         return;
 140                 }
 141
 142                 $this->assertValidJobs( $jobs );
 143
 144                 $jobsByType = []; // (job type => list of jobs)
 145                 foreach ( $jobs as $job ) {
 146                         $jobsByType[$job->getType()][] = $job;
 147                 }
 148
 149                 foreach ( $jobsByType as $type => $jobs ) {
 150                         $this->get( $type )->push( $jobs );
 151                 }
 152
 153                 if ( $this->cache->hasField( 'queues-ready', 'list' ) ) {
 154                         $list = $this->cache->getField( 'queues-ready', 'list' );
 155                         if ( count( array_diff( array_keys( $jobsByType ), $list ) ) ) {
 156                                 $this->cache->clear( 'queues-ready' );
 157                         }
 158                 }
 159
 160                 $cache = ObjectCache::getLocalClusterInstance();
 161                 $cache->set(
 162                         $cache->makeGlobalKey( 'jobqueue', $this->wiki, 'hasjobs', self::TYPE_ANY ),
 163                         'true',
 164                         15
 165                 );
 166                 if ( array_diff( array_keys( $jobsByType ), $wgJobTypesExcludedFromDefaultQueue ) ) {
 167                         $cache->set(
 168                                 $cache->makeGlobalKey( 'jobqueue', $this->wiki, 'hasjobs', self::TYPE_DEFAULT ),
 169                                 'true',
 170                                 15
 171                         );
 172                 }
 173         }
 174
 175         /**
 176          * Buffer jobs for insertion via push() or call it now if in CLI mode
 177          *
 178          * Note that pushLazyJobs() is registered as a deferred update just before
 179          * DeferredUpdates::doUpdates() in MediaWiki and JobRunner classes in order
 180          * to be executed as the very last deferred update (T100085, T154425).
 181          *
 182          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 183          * @return void
 184          * @since 1.26
 185          */
 186         public function lazyPush( $jobs ) {
 187                 if ( $this->invalidWiki ) {
 188                         // Do not enqueue job that cannot be run (T171371)
 189                         throw new LogicException( "Domain '{$this->wiki}' is not recognized." );
 190                 }
 191
 192                 if ( PHP_SAPI === 'cli' || PHP_SAPI === 'phpdbg' ) {
 193                         $this->push( $jobs );
 194                         return;
 195                 }
 196
 197                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 198
 199                 // Throw errors now instead of on push(), when other jobs may be buffered
 200                 $this->assertValidJobs( $jobs );
 201
 202                 $this->bufferedJobs = array_merge( $this->bufferedJobs, $jobs );
 203         }
 204
 205         /**
 206          * Push all jobs buffered via lazyPush() into their respective queues
 207          *
 208          * @return void
 209          * @since 1.26
 210          */
 211         public static function pushLazyJobs() {
 212                 foreach ( self::$instances as $group ) {
 213                         try {
 214                                 $group->push( $group->bufferedJobs );
 215                                 $group->bufferedJobs = [];
 216                         } catch ( Exception $e ) {
 217                                 // Get in as many jobs as possible and let other post-send updates happen
 218                                 MWExceptionHandler::logException( $e );
 219                         }
 220                 }
 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->wiki, '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 array List of job types that have non-empty queues
 356          */
 357         public function getQueuesWithJobs() {
 358                 $types = [];
 359                 foreach ( $this->getCoalescedQueues() as $info ) {
 360                         $nonEmpty = $info['queue']->getSiblingQueuesWithJobs( $this->getQueueTypes() );
 361                         if ( is_array( $nonEmpty ) ) { // batching features supported
 362                                 $types = array_merge( $types, $nonEmpty );
 363                         } else { // we have to go through the queues in the bucket one-by-one
 364                                 foreach ( $info['types'] as $type ) {
 365                                         if ( !$this->get( $type )->isEmpty() ) {
 366                                                 $types[] = $type;
 367                                         }
 368                                 }
 369                         }
 370                 }
 371
 372                 return $types;
 373         }
 374
 375         /**
 376          * Get the size of the queus for a list of job types
 377          *
 378          * @return array Map of (job type => size)
 379          */
 380         public function getQueueSizes() {
 381                 $sizeMap = [];
 382                 foreach ( $this->getCoalescedQueues() as $info ) {
 383                         $sizes = $info['queue']->getSiblingQueueSizes( $this->getQueueTypes() );
 384                         if ( is_array( $sizes ) ) { // batching features supported
 385                                 $sizeMap = $sizeMap + $sizes;
 386                         } else { // we have to go through the queues in the bucket one-by-one
 387                                 foreach ( $info['types'] as $type ) {
 388                                         $sizeMap[$type] = $this->get( $type )->getSize();
 389                                 }
 390                         }
 391                 }
 392
 393                 return $sizeMap;
 394         }
 395
 396         /**
 397          * @return array
 398          */
 399         protected function getCoalescedQueues() {
 400                 global $wgJobTypeConf;
 401
 402                 if ( $this->coalescedQueues === null ) {
 403                         $this->coalescedQueues = [];
 404                         foreach ( $wgJobTypeConf as $type => $conf ) {
 405                                 $queue = JobQueue::factory(
 406                                         [ 'wiki' => $this->wiki, 'type' => 'null' ] + $conf );
 407                                 $loc = $queue->getCoalesceLocationInternal();
 408                                 if ( !isset( $this->coalescedQueues[$loc] ) ) {
 409                                         $this->coalescedQueues[$loc]['queue'] = $queue;
 410                                         $this->coalescedQueues[$loc]['types'] = [];
 411                                 }
 412                                 if ( $type === 'default' ) {
 413                                         $this->coalescedQueues[$loc]['types'] = array_merge(
 414                                                 $this->coalescedQueues[$loc]['types'],
 415                                                 array_diff( $this->getQueueTypes(), array_keys( $wgJobTypeConf ) )
 416                                         );
 417                                 } else {
 418                                         $this->coalescedQueues[$loc]['types'][] = $type;
 419                                 }
 420                         }
 421                 }
 422
 423                 return $this->coalescedQueues;
 424         }
 425
 426         /**
 427          * @param string $name
 428          * @return mixed
 429          */
 430         private function getCachedConfigVar( $name ) {
 431                 // @TODO: cleanup this whole method with a proper config system
 432                 if ( $this->wiki === wfWikiID() ) {
 433                         return $GLOBALS[$name]; // common case
 434                 } else {
 435                         $wiki = $this->wiki;
 436                         $cache = ObjectCache::getMainWANInstance();
 437                         $value = $cache->getWithSetCallback(
 438                                 $cache->makeGlobalKey( 'jobqueue', 'configvalue', $wiki, $name ),
 439                                 $cache::TTL_DAY + mt_rand( 0, $cache::TTL_DAY ),
 440                                 function () use ( $wiki, $name ) {
 441                                         global $wgConf;
 442
 443                                         return [ 'v' => $wgConf->getConfig( $wiki, $name ) ];
 444                                 },
 445                                 [ 'pcTTL' => WANObjectCache::TTL_PROC_LONG ]
 446                         );
 447
 448                         return $value['v'];
 449                 }
 450         }
 451
 452         /**
 453          * @param array $jobs
 454          * @throws InvalidArgumentException
 455          */
 456         private function assertValidJobs( array $jobs ) {
 457                 foreach ( $jobs as $job ) { // sanity checks
 458                         if ( !( $job instanceof IJobSpecification ) ) {
 459                                 throw new InvalidArgumentException( "Expected IJobSpecification objects" );
 460                         }
 461                 }
 462         }
 463
 464         function __destruct() {
 465                 $n = count( $this->bufferedJobs );
 466                 if ( $n > 0 ) {
 467                         $type = implode( ', ', array_unique( array_map( 'get_class', $this->bufferedJobs ) ) );
 468                         trigger_error( __METHOD__ . ": $n buffered job(s) of type(s) $type never inserted." );
 469                 }
 470         }
 471 }