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  * @author Aaron Schulz
  22  */
  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 ID */
  38         protected $wiki;
  39         /** @var string|bool Read only rationale (or false if r/w) */
  40         protected $readOnlyReason;
  41
  42         /** @var array Map of (bucket => (queue => JobQueue, types => list of types) */
  43         protected $coalescedQueues;
  44
  45         /** @var Job[] */
  46         protected $bufferedJobs = [];
  47
  48         const TYPE_DEFAULT = 1; // integer; jobs popped by default
  49         const TYPE_ANY = 2; // integer; any job
  50
  51         const USE_CACHE = 1; // integer; use process or persistent cache
  52
  53         const PROC_CACHE_TTL = 15; // integer; seconds
  54
  55         const CACHE_VERSION = 1; // integer; cache version
  56
  57         /**
  58          * @param string $wiki Wiki ID
  59          * @param string|bool $readOnlyReason Read-only reason or false
  60          */
  61         protected function __construct( $wiki, $readOnlyReason ) {
  62                 $this->wiki = $wiki;
  63                 $this->readOnlyReason = $readOnlyReason;
  64                 $this->cache = new ProcessCacheLRU( 10 );
  65         }
  66
  67         /**
  68          * @param bool|string $wiki Wiki ID
  69          * @return JobQueueGroup
  70          */
  71         public static function singleton( $wiki = false ) {
  72                 $wiki = ( $wiki === false ) ? wfWikiID() : $wiki;
  73                 if ( !isset( self::$instances[$wiki] ) ) {
  74                         self::$instances[$wiki] = new self( $wiki, wfConfiguredReadOnlyReason() );
  75                 }
  76
  77                 return self::$instances[$wiki];
  78         }
  79
  80         /**
  81          * Destroy the singleton instances
  82          *
  83          * @return void
  84          */
  85         public static function destroySingletons() {
  86                 self::$instances = [];
  87         }
  88
  89         /**
  90          * Get the job queue object for a given queue type
  91          *
  92          * @param string $type
  93          * @return JobQueue
  94          */
  95         public function get( $type ) {
  96                 global $wgJobTypeConf;
  97
  98                 $conf = [ 'wiki' => $this->wiki, 'type' => $type ];
  99                 if ( isset( $wgJobTypeConf[$type] ) ) {
 100                         $conf = $conf + $wgJobTypeConf[$type];
 101                 } else {
 102                         $conf = $conf + $wgJobTypeConf['default'];
 103                 }
 104                 $conf['aggregator'] = JobQueueAggregator::singleton();
 105                 if ( $this->readOnlyReason !== false ) {
 106                         $conf['readOnlyReason'] = $this->readOnlyReason;
 107                 }
 108
 109                 return JobQueue::factory( $conf );
 110         }
 111
 112         /**
 113          * Insert jobs into the respective queues of which they belong
 114          *
 115          * This inserts the jobs into the queue specified by $wgJobTypeConf
 116          * and updates the aggregate job queue information cache as needed.
 117          *
 118          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 119          * @throws InvalidArgumentException
 120          * @return void
 121          */
 122         public function push( $jobs ) {
 123                 global $wgJobTypesExcludedFromDefaultQueue;
 124
 125                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 126                 if ( !count( $jobs ) ) {
 127                         return;
 128                 }
 129
 130                 $this->assertValidJobs( $jobs );
 131
 132                 $jobsByType = []; // (job type => list of jobs)
 133                 foreach ( $jobs as $job ) {
 134                         $jobsByType[$job->getType()][] = $job;
 135                 }
 136
 137                 foreach ( $jobsByType as $type => $jobs ) {
 138                         $this->get( $type )->push( $jobs );
 139                 }
 140
 141                 if ( $this->cache->has( 'queues-ready', 'list' ) ) {
 142                         $list = $this->cache->get( 'queues-ready', 'list' );
 143                         if ( count( array_diff( array_keys( $jobsByType ), $list ) ) ) {
 144                                 $this->cache->clear( 'queues-ready' );
 145                         }
 146                 }
 147
 148                 $cache = ObjectCache::getLocalClusterInstance();
 149                 $cache->set(
 150                         $cache->makeGlobalKey( 'jobqueue', $this->wiki, 'hasjobs', self::TYPE_ANY ),
 151                         'true',
 152                         15
 153                 );
 154                 if ( array_diff( array_keys( $jobsByType ), $wgJobTypesExcludedFromDefaultQueue ) ) {
 155                         $cache->set(
 156                                 $cache->makeGlobalKey( 'jobqueue', $this->wiki, 'hasjobs', self::TYPE_DEFAULT ),
 157                                 'true',
 158                                 15
 159                         );
 160                 }
 161         }
 162
 163         /**
 164          * Buffer jobs for insertion via push() or call it now if in CLI mode
 165          *
 166          * Note that pushLazyJobs() is registered as a deferred update just before
 167          * DeferredUpdates::doUpdates() in MediaWiki and JobRunner classes in order
 168          * to be executed as the very last deferred update (T100085, T154425).
 169          *
 170          * @param IJobSpecification|IJobSpecification[] $jobs A single Job or a list of Jobs
 171          * @return void
 172          * @since 1.26
 173          */
 174         public function lazyPush( $jobs ) {
 175                 if ( PHP_SAPI === 'cli' || PHP_SAPI === 'phpdbg' ) {
 176                         $this->push( $jobs );
 177                         return;
 178                 }
 179
 180                 $jobs = is_array( $jobs ) ? $jobs : [ $jobs ];
 181
 182                 // Throw errors now instead of on push(), when other jobs may be buffered
 183                 $this->assertValidJobs( $jobs );
 184
 185                 $this->bufferedJobs = array_merge( $this->bufferedJobs, $jobs );
 186         }
 187
 188         /**
 189          * Push all jobs buffered via lazyPush() into their respective queues
 190          *
 191          * @return void
 192          * @since 1.26
 193          */
 194         public static function pushLazyJobs() {
 195                 foreach ( self::$instances as $group ) {
 196                         try {
 197                                 $group->push( $group->bufferedJobs );
 198                                 $group->bufferedJobs = [];
 199                         } catch ( Exception $e ) {
 200                                 // Get in as many jobs as possible and let other post-send updates happen
 201                                 MWExceptionHandler::logException( $e );
 202                         }
 203                 }
 204         }
 205
 206         /**
 207          * Pop a job off one of the job queues
 208          *
 209          * This pops a job off a queue as specified by $wgJobTypeConf and
 210          * updates the aggregate job queue information cache as needed.
 211          *
 212          * @param int|string $qtype JobQueueGroup::TYPE_* constant or job type string
 213          * @param int $flags Bitfield of JobQueueGroup::USE_* constants
 214          * @param array $blacklist List of job types to ignore
 215          * @return Job|bool Returns false on failure
 216          */
 217         public function pop( $qtype = self::TYPE_DEFAULT, $flags = 0, array $blacklist = [] ) {
 218                 $job = false;
 219
 220                 if ( is_string( $qtype ) ) { // specific job type
 221                         if ( !in_array( $qtype, $blacklist ) ) {
 222                                 $job = $this->get( $qtype )->pop();
 223                         }
 224                 } else { // any job in the "default" jobs types
 225                         if ( $flags & self::USE_CACHE ) {
 226                                 if ( !$this->cache->has( 'queues-ready', 'list', self::PROC_CACHE_TTL ) ) {
 227                                         $this->cache->set( 'queues-ready', 'list', $this->getQueuesWithJobs() );
 228                                 }
 229                                 $types = $this->cache->get( 'queues-ready', 'list' );
 230                         } else {
 231                                 $types = $this->getQueuesWithJobs();
 232                         }
 233
 234                         if ( $qtype == self::TYPE_DEFAULT ) {
 235                                 $types = array_intersect( $types, $this->getDefaultQueueTypes() );
 236                         }
 237
 238                         $types = array_diff( $types, $blacklist ); // avoid selected types
 239                         shuffle( $types ); // avoid starvation
 240
 241                         foreach ( $types as $type ) { // for each queue...
 242                                 $job = $this->get( $type )->pop();
 243                                 if ( $job ) { // found
 244                                         break;
 245                                 } else { // not found
 246                                         $this->cache->clear( 'queues-ready' );
 247                                 }
 248                         }
 249                 }
 250
 251                 return $job;
 252         }
 253
 254         /**
 255          * Acknowledge that a job was completed
 256          *
 257          * @param Job $job
 258          * @return void
 259          */
 260         public function ack( Job $job ) {
 261                 $this->get( $job->getType() )->ack( $job );
 262         }
 263
 264         /**
 265          * Register the "root job" of a given job into the queue for de-duplication.
 266          * This should only be called right *after* all the new jobs have been inserted.
 267          *
 268          * @param Job $job
 269          * @return bool
 270          */
 271         public function deduplicateRootJob( Job $job ) {
 272                 return $this->get( $job->getType() )->deduplicateRootJob( $job );
 273         }
 274
 275         /**
 276          * Wait for any replica DBs or backup queue servers to catch up.
 277          *
 278          * This does nothing for certain queue classes.
 279          *
 280          * @return void
 281          */
 282         public function waitForBackups() {
 283                 global $wgJobTypeConf;
 284
 285                 // Try to avoid doing this more than once per queue storage medium
 286                 foreach ( $wgJobTypeConf as $type => $conf ) {
 287                         $this->get( $type )->waitForBackups();
 288                 }
 289         }
 290
 291         /**
 292          * Get the list of queue types
 293          *
 294          * @return array List of strings
 295          */
 296         public function getQueueTypes() {
 297                 return array_keys( $this->getCachedConfigVar( 'wgJobClasses' ) );
 298         }
 299
 300         /**
 301          * Get the list of default queue types
 302          *
 303          * @return array List of strings
 304          */
 305         public function getDefaultQueueTypes() {
 306                 global $wgJobTypesExcludedFromDefaultQueue;
 307
 308                 return array_diff( $this->getQueueTypes(), $wgJobTypesExcludedFromDefaultQueue );
 309         }
 310
 311         /**
 312          * Check if there are any queues with jobs (this is cached)
 313          *
 314          * @param int $type JobQueueGroup::TYPE_* constant
 315          * @return bool
 316          * @since 1.23
 317          */
 318         public function queuesHaveJobs( $type = self::TYPE_ANY ) {
 319                 $cache = ObjectCache::getLocalClusterInstance();
 320                 $key = $cache->makeGlobalKey( 'jobqueue', $this->wiki, 'hasjobs', $type );
 321
 322                 $value = $cache->get( $key );
 323                 if ( $value === false ) {
 324                         $queues = $this->getQueuesWithJobs();
 325                         if ( $type == self::TYPE_DEFAULT ) {
 326                                 $queues = array_intersect( $queues, $this->getDefaultQueueTypes() );
 327                         }
 328                         $value = count( $queues ) ? 'true' : 'false';
 329                         $cache->add( $key, $value, 15 );
 330                 }
 331
 332                 return ( $value === 'true' );
 333         }
 334
 335         /**
 336          * Get the list of job types that have non-empty queues
 337          *
 338          * @return array List of job types that have non-empty queues
 339          */
 340         public function getQueuesWithJobs() {
 341                 $types = [];
 342                 foreach ( $this->getCoalescedQueues() as $info ) {
 343                         $nonEmpty = $info['queue']->getSiblingQueuesWithJobs( $this->getQueueTypes() );
 344                         if ( is_array( $nonEmpty ) ) { // batching features supported
 345                                 $types = array_merge( $types, $nonEmpty );
 346                         } else { // we have to go through the queues in the bucket one-by-one
 347                                 foreach ( $info['types'] as $type ) {
 348                                         if ( !$this->get( $type )->isEmpty() ) {
 349                                                 $types[] = $type;
 350                                         }
 351                                 }
 352                         }
 353                 }
 354
 355                 return $types;
 356         }
 357
 358         /**
 359          * Get the size of the queus for a list of job types
 360          *
 361          * @return array Map of (job type => size)
 362          */
 363         public function getQueueSizes() {
 364                 $sizeMap = [];
 365                 foreach ( $this->getCoalescedQueues() as $info ) {
 366                         $sizes = $info['queue']->getSiblingQueueSizes( $this->getQueueTypes() );
 367                         if ( is_array( $sizes ) ) { // batching features supported
 368                                 $sizeMap = $sizeMap + $sizes;
 369                         } else { // we have to go through the queues in the bucket one-by-one
 370                                 foreach ( $info['types'] as $type ) {
 371                                         $sizeMap[$type] = $this->get( $type )->getSize();
 372                                 }
 373                         }
 374                 }
 375
 376                 return $sizeMap;
 377         }
 378
 379         /**
 380          * @return array
 381          */
 382         protected function getCoalescedQueues() {
 383                 global $wgJobTypeConf;
 384
 385                 if ( $this->coalescedQueues === null ) {
 386                         $this->coalescedQueues = [];
 387                         foreach ( $wgJobTypeConf as $type => $conf ) {
 388                                 $queue = JobQueue::factory(
 389                                         [ 'wiki' => $this->wiki, 'type' => 'null' ] + $conf );
 390                                 $loc = $queue->getCoalesceLocationInternal();
 391                                 if ( !isset( $this->coalescedQueues[$loc] ) ) {
 392                                         $this->coalescedQueues[$loc]['queue'] = $queue;
 393                                         $this->coalescedQueues[$loc]['types'] = [];
 394                                 }
 395                                 if ( $type === 'default' ) {
 396                                         $this->coalescedQueues[$loc]['types'] = array_merge(
 397                                                 $this->coalescedQueues[$loc]['types'],
 398                                                 array_diff( $this->getQueueTypes(), array_keys( $wgJobTypeConf ) )
 399                                         );
 400                                 } else {
 401                                         $this->coalescedQueues[$loc]['types'][] = $type;
 402                                 }
 403                         }
 404                 }
 405
 406                 return $this->coalescedQueues;
 407         }
 408
 409         /**
 410          * @param string $name
 411          * @return mixed
 412          */
 413         private function getCachedConfigVar( $name ) {
 414                 // @TODO: cleanup this whole method with a proper config system
 415                 if ( $this->wiki === wfWikiID() ) {
 416                         return $GLOBALS[$name]; // common case
 417                 } else {
 418                         $wiki = $this->wiki;
 419                         $cache = ObjectCache::getMainWANInstance();
 420                         $value = $cache->getWithSetCallback(
 421                                 $cache->makeGlobalKey( 'jobqueue', 'configvalue', $wiki, $name ),
 422                                 $cache::TTL_DAY + mt_rand( 0, $cache::TTL_DAY ),
 423                                 function () use ( $wiki, $name ) {
 424                                         global $wgConf;
 425
 426                                         return [ 'v' => $wgConf->getConfig( $wiki, $name ) ];
 427                                 },
 428                                 [ 'pcTTL' => WANObjectCache::TTL_PROC_LONG ]
 429                         );
 430
 431                         return $value['v'];
 432                 }
 433         }
 434
 435         /**
 436          * @param array $jobs
 437          * @throws InvalidArgumentException
 438          */
 439         private function assertValidJobs( array $jobs ) {
 440                 foreach ( $jobs as $job ) { // sanity checks
 441                         if ( !( $job instanceof IJobSpecification ) ) {
 442                                 throw new InvalidArgumentException( "Expected IJobSpecification objects" );
 443                         }
 444                 }
 445         }
 446
 447         function __destruct() {
 448                 $n = count( $this->bufferedJobs );
 449                 if ( $n > 0 ) {
 450                         $type = implode( ', ', array_unique( array_map( 'get_class', $this->bufferedJobs ) ) );
 451                         trigger_error( __METHOD__ . ": $n buffered job(s) of type(s) $type never inserted." );
 452                 }
 453         }
 454 }