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 ( $wiki !== wfWikiID() && !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                 if ( isset( $wgJobTypeConf[$type] ) ) {
 106                         $conf = $conf + $wgJobTypeConf[$type];
 107                 } else {
 108                         $conf = $conf + $wgJobTypeConf['default'];
 109                 }
 110                 $conf['aggregator'] = JobQueueAggregator::singleton();
 111                 if ( !isset( $conf['readOnlyReason'] ) ) {
 112                         $conf['readOnlyReason'] = $this->readOnlyReason;
 113                 }
 114
 115                 return JobQueue::factory( $conf );
 116         }
 117
 118         /**
 119          * Insert jobs into the respective queues of which they belong
 120          *
 121          * This inserts the jobs into the queue specified by $wgJobTypeConf
 122          * and updates the aggregate job queue information cache as needed.
 123          *
 124          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 125          * @throws InvalidArgumentException
 126          * @return void
 127          */
 128         public function push( $jobs ) {
 129                 global $wgJobTypesExcludedFromDefaultQueue;
 130
 131                 if ( $this->invalidWiki ) {
 132                         // Do not enqueue job that cannot be run (T171371)
 133                         $e = new LogicException( "Domain '{$this->wiki}' is not recognized." );
 134                         MWExceptionHandler::logException( $e );
 135                         return;
 136                 }
 137
 138                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 139                 if ( !count( $jobs ) ) {
 140                         return;
 141                 }
 142
 143                 $this->assertValidJobs( $jobs );
 144
 145                 $jobsByType = []; // (job type => list of jobs)
 146                 foreach ( $jobs as $job ) {
 147                         $jobsByType[$job->getType()][] = $job;
 148                 }
 149
 150                 foreach ( $jobsByType as $type => $jobs ) {
 151                         $this->get( $type )->push( $jobs );
 152                 }
 153
 154                 if ( $this->cache->hasField( 'queues-ready', 'list' ) ) {
 155                         $list = $this->cache->getField( 'queues-ready', 'list' );
 156                         if ( count( array_diff( array_keys( $jobsByType ), $list ) ) ) {
 157                                 $this->cache->clear( 'queues-ready' );
 158                         }
 159                 }
 160
 161                 $cache = ObjectCache::getLocalClusterInstance();
 162                 $cache->set(
 163                         $cache->makeGlobalKey( 'jobqueue', $this->wiki, 'hasjobs', self::TYPE_ANY ),
 164                         'true',
 165                         15
 166                 );
 167                 if ( array_diff( array_keys( $jobsByType ), $wgJobTypesExcludedFromDefaultQueue ) ) {
 168                         $cache->set(
 169                                 $cache->makeGlobalKey( 'jobqueue', $this->wiki, 'hasjobs', self::TYPE_DEFAULT ),
 170                                 'true',
 171                                 15
 172                         );
 173                 }
 174         }
 175
 176         /**
 177          * Buffer jobs for insertion via push() or call it now if in CLI mode
 178          *
 179          * Note that pushLazyJobs() is registered as a deferred update just before
 180          * DeferredUpdates::doUpdates() in MediaWiki and JobRunner classes in order
 181          * to be executed as the very last deferred update (T100085, T154425).
 182          *
 183          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 184          * @return void
 185          * @since 1.26
 186          */
 187         public function lazyPush( $jobs ) {
 188                 if ( $this->invalidWiki ) {
 189                         // Do not enqueue job that cannot be run (T171371)
 190                         throw new LogicException( "Domain '{$this->wiki}' is not recognized." );
 191                 }
 192
 193                 if ( PHP_SAPI === 'cli' || PHP_SAPI === 'phpdbg' ) {
 194                         $this->push( $jobs );
 195                         return;
 196                 }
 197
 198                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 199
 200                 // Throw errors now instead of on push(), when other jobs may be buffered
 201                 $this->assertValidJobs( $jobs );
 202
 203                 DeferredUpdates::addUpdate( new JobQueueEnqueueUpdate( $this->wiki, $jobs ) );
 204         }
 205
 206         /**
 207          * Push all jobs buffered via lazyPush() into their respective queues
 208          *
 209          * @return void
 210          * @since 1.26
 211          * @deprecated Since 1.33 Not needed anymore
 212          */
 213         public static function pushLazyJobs() {
 214                 wfDeprecated( __METHOD__, '1.33' );
 215         }
 216
 217         /**
 218          * Pop a job off one of the job queues
 219          *
 220          * This pops a job off a queue as specified by $wgJobTypeConf and
 221          * updates the aggregate job queue information cache as needed.
 222          *
 223          * @param int|string $qtype JobQueueGroup::TYPE_* constant or job type string
 224          * @param int $flags Bitfield of JobQueueGroup::USE_* constants
 225          * @param array $blacklist List of job types to ignore
 226          * @return Job|bool Returns false on failure
 227          */
 228         public function pop( $qtype = self::TYPE_DEFAULT, $flags = 0, array $blacklist = [] ) {
 229                 $job = false;
 230
 231                 if ( is_string( $qtype ) ) { // specific job type
 232                         if ( !in_array( $qtype, $blacklist ) ) {
 233                                 $job = $this->get( $qtype )->pop();
 234                         }
 235                 } else { // any job in the "default" jobs types
 236                         if ( $flags & self::USE_CACHE ) {
 237                                 if ( !$this->cache->hasField( 'queues-ready', 'list', self::PROC_CACHE_TTL ) ) {
 238                                         $this->cache->setField( 'queues-ready', 'list', $this->getQueuesWithJobs() );
 239                                 }
 240                                 $types = $this->cache->getField( 'queues-ready', 'list' );
 241                         } else {
 242                                 $types = $this->getQueuesWithJobs();
 243                         }
 244
 245                         if ( $qtype == self::TYPE_DEFAULT ) {
 246                                 $types = array_intersect( $types, $this->getDefaultQueueTypes() );
 247                         }
 248
 249                         $types = array_diff( $types, $blacklist ); // avoid selected types
 250                         shuffle( $types ); // avoid starvation
 251
 252                         foreach ( $types as $type ) { // for each queue...
 253                                 $job = $this->get( $type )->pop();
 254                                 if ( $job ) { // found
 255                                         break;
 256                                 } else { // not found
 257                                         $this->cache->clear( 'queues-ready' );
 258                                 }
 259                         }
 260                 }
 261
 262                 return $job;
 263         }
 264
 265         /**
 266          * Acknowledge that a job was completed
 267          *
 268          * @param Job $job
 269          * @return void
 270          */
 271         public function ack( Job $job ) {
 272                 $this->get( $job->getType() )->ack( $job );
 273         }
 274
 275         /**
 276          * Register the "root job" of a given job into the queue for de-duplication.
 277          * This should only be called right *after* all the new jobs have been inserted.
 278          *
 279          * @param Job $job
 280          * @return bool
 281          */
 282         public function deduplicateRootJob( Job $job ) {
 283                 return $this->get( $job->getType() )->deduplicateRootJob( $job );
 284         }
 285
 286         /**
 287          * Wait for any replica DBs or backup queue servers to catch up.
 288          *
 289          * This does nothing for certain queue classes.
 290          *
 291          * @return void
 292          */
 293         public function waitForBackups() {
 294                 global $wgJobTypeConf;
 295
 296                 // Try to avoid doing this more than once per queue storage medium
 297                 foreach ( $wgJobTypeConf as $type => $conf ) {
 298                         $this->get( $type )->waitForBackups();
 299                 }
 300         }
 301
 302         /**
 303          * Get the list of queue types
 304          *
 305          * @return array List of strings
 306          */
 307         public function getQueueTypes() {
 308                 return array_keys( $this->getCachedConfigVar( 'wgJobClasses' ) );
 309         }
 310
 311         /**
 312          * Get the list of default queue types
 313          *
 314          * @return array List of strings
 315          */
 316         public function getDefaultQueueTypes() {
 317                 global $wgJobTypesExcludedFromDefaultQueue;
 318
 319                 return array_diff( $this->getQueueTypes(), $wgJobTypesExcludedFromDefaultQueue );
 320         }
 321
 322         /**
 323          * Check if there are any queues with jobs (this is cached)
 324          *
 325          * @param int $type JobQueueGroup::TYPE_* constant
 326          * @return bool
 327          * @since 1.23
 328          */
 329         public function queuesHaveJobs( $type = self::TYPE_ANY ) {
 330                 $cache = ObjectCache::getLocalClusterInstance();
 331                 $key = $cache->makeGlobalKey( 'jobqueue', $this->wiki, 'hasjobs', $type );
 332
 333                 $value = $cache->get( $key );
 334                 if ( $value === false ) {
 335                         $queues = $this->getQueuesWithJobs();
 336                         if ( $type == self::TYPE_DEFAULT ) {
 337                                 $queues = array_intersect( $queues, $this->getDefaultQueueTypes() );
 338                         }
 339                         $value = count( $queues ) ? 'true' : 'false';
 340                         $cache->add( $key, $value, 15 );
 341                 }
 342
 343                 return ( $value === 'true' );
 344         }
 345
 346         /**
 347          * Get the list of job types that have non-empty queues
 348          *
 349          * @return array List of job types that have non-empty queues
 350          */
 351         public function getQueuesWithJobs() {
 352                 $types = [];
 353                 foreach ( $this->getCoalescedQueues() as $info ) {
 354                         $nonEmpty = $info['queue']->getSiblingQueuesWithJobs( $this->getQueueTypes() );
 355                         if ( is_array( $nonEmpty ) ) { // batching features supported
 356                                 $types = array_merge( $types, $nonEmpty );
 357                         } else { // we have to go through the queues in the bucket one-by-one
 358                                 foreach ( $info['types'] as $type ) {
 359                                         if ( !$this->get( $type )->isEmpty() ) {
 360                                                 $types[] = $type;
 361                                         }
 362                                 }
 363                         }
 364                 }
 365
 366                 return $types;
 367         }
 368
 369         /**
 370          * Get the size of the queus for a list of job types
 371          *
 372          * @return array Map of (job type => size)
 373          */
 374         public function getQueueSizes() {
 375                 $sizeMap = [];
 376                 foreach ( $this->getCoalescedQueues() as $info ) {
 377                         $sizes = $info['queue']->getSiblingQueueSizes( $this->getQueueTypes() );
 378                         if ( is_array( $sizes ) ) { // batching features supported
 379                                 $sizeMap = $sizeMap + $sizes;
 380                         } else { // we have to go through the queues in the bucket one-by-one
 381                                 foreach ( $info['types'] as $type ) {
 382                                         $sizeMap[$type] = $this->get( $type )->getSize();
 383                                 }
 384                         }
 385                 }
 386
 387                 return $sizeMap;
 388         }
 389
 390         /**
 391          * @return array
 392          */
 393         protected function getCoalescedQueues() {
 394                 global $wgJobTypeConf;
 395
 396                 if ( $this->coalescedQueues === null ) {
 397                         $this->coalescedQueues = [];
 398                         foreach ( $wgJobTypeConf as $type => $conf ) {
 399                                 $queue = JobQueue::factory(
 400                                         [ 'wiki' => $this->wiki, 'type' => 'null' ] + $conf );
 401                                 $loc = $queue->getCoalesceLocationInternal();
 402                                 if ( !isset( $this->coalescedQueues[$loc] ) ) {
 403                                         $this->coalescedQueues[$loc]['queue'] = $queue;
 404                                         $this->coalescedQueues[$loc]['types'] = [];
 405                                 }
 406                                 if ( $type === 'default' ) {
 407                                         $this->coalescedQueues[$loc]['types'] = array_merge(
 408                                                 $this->coalescedQueues[$loc]['types'],
 409                                                 array_diff( $this->getQueueTypes(), array_keys( $wgJobTypeConf ) )
 410                                         );
 411                                 } else {
 412                                         $this->coalescedQueues[$loc]['types'][] = $type;
 413                                 }
 414                         }
 415                 }
 416
 417                 return $this->coalescedQueues;
 418         }
 419
 420         /**
 421          * @param string $name
 422          * @return mixed
 423          */
 424         private function getCachedConfigVar( $name ) {
 425                 // @TODO: cleanup this whole method with a proper config system
 426                 if ( $this->wiki === wfWikiID() ) {
 427                         return $GLOBALS[$name]; // common case
 428                 } else {
 429                         $wiki = $this->wiki;
 430                         $cache = ObjectCache::getMainWANInstance();
 431                         $value = $cache->getWithSetCallback(
 432                                 $cache->makeGlobalKey( 'jobqueue', 'configvalue', $wiki, $name ),
 433                                 $cache::TTL_DAY + mt_rand( 0, $cache::TTL_DAY ),
 434                                 function () use ( $wiki, $name ) {
 435                                         global $wgConf;
 436
 437                                         return [ 'v' => $wgConf->getConfig( $wiki, $name ) ];
 438                                 },
 439                                 [ 'pcTTL' => WANObjectCache::TTL_PROC_LONG ]
 440                         );
 441
 442                         return $value['v'];
 443                 }
 444         }
 445
 446         /**
 447          * @param array $jobs
 448          * @throws InvalidArgumentException
 449          */
 450         private function assertValidJobs( array $jobs ) {
 451                 foreach ( $jobs as $job ) { // sanity checks
 452                         if ( !( $job instanceof IJobSpecification ) ) {
 453                                 throw new InvalidArgumentException( "Expected IJobSpecification objects" );
 454                         }
 455                 }
 456         }
 457 }