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
  25 /**
  26  * Job to update link tables for pages
  27  *
  28  * This job comes in a few variants:
  29  *   - a) Recursive jobs to update links for backlink pages for a given title.
  30  *        These jobs have (recursive:true,table:<table>) set.
  31  *   - b) Jobs to update links for a set of pages (the job title is ignored).
  32  *        These jobs have (pages:(<page ID>:(<namespace>,<title>),...) set.
  33  *   - c) Jobs to update links for a single page (the job title)
  34  *        These jobs need no extra fields set.
  35  *
  36  * @ingroup JobQueue
  37  */
  38 class RefreshLinksJob extends Job {
  39         /** @var float Cache parser output when it takes this long to render */
  40         const PARSE_THRESHOLD_SEC = 1.0;
  41         /** @var int Lag safety margin when comparing root job times to last-refresh times */
  42         const CLOCK_FUDGE = 10;
  43         /** @var int How many seconds to wait for replica DBs to catch up */
  44         const LAG_WAIT_TIMEOUT = 15;
  45
  46         function __construct( Title $title, array $params ) {
  47                 parent::__construct( 'refreshLinks', $title, $params );
  48                 // Avoid the overhead of de-duplication when it would be pointless
  49                 $this->removeDuplicates = (
  50                         // Ranges rarely will line up
  51                         !isset( $params['range'] ) &&
  52                         // Multiple pages per job make matches unlikely
  53                         !( isset( $params['pages'] ) && count( $params['pages'] ) != 1 )
  54                 );
  55                 $this->params += [ 'causeAction' => 'unknown', 'causeAgent' => 'unknown' ];
  56         }
  57
  58         /**
  59          * @param Title $title
  60          * @param array $params
  61          * @return RefreshLinksJob
  62          */
  63         public static function newPrioritized( Title $title, array $params ) {
  64                 $job = new self( $title, $params );
  65                 $job->command = 'refreshLinksPrioritized';
  66
  67                 return $job;
  68         }
  69
  70         /**
  71          * @param Title $title
  72          * @param array $params
  73          * @return RefreshLinksJob
  74          */
  75         public static function newDynamic( Title $title, array $params ) {
  76                 $job = new self( $title, $params );
  77                 $job->command = 'refreshLinksDynamic';
  78
  79                 return $job;
  80         }
  81
  82         function run() {
  83                 global $wgUpdateRowsPerJob;
  84
  85                 // Job to update all (or a range of) backlink pages for a page
  86                 if ( !empty( $this->params['recursive'] ) ) {
  87                         // When the base job branches, wait for the replica DBs to catch up to the master.
  88                         // From then on, we know that any template changes at the time the base job was
  89                         // enqueued will be reflected in backlink page parses when the leaf jobs run.
  90                         if ( !isset( $this->params['range'] ) ) {
  91                                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
  92                                 if ( !$lbFactory->waitForReplication( [
  93                                                 'wiki'    => wfWikiID(),
  94                                                 'timeout' => self::LAG_WAIT_TIMEOUT
  95                                 ] ) ) { // only try so hard
  96                                         $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
  97                                         $stats->increment( 'refreshlinks.lag_wait_failed' );
  98                                 }
  99                         }
 100                         // Carry over information for de-duplication
 101                         $extraParams = $this->getRootJobParams();
 102                         $extraParams['triggeredRecursive'] = true;
 103                         // Carry over cause information for logging
 104                         $extraParams['causeAction'] = $this->params['causeAction'];
 105                         $extraParams['causeAgent'] = $this->params['causeAgent'];
 106                         // Convert this into no more than $wgUpdateRowsPerJob RefreshLinks per-title
 107                         // jobs and possibly a recursive RefreshLinks job for the rest of the backlinks
 108                         $jobs = BacklinkJobUtils::partitionBacklinkJob(
 109                                 $this,
 110                                 $wgUpdateRowsPerJob,
 111                                 1, // job-per-title
 112                                 [ 'params' => $extraParams ]
 113                         );
 114                         JobQueueGroup::singleton()->push( $jobs );
 115                 // Job to update link tables for a set of titles
 116                 } elseif ( isset( $this->params['pages'] ) ) {
 117                         foreach ( $this->params['pages'] as $nsAndKey ) {
 118                                 list( $ns, $dbKey ) = $nsAndKey;
 119                                 $this->runForTitle( Title::makeTitleSafe( $ns, $dbKey ) );
 120                         }
 121                 // Job to update link tables for a given title
 122                 } else {
 123                         $this->runForTitle( $this->title );
 124                 }
 125
 126                 return true;
 127         }
 128
 129         /**
 130          * @param Title $title
 131          * @return bool
 132          */
 133         protected function runForTitle( Title $title ) {
 134                 $services = MediaWikiServices::getInstance();
 135                 $stats = $services->getStatsdDataFactory();
 136                 $lbFactory = $services->getDBLoadBalancerFactory();
 137                 $ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
 138
 139                 $page = WikiPage::factory( $title );
 140                 $page->loadPageData( WikiPage::READ_LATEST );
 141
 142                 // Serialize links updates by page ID so they see each others' changes
 143                 $dbw = $lbFactory->getMainLB()->getConnection( DB_MASTER );
 144                 /** @noinspection PhpUnusedLocalVariableInspection */
 145                 $scopedLock = LinksUpdate::acquirePageLock( $dbw, $page->getId(), 'job' );
 146                 if ( $scopedLock === null ) {
 147                         // Another job is already updating the page, likely for an older revision (T170596).
 148                         $this->setLastError( 'LinksUpdate already running for this page, try again later.' );
 149                         return false;
 150                 }
 151                 // Get the latest ID *after* acquirePageLock() flushed the transaction.
 152                 // This is used to detect edits/moves after loadPageData() but before the scope lock.
 153                 // The works around the chicken/egg problem of determining the scope lock key.
 154                 $latest = $title->getLatestRevID( Title::GAID_FOR_UPDATE );
 155
 156                 if ( !empty( $this->params['triggeringRevisionId'] ) ) {
 157                         // Fetch the specified revision; lockAndGetLatest() below detects if the page
 158                         // was edited since and aborts in order to avoid corrupting the link tables
 159                         $revision = Revision::newFromId(
 160                                 $this->params['triggeringRevisionId'],
 161                                 Revision::READ_LATEST
 162                         );
 163                 } else {
 164                         // Fetch current revision; READ_LATEST reduces lockAndGetLatest() check failures
 165                         $revision = Revision::newFromTitle( $title, false, Revision::READ_LATEST );
 166                 }
 167
 168                 if ( !$revision ) {
 169                         $stats->increment( 'refreshlinks.rev_not_found' );
 170                         $this->setLastError( "Revision not found for {$title->getPrefixedDBkey()}" );
 171                         return false; // just deleted?
 172                 } elseif ( $revision->getId() != $latest || $revision->getPage() !== $page->getId() ) {
 173                         // Do not clobber over newer updates with older ones. If all jobs where FIFO and
 174                         // serialized, it would be OK to update links based on older revisions since it
 175                         // would eventually get to the latest. Since that is not the case (by design),
 176                         // only update the link tables to a state matching the current revision's output.
 177                         $stats->increment( 'refreshlinks.rev_not_current' );
 178                         $this->setLastError( "Revision {$revision->getId()} is not current" );
 179                         return false;
 180                 }
 181
 182                 $content = $revision->getContent( Revision::RAW );
 183                 if ( !$content ) {
 184                         // If there is no content, pretend the content is empty
 185                         $content = $revision->getContentHandler()->makeEmptyContent();
 186                 }
 187
 188                 $parserOutput = false;
 189                 $parserOptions = $page->makeParserOptions( 'canonical' );
 190                 // If page_touched changed after this root job, then it is likely that
 191                 // any views of the pages already resulted in re-parses which are now in
 192                 // cache. The cache can be reused to avoid expensive parsing in some cases.
 193                 if ( isset( $this->params['rootJobTimestamp'] ) ) {
 194                         $opportunistic = !empty( $this->params['isOpportunistic'] );
 195
 196                         $skewedTimestamp = $this->params['rootJobTimestamp'];
 197                         if ( $opportunistic ) {
 198                                 // Neither clock skew nor DB snapshot/replica DB lag matter much for such
 199                                 // updates; focus on reusing the (often recently updated) cache
 200                         } else {
 201                                 // For transclusion updates, the template changes must be reflected
 202                                 $skewedTimestamp = wfTimestamp( TS_MW,
 203                                         wfTimestamp( TS_UNIX, $skewedTimestamp ) + self::CLOCK_FUDGE
 204                                 );
 205                         }
 206
 207                         if ( $page->getLinksTimestamp() > $skewedTimestamp ) {
 208                                 // Something already updated the backlinks since this job was made
 209                                 $stats->increment( 'refreshlinks.update_skipped' );
 210                                 return true;
 211                         }
 212
 213                         if ( $page->getTouched() >= $this->params['rootJobTimestamp'] || $opportunistic ) {
 214                                 // Cache is suspected to be up-to-date. As long as the cache rev ID matches
 215                                 // and it reflects the job's triggering change, then it is usable.
 216                                 $parserOutput = $services->getParserCache()->getDirty( $page, $parserOptions );
 217                                 if ( !$parserOutput
 218                                         || $parserOutput->getCacheRevisionId() != $revision->getId()
 219                                         || $parserOutput->getCacheTime() < $skewedTimestamp
 220                                 ) {
 221                                         $parserOutput = false; // too stale
 222                                 }
 223                         }
 224                 }
 225
 226                 // Fetch the current revision and parse it if necessary...
 227                 if ( $parserOutput ) {
 228                         $stats->increment( 'refreshlinks.parser_cached' );
 229                 } else {
 230                         $start = microtime( true );
 231                         // Revision ID must be passed to the parser output to get revision variables correct
 232                         $parserOutput = $content->getParserOutput(
 233                                 $title, $revision->getId(), $parserOptions, false );
 234                         $elapsed = microtime( true ) - $start;
 235                         // If it took a long time to render, then save this back to the cache to avoid
 236                         // wasted CPU by other apaches or job runners. We don't want to always save to
 237                         // cache as this can cause high cache I/O and LRU churn when a template changes.
 238                         if ( $elapsed >= self::PARSE_THRESHOLD_SEC
 239                                 && $page->shouldCheckParserCache( $parserOptions, $revision->getId() )
 240                                 && $parserOutput->isCacheable()
 241                         ) {
 242                                 $ctime = wfTimestamp( TS_MW, (int)$start ); // cache time
 243                                 $services->getParserCache()->save(
 244                                         $parserOutput, $page, $parserOptions, $ctime, $revision->getId()
 245                                 );
 246                         }
 247                         $stats->increment( 'refreshlinks.parser_uncached' );
 248                 }
 249
 250                 $options = [
 251                         'recursive' => !empty( $this->params['useRecursiveLinksUpdate'] ),
 252                         // Carry over cause so the update can do extra logging
 253                         'causeAction' => $this->params['causeAction'],
 254                         'causeAgent' => $this->params['causeAgent'],
 255                         'defer' => false,
 256                         'transactionTicket' => $ticket,
 257                 ];
 258                 if ( !empty( $this->params['triggeringUser'] ) ) {
 259                         $userInfo = $this->params['triggeringUser'];
 260                         if ( $userInfo['userId'] ) {
 261                                 $options['triggeringUser'] = User::newFromId( $userInfo['userId'] );
 262                         } else {
 263                                 // Anonymous, use the username
 264                                 $options['triggeringUser'] = User::newFromName( $userInfo['userName'], false );
 265                         }
 266                 }
 267                 $page->doSecondaryDataUpdates( $options );
 268
 269                 InfoAction::invalidateCache( $title );
 270
 271                 // Commit any writes here in case this method is called in a loop.
 272                 // In that case, the scoped lock will fail to be acquired.
 273                 $lbFactory->commitAndWaitForReplication( __METHOD__, $ticket );
 274
 275                 return true;
 276         }
 277
 278         public function getDeduplicationInfo() {
 279                 $info = parent::getDeduplicationInfo();
 280                 unset( $info['causeAction'] );
 281                 unset( $info['causeAgent'] );
 282                 if ( is_array( $info['params'] ) ) {
 283                         // For per-pages jobs, the job title is that of the template that changed
 284                         // (or similar), so remove that since it ruins duplicate detection
 285                         if ( isset( $info['params']['pages'] ) ) {
 286                                 unset( $info['namespace'] );
 287                                 unset( $info['title'] );
 288                         }
 289                 }
 290
 291                 return $info;
 292         }
 293
 294         public function workItemCount() {
 295                 if ( !empty( $this->params['recursive'] ) ) {
 296                         return 0; // nothing actually refreshed
 297                 } elseif ( isset( $this->params['pages'] ) ) {
 298                         return count( $this->params['pages'] );
 299                 }
 300
 301                 return 1; // one title
 302         }
 303 }