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 DB domain ID */
  37         protected $domain;
  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 $domain Wiki DB domain ID
  60          * @param string|bool $readOnlyReason Read-only reason or false
  61          */
  62         protected function __construct( $domain, $readOnlyReason ) {
  63                 $this->domain = $domain;
  64                 $this->readOnlyReason = $readOnlyReason;
  65                 $this->cache = new ProcessCacheLRU( 10 );
  66         }
  67
  68         /**
  69          * @param bool|string $domain Wiki domain ID
  70          * @return JobQueueGroup
  71          */
  72         public static function singleton( $domain = false ) {
  73                 global $wgLocalDatabases;
  74
  75                 if ( $domain === false ) {
  76                         $domain = WikiMap::getCurrentWikiDbDomain()->getId();
  77                 }
  78
  79                 if ( !isset( self::$instances[$domain] ) ) {
  80                         self::$instances[$domain] = new self( $domain, wfConfiguredReadOnlyReason() );
  81                         // Make sure jobs are not getting pushed to bogus wikis. This can confuse
  82                         // the job runner system into spawning endless RPC requests that fail (T171371).
  83                         $wikiId = WikiMap::getWikiIdFromDomain( $domain );
  84                         if (
  85                                 !WikiMap::isCurrentWikiDbDomain( $domain ) &&
  86                                 !in_array( $wikiId, $wgLocalDatabases )
  87                         ) {
  88                                 self::$instances[$domain]->invalidWiki = true;
  89                         }
  90                 }
  91
  92                 return self::$instances[$domain];
  93         }
  94
  95         /**
  96          * Destroy the singleton instances
  97          *
  98          * @return void
  99          */
 100         public static function destroySingletons() {
 101                 self::$instances = [];
 102         }
 103
 104         /**
 105          * Get the job queue object for a given queue type
 106          *
 107          * @param string $type
 108          * @return JobQueue
 109          */
 110         public function get( $type ) {
 111                 global $wgJobTypeConf;
 112
 113                 $conf = [ 'wiki' => $this->domain, 'type' => $type ];
 114                 if ( isset( $wgJobTypeConf[$type] ) ) {
 115                         $conf = $conf + $wgJobTypeConf[$type];
 116                 } else {
 117                         $conf = $conf + $wgJobTypeConf['default'];
 118                 }
 119                 $conf['aggregator'] = JobQueueAggregator::singleton();
 120                 if ( $this->readOnlyReason !== false ) {
 121                         $conf['readOnlyReason'] = $this->readOnlyReason;
 122                 }
 123
 124                 return JobQueue::factory( $conf );
 125         }
 126
 127         /**
 128          * Insert jobs into the respective queues of which they belong
 129          *
 130          * This inserts the jobs into the queue specified by $wgJobTypeConf
 131          * and updates the aggregate job queue information cache as needed.
 132          *
 133          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 134          * @throws InvalidArgumentException
 135          * @return void
 136          */
 137         public function push( $jobs ) {
 138                 global $wgJobTypesExcludedFromDefaultQueue;
 139
 140                 if ( $this->invalidWiki ) {
 141                         // Do not enqueue job that cannot be run (T171371)
 142                         $e = new LogicException( "Domain '{$this->domain}' is not recognized." );
 143                         MWExceptionHandler::logException( $e );
 144                         return;
 145                 }
 146
 147                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 148                 if ( !count( $jobs ) ) {
 149                         return;
 150                 }
 151
 152                 $this->assertValidJobs( $jobs );
 153
 154                 $jobsByType = []; // (job type => list of jobs)
 155                 foreach ( $jobs as $job ) {
 156                         $jobsByType[$job->getType()][] = $job;
 157                 }
 158
 159                 foreach ( $jobsByType as $type => $jobs ) {
 160                         $this->get( $type )->push( $jobs );
 161                 }
 162
 163                 if ( $this->cache->has( 'queues-ready', 'list' ) ) {
 164                         $list = $this->cache->get( 'queues-ready', 'list' );
 165                         if ( count( array_diff( array_keys( $jobsByType ), $list ) ) ) {
 166                                 $this->cache->clear( 'queues-ready' );
 167                         }
 168                 }
 169
 170                 $cache = ObjectCache::getLocalClusterInstance();
 171                 $cache->set(
 172                         $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', self::TYPE_ANY ),
 173                         'true',
 174                         15
 175                 );
 176                 if ( array_diff( array_keys( $jobsByType ), $wgJobTypesExcludedFromDefaultQueue ) ) {
 177                         $cache->set(
 178                                 $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', self::TYPE_DEFAULT ),
 179                                 'true',
 180                                 15
 181                         );
 182                 }
 183         }
 184
 185         /**
 186          * Buffer jobs for insertion via push() or call it now if in CLI mode
 187          *
 188          * Note that pushLazyJobs() is registered as a deferred update just before
 189          * DeferredUpdates::doUpdates() in MediaWiki and JobRunner classes in order
 190          * to be executed as the very last deferred update (T100085, T154425).
 191          *
 192          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 193          * @return void
 194          * @since 1.26
 195          */
 196         public function lazyPush( $jobs ) {
 197                 if ( $this->invalidWiki ) {
 198                         // Do not enqueue job that cannot be run (T171371)
 199                         throw new LogicException( "Domain '{$this->domain}' is not recognized." );
 200                 }
 201
 202                 if ( PHP_SAPI === 'cli' || PHP_SAPI === 'phpdbg' ) {
 203                         $this->push( $jobs );
 204                         return;
 205                 }
 206
 207                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 208
 209                 // Throw errors now instead of on push(), when other jobs may be buffered
 210                 $this->assertValidJobs( $jobs );
 211
 212                 $this->bufferedJobs = array_merge( $this->bufferedJobs, $jobs );
 213         }
 214
 215         /**
 216          * Push all jobs buffered via lazyPush() into their respective queues
 217          *
 218          * @return void
 219          * @since 1.26
 220          */
 221         public static function pushLazyJobs() {
 222                 foreach ( self::$instances as $group ) {
 223                         try {
 224                                 $group->push( $group->bufferedJobs );
 225                                 $group->bufferedJobs = [];
 226                         } catch ( Exception $e ) {
 227                                 // Get in as many jobs as possible and let other post-send updates happen
 228                                 MWExceptionHandler::logException( $e );
 229                         }
 230                 }
 231         }
 232
 233         /**
 234          * Pop a job off one of the job queues
 235          *
 236          * This pops a job off a queue as specified by $wgJobTypeConf and
 237          * updates the aggregate job queue information cache as needed.
 238          *
 239          * @param int|string $qtype JobQueueGroup::TYPE_* constant or job type string
 240          * @param int $flags Bitfield of JobQueueGroup::USE_* constants
 241          * @param array $blacklist List of job types to ignore
 242          * @return Job|bool Returns false on failure
 243          */
 244         public function pop( $qtype = self::TYPE_DEFAULT, $flags = 0, array $blacklist = [] ) {
 245                 $job = false;
 246
 247                 if ( is_string( $qtype ) ) { // specific job type
 248                         if ( !in_array( $qtype, $blacklist ) ) {
 249                                 $job = $this->get( $qtype )->pop();
 250                         }
 251                 } else { // any job in the "default" jobs types
 252                         if ( $flags & self::USE_CACHE ) {
 253                                 if ( !$this->cache->has( 'queues-ready', 'list', self::PROC_CACHE_TTL ) ) {
 254                                         $this->cache->set( 'queues-ready', 'list', $this->getQueuesWithJobs() );
 255                                 }
 256                                 $types = $this->cache->get( 'queues-ready', 'list' );
 257                         } else {
 258                                 $types = $this->getQueuesWithJobs();
 259                         }
 260
 261                         if ( $qtype == self::TYPE_DEFAULT ) {
 262                                 $types = array_intersect( $types, $this->getDefaultQueueTypes() );
 263                         }
 264
 265                         $types = array_diff( $types, $blacklist ); // avoid selected types
 266                         shuffle( $types ); // avoid starvation
 267
 268                         foreach ( $types as $type ) { // for each queue...
 269                                 $job = $this->get( $type )->pop();
 270                                 if ( $job ) { // found
 271                                         break;
 272                                 } else { // not found
 273                                         $this->cache->clear( 'queues-ready' );
 274                                 }
 275                         }
 276                 }
 277
 278                 return $job;
 279         }
 280
 281         /**
 282          * Acknowledge that a job was completed
 283          *
 284          * @param Job $job
 285          * @return void
 286          */
 287         public function ack( Job $job ) {
 288                 $this->get( $job->getType() )->ack( $job );
 289         }
 290
 291         /**
 292          * Register the "root job" of a given job into the queue for de-duplication.
 293          * This should only be called right *after* all the new jobs have been inserted.
 294          *
 295          * @param Job $job
 296          * @return bool
 297          */
 298         public function deduplicateRootJob( Job $job ) {
 299                 return $this->get( $job->getType() )->deduplicateRootJob( $job );
 300         }
 301
 302         /**
 303          * Wait for any replica DBs or backup queue servers to catch up.
 304          *
 305          * This does nothing for certain queue classes.
 306          *
 307          * @return void
 308          */
 309         public function waitForBackups() {
 310                 global $wgJobTypeConf;
 311
 312                 // Try to avoid doing this more than once per queue storage medium
 313                 foreach ( $wgJobTypeConf as $type => $conf ) {
 314                         $this->get( $type )->waitForBackups();
 315                 }
 316         }
 317
 318         /**
 319          * Get the list of queue types
 320          *
 321          * @return array List of strings
 322          */
 323         public function getQueueTypes() {
 324                 return array_keys( $this->getCachedConfigVar( 'wgJobClasses' ) );
 325         }
 326
 327         /**
 328          * Get the list of default queue types
 329          *
 330          * @return array List of strings
 331          */
 332         public function getDefaultQueueTypes() {
 333                 global $wgJobTypesExcludedFromDefaultQueue;
 334
 335                 return array_diff( $this->getQueueTypes(), $wgJobTypesExcludedFromDefaultQueue );
 336         }
 337
 338         /**
 339          * Check if there are any queues with jobs (this is cached)
 340          *
 341          * @param int $type JobQueueGroup::TYPE_* constant
 342          * @return bool
 343          * @since 1.23
 344          */
 345         public function queuesHaveJobs( $type = self::TYPE_ANY ) {
 346                 $cache = ObjectCache::getLocalClusterInstance();
 347                 $key = $cache->makeGlobalKey( 'jobqueue', $this->domain, 'hasjobs', $type );
 348
 349                 $value = $cache->get( $key );
 350                 if ( $value === false ) {
 351                         $queues = $this->getQueuesWithJobs();
 352                         if ( $type == self::TYPE_DEFAULT ) {
 353                                 $queues = array_intersect( $queues, $this->getDefaultQueueTypes() );
 354                         }
 355                         $value = count( $queues ) ? 'true' : 'false';
 356                         $cache->add( $key, $value, 15 );
 357                 }
 358
 359                 return ( $value === 'true' );
 360         }
 361
 362         /**
 363          * Get the list of job types that have non-empty queues
 364          *
 365          * @return array List of job types that have non-empty queues
 366          */
 367         public function getQueuesWithJobs() {
 368                 $types = [];
 369                 foreach ( $this->getCoalescedQueues() as $info ) {
 370                         $nonEmpty = $info['queue']->getSiblingQueuesWithJobs( $this->getQueueTypes() );
 371                         if ( is_array( $nonEmpty ) ) { // batching features supported
 372                                 $types = array_merge( $types, $nonEmpty );
 373                         } else { // we have to go through the queues in the bucket one-by-one
 374                                 foreach ( $info['types'] as $type ) {
 375                                         if ( !$this->get( $type )->isEmpty() ) {
 376                                                 $types[] = $type;
 377                                         }
 378                                 }
 379                         }
 380                 }
 381
 382                 return $types;
 383         }
 384
 385         /**
 386          * Get the size of the queus for a list of job types
 387          *
 388          * @return array Map of (job type => size)
 389          */
 390         public function getQueueSizes() {
 391                 $sizeMap = [];
 392                 foreach ( $this->getCoalescedQueues() as $info ) {
 393                         $sizes = $info['queue']->getSiblingQueueSizes( $this->getQueueTypes() );
 394                         if ( is_array( $sizes ) ) { // batching features supported
 395                                 $sizeMap = $sizeMap + $sizes;
 396                         } else { // we have to go through the queues in the bucket one-by-one
 397                                 foreach ( $info['types'] as $type ) {
 398                                         $sizeMap[$type] = $this->get( $type )->getSize();
 399                                 }
 400                         }
 401                 }
 402
 403                 return $sizeMap;
 404         }
 405
 406         /**
 407          * @return array
 408          */
 409         protected function getCoalescedQueues() {
 410                 global $wgJobTypeConf;
 411
 412                 if ( $this->coalescedQueues === null ) {
 413                         $this->coalescedQueues = [];
 414                         foreach ( $wgJobTypeConf as $type => $conf ) {
 415                                 $queue = JobQueue::factory(
 416                                         [ 'wiki' => $this->domain, 'type' => 'null' ] + $conf );
 417                                 $loc = $queue->getCoalesceLocationInternal();
 418                                 if ( !isset( $this->coalescedQueues[$loc] ) ) {
 419                                         $this->coalescedQueues[$loc]['queue'] = $queue;
 420                                         $this->coalescedQueues[$loc]['types'] = [];
 421                                 }
 422                                 if ( $type === 'default' ) {
 423                                         $this->coalescedQueues[$loc]['types'] = array_merge(
 424                                                 $this->coalescedQueues[$loc]['types'],
 425                                                 array_diff( $this->getQueueTypes(), array_keys( $wgJobTypeConf ) )
 426                                         );
 427                                 } else {
 428                                         $this->coalescedQueues[$loc]['types'][] = $type;
 429                                 }
 430                         }
 431                 }
 432
 433                 return $this->coalescedQueues;
 434         }
 435
 436         /**
 437          * @param string $name
 438          * @return mixed
 439          */
 440         private function getCachedConfigVar( $name ) {
 441                 // @TODO: cleanup this whole method with a proper config system
 442                 if ( WikiMap::isCurrentWikiDbDomain( $this->domain ) ) {
 443                         return $GLOBALS[$name]; // common case
 444                 } else {
 445                         $wiki = WikiMap::getWikiIdFromDomain( $this->domain );
 446                         $cache = ObjectCache::getMainWANInstance();
 447                         $value = $cache->getWithSetCallback(
 448                                 $cache->makeGlobalKey( 'jobqueue', 'configvalue', $this->domain, $name ),
 449                                 $cache::TTL_DAY + mt_rand( 0, $cache::TTL_DAY ),
 450                                 function () use ( $wiki, $name ) {
 451                                         global $wgConf;
 452                                         // @TODO: use the full domain ID here
 453                                         return [ 'v' => $wgConf->getConfig( $wiki, $name ) ];
 454                                 },
 455                                 [ 'pcTTL' => WANObjectCache::TTL_PROC_LONG ]
 456                         );
 457                         return $value['v'];
 458                 }
 459         }
 460
 461         /**
 462          * @param array $jobs
 463          * @throws InvalidArgumentException
 464          */
 465         private function assertValidJobs( array $jobs ) {
 466                 foreach ( $jobs as $job ) { // sanity checks
 467                         if ( !( $job instanceof IJobSpecification ) ) {
 468                                 throw new InvalidArgumentException( "Expected IJobSpecification objects" );
 469                         }
 470                 }
 471         }
 472
 473         function __destruct() {
 474                 $n = count( $this->bufferedJobs );
 475                 if ( $n > 0 ) {
 476                         $type = implode( ', ', array_unique( array_map( 'get_class', $this->bufferedJobs ) ) );
 477                         trigger_error( __METHOD__ . ": $n buffered job(s) of type(s) $type never inserted." );
 478                 }
 479         }
 480 }