includes/jobqueue/jobs/RefreshLinksJob.php

   1 <?php
   2 /**
   3  * Job to update link tables for pages
   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  * @ingroup JobQueue
  22  */
  23 use MediaWiki\MediaWikiServices;
  24 use MediaWiki\Revision\RevisionRecord;
  25 use MediaWiki\Revision\RevisionRenderer;
  26 use Liuggio\StatsdClient\Factory\StatsdDataFactoryInterface;
  27
  28 /**
  29  * Job to update link tables for pages
  30  *
  31  * This job comes in a few variants:
  32  *   - a) Recursive jobs to update links for backlink pages for a given title.
  33  *        These jobs have (recursive:true,table:<table>) set.
  34  *   - b) Jobs to update links for a set of pages (the job title is ignored).
  35  *        These jobs have (pages:(<page ID>:(<namespace>,<title>),...) set.
  36  *   - c) Jobs to update links for a single page (the job title)
  37  *        These jobs need no extra fields set.
  38  *
  39  * @ingroup JobQueue
  40  */
  41 class RefreshLinksJob extends Job {
  42         /** @var int Lag safety margin when comparing root job times to last-refresh times */
  43         const NORMAL_MAX_LAG = 10;
  44         /** @var int How many seconds to wait for replica DBs to catch up */
  45         const LAG_WAIT_TIMEOUT = 15;
  46
  47         function __construct( Title $title, array $params ) {
  48                 parent::__construct( 'refreshLinks', $title, $params );
  49                 // Avoid the overhead of de-duplication when it would be pointless
  50                 $this->removeDuplicates = (
  51                         // Ranges rarely will line up
  52                         !isset( $params['range'] ) &&
  53                         // Multiple pages per job make matches unlikely
  54                         !( isset( $params['pages'] ) && count( $params['pages'] ) != 1 )
  55                 );
  56                 $this->params += [ 'causeAction' => 'unknown', 'causeAgent' => 'unknown' ];
  57                 // Tell JobRunner to not automatically wrap run() in a transaction round.
  58                 // Each runForTitle() call will manage its own rounds in order to run DataUpdates
  59                 // and to avoid contention as well.
  60                 $this->executionFlags |= self::JOB_NO_EXPLICIT_TRX_ROUND;
  61         }
  62
  63         /**
  64          * @param Title $title
  65          * @param array $params
  66          * @return RefreshLinksJob
  67          */
  68         public static function newPrioritized( Title $title, array $params ) {
  69                 $job = new self( $title, $params );
  70                 $job->command = 'refreshLinksPrioritized';
  71
  72                 return $job;
  73         }
  74
  75         /**
  76          * @param Title $title
  77          * @param array $params
  78          * @return RefreshLinksJob
  79          */
  80         public static function newDynamic( Title $title, array $params ) {
  81                 $job = new self( $title, $params );
  82                 $job->command = 'refreshLinksDynamic';
  83
  84                 return $job;
  85         }
  86
  87         function run() {
  88                 $ok = true;
  89
  90                 // Job to update all (or a range of) backlink pages for a page
  91                 if ( !empty( $this->params['recursive'] ) ) {
  92                         $services = MediaWikiServices::getInstance();
  93                         // When the base job branches, wait for the replica DBs to catch up to the master.
  94                         // From then on, we know that any template changes at the time the base job was
  95                         // enqueued will be reflected in backlink page parses when the leaf jobs run.
  96                         if ( !isset( $this->params['range'] ) ) {
  97                                 $lbFactory = $services->getDBLoadBalancerFactory();
  98                                 if ( !$lbFactory->waitForReplication( [
  99                                         'domain'  => $lbFactory->getLocalDomainID(),
 100                                         'timeout' => self::LAG_WAIT_TIMEOUT
 101                                 ] ) ) { // only try so hard
 102                                         $stats = $services->getStatsdDataFactory();
 103                                         $stats->increment( 'refreshlinks.lag_wait_failed' );
 104                                 }
 105                         }
 106                         // Carry over information for de-duplication
 107                         $extraParams = $this->getRootJobParams();
 108                         $extraParams['triggeredRecursive'] = true;
 109                         // Carry over cause information for logging
 110                         $extraParams['causeAction'] = $this->params['causeAction'];
 111                         $extraParams['causeAgent'] = $this->params['causeAgent'];
 112                         // Convert this into no more than $wgUpdateRowsPerJob RefreshLinks per-title
 113                         // jobs and possibly a recursive RefreshLinks job for the rest of the backlinks
 114                         $jobs = BacklinkJobUtils::partitionBacklinkJob(
 115                                 $this,
 116                                 $services->getMainConfig()->get( 'UpdateRowsPerJob' ),
 117                                 1, // job-per-title
 118                                 [ 'params' => $extraParams ]
 119                         );
 120                         JobQueueGroup::singleton()->push( $jobs );
 121                 // Job to update link tables for a set of titles
 122                 } elseif ( isset( $this->params['pages'] ) ) {
 123                         foreach ( $this->params['pages'] as list( $ns, $dbKey ) ) {
 124                                 $title = Title::makeTitleSafe( $ns, $dbKey );
 125                                 if ( $title ) {
 126                                         $ok = $this->runForTitle( $title ) && $ok;
 127                                 } else {
 128                                         $ok = false;
 129                                         $this->setLastError( "Invalid title ($ns,$dbKey)." );
 130                                 }
 131                         }
 132                 // Job to update link tables for a given title
 133                 } else {
 134                         $ok = $this->runForTitle( $this->title );
 135                 }
 136
 137                 return $ok;
 138         }
 139
 140         /**
 141          * @param Title $title
 142          * @return bool
 143          */
 144         protected function runForTitle( Title $title ) {
 145                 $services = MediaWikiServices::getInstance();
 146                 $stats = $services->getStatsdDataFactory();
 147                 $renderer = $services->getRevisionRenderer();
 148                 $parserCache = $services->getParserCache();
 149                 $lbFactory = $services->getDBLoadBalancerFactory();
 150                 $ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
 151
 152                 // Load the page from the master DB
 153                 $page = WikiPage::factory( $title );
 154                 $page->loadPageData( WikiPage::READ_LATEST );
 155
 156                 // Serialize link update job by page ID so they see each others' changes.
 157                 // The page ID and latest revision ID will be queried again after the lock
 158                 // is acquired to bail if they are changed from that of loadPageData() above.
 159                 // Serialize links updates by page ID so they see each others' changes
 160                 $dbw = $lbFactory->getMainLB()->getConnectionRef( DB_MASTER );
 161                 /** @noinspection PhpUnusedLocalVariableInspection */
 162                 $scopedLock = LinksUpdate::acquirePageLock( $dbw, $page->getId(), 'job' );
 163                 if ( $scopedLock === null ) {
 164                         // Another job is already updating the page, likely for a prior revision (T170596)
 165                         $this->setLastError( 'LinksUpdate already running for this page, try again later.' );
 166                         $stats->increment( 'refreshlinks.lock_failure' );
 167
 168                         return false;
 169                 }
 170
 171                 if ( $this->isAlreadyRefreshed( $page ) ) {
 172                         $stats->increment( 'refreshlinks.update_skipped' );
 173
 174                         return true;
 175                 }
 176
 177                 // Parse during a fresh transaction round for better read consistency
 178                 $lbFactory->beginMasterChanges( __METHOD__ );
 179                 $output = $this->getParserOutput( $renderer, $parserCache, $page, $stats );
 180                 $options = $this->getDataUpdateOptions();
 181                 $lbFactory->commitMasterChanges( __METHOD__ );
 182
 183                 if ( !$output ) {
 184                         return false; // raced out?
 185                 }
 186
 187                 // Tell DerivedPageDataUpdater to use this parser output
 188                 $options['known-revision-output'] = $output;
 189                 // Execute corresponding DataUpdates immediately
 190                 $page->doSecondaryDataUpdates( $options );
 191                 InfoAction::invalidateCache( $title );
 192
 193                 // Commit any writes here in case this method is called in a loop.
 194                 // In that case, the scoped lock will fail to be acquired.
 195                 $lbFactory->commitAndWaitForReplication( __METHOD__, $ticket );
 196
 197                 return true;
 198         }
 199
 200         /**
 201          * @param WikiPage $page
 202          * @return bool Whether something updated the backlinks with data newer than this job
 203          */
 204         private function isAlreadyRefreshed( WikiPage $page ) {
 205                 // Get the timestamp of the change that triggered this job
 206                 $rootTimestamp = $this->params['rootJobTimestamp'] ?? null;
 207                 if ( $rootTimestamp === null ) {
 208                         return false;
 209                 }
 210
 211                 if ( !empty( $this->params['isOpportunistic'] ) ) {
 212                         // Neither clock skew nor DB snapshot/replica DB lag matter much for
 213                         // such updates; focus on reusing the (often recently updated) cache
 214                         $lagAwareTimestamp = $rootTimestamp;
 215                 } else {
 216                         // For transclusion updates, the template changes must be reflected
 217                         $lagAwareTimestamp = wfTimestamp(
 218                                 TS_MW,
 219                                 wfTimestamp( TS_UNIX, $rootTimestamp ) + self::NORMAL_MAX_LAG
 220                         );
 221                 }
 222
 223                 return ( $page->getLinksTimestamp() > $lagAwareTimestamp );
 224         }
 225
 226         /**
 227          * Get the parser output if the page is unchanged from what was loaded in $page
 228          *
 229          * @param RevisionRenderer $renderer
 230          * @param ParserCache $parserCache
 231          * @param WikiPage $page Page already loaded with READ_LATEST
 232          * @param StatsdDataFactoryInterface $stats
 233          * @return ParserOutput|null Combined output for all slots; might only contain metadata
 234          */
 235         private function getParserOutput(
 236                 RevisionRenderer $renderer,
 237                 ParserCache $parserCache,
 238                 WikiPage $page,
 239                 StatsdDataFactoryInterface $stats
 240         ) {
 241                 $revision = $this->getCurrentRevisionIfUnchanged( $page, $stats );
 242                 if ( !$revision ) {
 243                         return null; // race condition?
 244                 }
 245
 246                 $cachedOutput = $this->getParserOutputFromCache( $parserCache, $page, $revision, $stats );
 247                 if ( $cachedOutput ) {
 248                         return $cachedOutput;
 249                 }
 250
 251                 $renderedRevision = $renderer->getRenderedRevision(
 252                         $revision,
 253                         $page->makeParserOptions( 'canonical' ),
 254                         null,
 255                         [ 'audience' => $revision::RAW ]
 256                 );
 257
 258                 $parseTimestamp = wfTimestampNow(); // timestamp that parsing started
 259                 $output = $renderedRevision->getRevisionParserOutput( [ 'generate-html' => false ] );
 260                 $output->setCacheTime( $parseTimestamp ); // notify LinksUpdate::doUpdate()
 261
 262                 return $output;
 263         }
 264
 265         /**
 266          * Get the current revision record if it is unchanged from what was loaded in $page
 267          *
 268          * @param WikiPage $page Page already loaded with READ_LATEST
 269          * @param StatsdDataFactoryInterface $stats
 270          * @return RevisionRecord|null The same instance that $page->getRevisionRecord() uses
 271          */
 272         private function getCurrentRevisionIfUnchanged(
 273                 WikiPage $page,
 274                 StatsdDataFactoryInterface $stats
 275         ) {
 276                 $title = $page->getTitle();
 277                 // Get the latest ID since acquirePageLock() in runForTitle() flushed the transaction.
 278                 // This is used to detect edits/moves after loadPageData() but before the scope lock.
 279                 // The works around the chicken/egg problem of determining the scope lock key name.
 280                 $latest = $title->getLatestRevID( Title::GAID_FOR_UPDATE );
 281
 282                 $triggeringRevisionId = $this->params['triggeringRevisionId'] ?? null;
 283                 if ( $triggeringRevisionId && $triggeringRevisionId !== $latest ) {
 284                         // This job is obsolete and one for the latest revision will handle updates
 285                         $stats->increment( 'refreshlinks.rev_not_current' );
 286                         $this->setLastError( "Revision $triggeringRevisionId is not current" );
 287
 288                         return null;
 289                 }
 290
 291                 // Load the current revision. Note that $page should have loaded with READ_LATEST.
 292                 // This instance will be reused in WikiPage::doSecondaryDataUpdates() later on.
 293                 $revision = $page->getRevisionRecord();
 294                 if ( !$revision ) {
 295                         $stats->increment( 'refreshlinks.rev_not_found' );
 296                         $this->setLastError( "Revision not found for {$title->getPrefixedDBkey()}" );
 297
 298                         return null; // just deleted?
 299                 } elseif ( $revision->getId() !== $latest || $revision->getPageId() !== $page->getId() ) {
 300                         // Do not clobber over newer updates with older ones. If all jobs where FIFO and
 301                         // serialized, it would be OK to update links based on older revisions since it
 302                         // would eventually get to the latest. Since that is not the case (by design),
 303                         // only update the link tables to a state matching the current revision's output.
 304                         $stats->increment( 'refreshlinks.rev_not_current' );
 305                         $this->setLastError( "Revision {$revision->getId()} is not current" );
 306
 307                         return null;
 308                 }
 309
 310                 return $revision;
 311         }
 312
 313         /**
 314          * Get the parser output from cache if it reflects the change that triggered this job
 315          *
 316          * @param ParserCache $parserCache
 317          * @param WikiPage $page
 318          * @param RevisionRecord $currentRevision
 319          * @param StatsdDataFactoryInterface $stats
 320          * @return ParserOutput|null
 321          */
 322         private function getParserOutputFromCache(
 323                 ParserCache $parserCache,
 324                 WikiPage $page,
 325                 RevisionRecord $currentRevision,
 326                 StatsdDataFactoryInterface $stats
 327         ) {
 328                 $cachedOutput = null;
 329                 // If page_touched changed after this root job, then it is likely that
 330                 // any views of the pages already resulted in re-parses which are now in
 331                 // cache. The cache can be reused to avoid expensive parsing in some cases.
 332                 $rootTimestamp = $this->params['rootJobTimestamp'] ?? null;
 333                 if ( $rootTimestamp !== null ) {
 334                         $opportunistic = !empty( $this->params['isOpportunistic'] );
 335                         if ( $opportunistic ) {
 336                                 // Neither clock skew nor DB snapshot/replica DB lag matter much for
 337                                 // such updates; focus on reusing the (often recently updated) cache
 338                                 $lagAwareTimestamp = $rootTimestamp;
 339                         } else {
 340                                 // For transclusion updates, the template changes must be reflected
 341                                 $lagAwareTimestamp = wfTimestamp(
 342                                         TS_MW,
 343                                         wfTimestamp( TS_UNIX, $rootTimestamp ) + self::NORMAL_MAX_LAG
 344                                 );
 345                         }
 346
 347                         if ( $page->getTouched() >= $rootTimestamp || $opportunistic ) {
 348                                 // Cache is suspected to be up-to-date so it's worth the I/O of checking.
 349                                 // As long as the cache rev ID matches the current rev ID and it reflects
 350                                 // the job's triggering change, then it is usable.
 351                                 $parserOptions = $page->makeParserOptions( 'canonical' );
 352                                 $output = $parserCache->getDirty( $page, $parserOptions );
 353                                 if (
 354                                         $output &&
 355                                         $output->getCacheRevisionId() == $currentRevision->getId() &&
 356                                         $output->getCacheTime() >= $lagAwareTimestamp
 357                                 ) {
 358                                         $cachedOutput = $output;
 359                                 }
 360                         }
 361                 }
 362
 363                 if ( $cachedOutput ) {
 364                         $stats->increment( 'refreshlinks.parser_cached' );
 365                 } else {
 366                         $stats->increment( 'refreshlinks.parser_uncached' );
 367                 }
 368
 369                 return $cachedOutput;
 370         }
 371
 372         /**
 373          * @return array
 374          */
 375         private function getDataUpdateOptions() {
 376                 $options = [
 377                         'recursive' => !empty( $this->params['useRecursiveLinksUpdate'] ),
 378                         // Carry over cause so the update can do extra logging
 379                         'causeAction' => $this->params['causeAction'],
 380                         'causeAgent' => $this->params['causeAgent']
 381                 ];
 382                 if ( !empty( $this->params['triggeringUser'] ) ) {
 383                         $userInfo = $this->params['triggeringUser'];
 384                         if ( $userInfo['userId'] ) {
 385                                 $options['triggeringUser'] = User::newFromId( $userInfo['userId'] );
 386                         } else {
 387                                 // Anonymous, use the username
 388                                 $options['triggeringUser'] = User::newFromName( $userInfo['userName'], false );
 389                         }
 390                 }
 391
 392                 return $options;
 393         }
 394
 395         public function getDeduplicationInfo() {
 396                 $info = parent::getDeduplicationInfo();
 397                 unset( $info['causeAction'] );
 398                 unset( $info['causeAgent'] );
 399                 if ( is_array( $info['params'] ) ) {
 400                         // For per-pages jobs, the job title is that of the template that changed
 401                         // (or similar), so remove that since it ruins duplicate detection
 402                         if ( isset( $info['params']['pages'] ) ) {
 403                                 unset( $info['namespace'] );
 404                                 unset( $info['title'] );
 405                         }
 406                 }
 407
 408                 return $info;
 409         }
 410
 411         public function workItemCount() {
 412                 if ( !empty( $this->params['recursive'] ) ) {
 413                         return 0; // nothing actually refreshed
 414                 } elseif ( isset( $this->params['pages'] ) ) {
 415                         return count( $this->params['pages'] );
 416                 }
 417
 418                 return 1; // one title
 419         }
 420 }