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
  26 /**
  27  * Job to update link tables for pages
  28  *
  29  * This job comes in a few variants:
  30  *   - a) Recursive jobs to update links for backlink pages for a given title.
  31  *        These jobs have (recursive:true,table:<table>) set.
  32  *   - b) Jobs to update links for a set of pages (the job title is ignored).
  33  *        These jobs have (pages:(<page ID>:(<namespace>,<title>),...) set.
  34  *   - c) Jobs to update links for a single page (the job title)
  35  *        These jobs need no extra fields set.
  36  *
  37  * @ingroup JobQueue
  38  */
  39 class RefreshLinksJob extends Job {
  40         /** @var float Cache parser output when it takes this long to render */
  41         const PARSE_THRESHOLD_SEC = 1.0;
  42         /** @var int Lag safety margin when comparing root job times to last-refresh times */
  43         const CLOCK_FUDGE = 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         }
  58
  59         /**
  60          * @param Title $title
  61          * @param array $params
  62          * @return RefreshLinksJob
  63          */
  64         public static function newPrioritized( Title $title, array $params ) {
  65                 $job = new self( $title, $params );
  66                 $job->command = 'refreshLinksPrioritized';
  67
  68                 return $job;
  69         }
  70
  71         /**
  72          * @param Title $title
  73          * @param array $params
  74          * @return RefreshLinksJob
  75          */
  76         public static function newDynamic( Title $title, array $params ) {
  77                 $job = new self( $title, $params );
  78                 $job->command = 'refreshLinksDynamic';
  79
  80                 return $job;
  81         }
  82
  83         function run() {
  84                 global $wgUpdateRowsPerJob;
  85
  86                 $ok = true;
  87                 // Job to update all (or a range of) backlink pages for a page
  88                 if ( !empty( $this->params['recursive'] ) ) {
  89                         // When the base job branches, wait for the replica DBs to catch up to the master.
  90                         // From then on, we know that any template changes at the time the base job was
  91                         // enqueued will be reflected in backlink page parses when the leaf jobs run.
  92                         if ( !isset( $this->params['range'] ) ) {
  93                                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
  94                                 if ( !$lbFactory->waitForReplication( [
  95                                                 'domain'  => $lbFactory->getLocalDomainID(),
  96                                                 'timeout' => self::LAG_WAIT_TIMEOUT
  97                                 ] ) ) { // only try so hard
  98                                         $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
  99                                         $stats->increment( 'refreshlinks.lag_wait_failed' );
 100                                 }
 101                         }
 102                         // Carry over information for de-duplication
 103                         $extraParams = $this->getRootJobParams();
 104                         $extraParams['triggeredRecursive'] = true;
 105                         // Carry over cause information for logging
 106                         $extraParams['causeAction'] = $this->params['causeAction'];
 107                         $extraParams['causeAgent'] = $this->params['causeAgent'];
 108                         // Convert this into no more than $wgUpdateRowsPerJob RefreshLinks per-title
 109                         // jobs and possibly a recursive RefreshLinks job for the rest of the backlinks
 110                         $jobs = BacklinkJobUtils::partitionBacklinkJob(
 111                                 $this,
 112                                 $wgUpdateRowsPerJob,
 113                                 1, // job-per-title
 114                                 [ 'params' => $extraParams ]
 115                         );
 116                         JobQueueGroup::singleton()->push( $jobs );
 117                 // Job to update link tables for a set of titles
 118                 } elseif ( isset( $this->params['pages'] ) ) {
 119                         foreach ( $this->params['pages'] as list( $ns, $dbKey ) ) {
 120                                 $title = Title::makeTitleSafe( $ns, $dbKey );
 121                                 if ( $title ) {
 122                                         $this->runForTitle( $title );
 123                                 } else {
 124                                         $ok = false;
 125                                         $this->setLastError( "Invalid title ($ns,$dbKey)." );
 126                                 }
 127                         }
 128                 // Job to update link tables for a given title
 129                 } else {
 130                         $this->runForTitle( $this->title );
 131                 }
 132
 133                 return $ok;
 134         }
 135
 136         /**
 137          * @param Title $title
 138          * @return bool
 139          */
 140         protected function runForTitle( Title $title ) {
 141                 $services = MediaWikiServices::getInstance();
 142                 $stats = $services->getStatsdDataFactory();
 143                 $lbFactory = $services->getDBLoadBalancerFactory();
 144                 $revisionStore = $services->getRevisionStore();
 145                 $renderer = $services->getRevisionRenderer();
 146                 $ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
 147
 148                 $page = WikiPage::factory( $title );
 149                 $page->loadPageData( WikiPage::READ_LATEST );
 150
 151                 // Serialize links updates by page ID so they see each others' changes
 152                 $dbw = $lbFactory->getMainLB()->getConnection( DB_MASTER );
 153                 /** @noinspection PhpUnusedLocalVariableInspection */
 154                 $scopedLock = LinksUpdate::acquirePageLock( $dbw, $page->getId(), 'job' );
 155                 if ( $scopedLock === null ) {
 156                         // Another job is already updating the page, likely for an older revision (T170596).
 157                         $this->setLastError( 'LinksUpdate already running for this page, try again later.' );
 158                         return false;
 159                 }
 160                 // Get the latest ID *after* acquirePageLock() flushed the transaction.
 161                 // This is used to detect edits/moves after loadPageData() but before the scope lock.
 162                 // The works around the chicken/egg problem of determining the scope lock key.
 163                 $latest = $title->getLatestRevID( Title::GAID_FOR_UPDATE );
 164
 165                 if ( !empty( $this->params['triggeringRevisionId'] ) ) {
 166                         // Fetch the specified revision; lockAndGetLatest() below detects if the page
 167                         // was edited since and aborts in order to avoid corrupting the link tables
 168                         $revision = $revisionStore->getRevisionById(
 169                                 (int)$this->params['triggeringRevisionId'],
 170                                 Revision::READ_LATEST
 171                         );
 172                 } else {
 173                         // Fetch current revision; READ_LATEST reduces lockAndGetLatest() check failures
 174                         $revision = $revisionStore->getRevisionByTitle( $title, 0, Revision::READ_LATEST );
 175                 }
 176
 177                 if ( !$revision ) {
 178                         $stats->increment( 'refreshlinks.rev_not_found' );
 179                         $this->setLastError( "Revision not found for {$title->getPrefixedDBkey()}" );
 180                         return false; // just deleted?
 181                 } elseif ( $revision->getId() != $latest || $revision->getPageId() !== $page->getId() ) {
 182                         // Do not clobber over newer updates with older ones. If all jobs where FIFO and
 183                         // serialized, it would be OK to update links based on older revisions since it
 184                         // would eventually get to the latest. Since that is not the case (by design),
 185                         // only update the link tables to a state matching the current revision's output.
 186                         $stats->increment( 'refreshlinks.rev_not_current' );
 187                         $this->setLastError( "Revision {$revision->getId()} is not current" );
 188                         return false;
 189                 }
 190
 191                 $parserOutput = false;
 192                 $parserOptions = $page->makeParserOptions( 'canonical' );
 193                 // If page_touched changed after this root job, then it is likely that
 194                 // any views of the pages already resulted in re-parses which are now in
 195                 // cache. The cache can be reused to avoid expensive parsing in some cases.
 196                 if ( isset( $this->params['rootJobTimestamp'] ) ) {
 197                         $opportunistic = !empty( $this->params['isOpportunistic'] );
 198
 199                         $skewedTimestamp = $this->params['rootJobTimestamp'];
 200                         if ( $opportunistic ) {
 201                                 // Neither clock skew nor DB snapshot/replica DB lag matter much for such
 202                                 // updates; focus on reusing the (often recently updated) cache
 203                         } else {
 204                                 // For transclusion updates, the template changes must be reflected
 205                                 $skewedTimestamp = wfTimestamp( TS_MW,
 206                                         wfTimestamp( TS_UNIX, $skewedTimestamp ) + self::CLOCK_FUDGE
 207                                 );
 208                         }
 209
 210                         if ( $page->getLinksTimestamp() > $skewedTimestamp ) {
 211                                 // Something already updated the backlinks since this job was made
 212                                 $stats->increment( 'refreshlinks.update_skipped' );
 213                                 return true;
 214                         }
 215
 216                         if ( $page->getTouched() >= $this->params['rootJobTimestamp'] || $opportunistic ) {
 217                                 // Cache is suspected to be up-to-date. As long as the cache rev ID matches
 218                                 // and it reflects the job's triggering change, then it is usable.
 219                                 $parserOutput = $services->getParserCache()->getDirty( $page, $parserOptions );
 220                                 if ( !$parserOutput
 221                                         || $parserOutput->getCacheRevisionId() != $revision->getId()
 222                                         || $parserOutput->getCacheTime() < $skewedTimestamp
 223                                 ) {
 224                                         $parserOutput = false; // too stale
 225                                 }
 226                         }
 227                 }
 228
 229                 // Fetch the current revision and parse it if necessary...
 230                 if ( $parserOutput ) {
 231                         $stats->increment( 'refreshlinks.parser_cached' );
 232                 } else {
 233                         $start = microtime( true );
 234
 235                         $checkCache = $page->shouldCheckParserCache( $parserOptions, $revision->getId() );
 236
 237                         // Revision ID must be passed to the parser output to get revision variables correct
 238                         $renderedRevision = $renderer->getRenderedRevision(
 239                                 $revision,
 240                                 $parserOptions,
 241                                 null,
 242                                 [
 243                                         // use master, for consistency with the getRevisionByTitle call above.
 244                                         'use-master' => true,
 245                                         // bypass audience checks, since we know that this is the current revision.
 246                                         'audience' => RevisionRecord::RAW
 247                                 ]
 248                         );
 249                         $parserOutput = $renderedRevision->getRevisionParserOutput(
 250                                 // HTML is only needed if the output is to be placed in the parser cache
 251                                 [ 'generate-html' => $checkCache ]
 252                         );
 253
 254                         // If it took a long time to render, then save this back to the cache to avoid
 255                         // wasted CPU by other apaches or job runners. We don't want to always save to
 256                         // cache as this can cause high cache I/O and LRU churn when a template changes.
 257                         $elapsed = microtime( true ) - $start;
 258
 259                         $parseThreshold = $this->params['parseThreshold'] ?? self::PARSE_THRESHOLD_SEC;
 260
 261                         if ( $checkCache && $elapsed >= $parseThreshold && $parserOutput->isCacheable() ) {
 262                                 $ctime = wfTimestamp( TS_MW, (int)$start ); // cache time
 263                                 $services->getParserCache()->save(
 264                                         $parserOutput, $page, $parserOptions, $ctime, $revision->getId()
 265                                 );
 266                         }
 267                         $stats->increment( 'refreshlinks.parser_uncached' );
 268                 }
 269
 270                 $options = [
 271                         'recursive' => !empty( $this->params['useRecursiveLinksUpdate'] ),
 272                         // Carry over cause so the update can do extra logging
 273                         'causeAction' => $this->params['causeAction'],
 274                         'causeAgent' => $this->params['causeAgent'],
 275                         'defer' => false,
 276                         'transactionTicket' => $ticket,
 277                 ];
 278                 if ( !empty( $this->params['triggeringUser'] ) ) {
 279                         $userInfo = $this->params['triggeringUser'];
 280                         if ( $userInfo['userId'] ) {
 281                                 $options['triggeringUser'] = User::newFromId( $userInfo['userId'] );
 282                         } else {
 283                                 // Anonymous, use the username
 284                                 $options['triggeringUser'] = User::newFromName( $userInfo['userName'], false );
 285                         }
 286                 }
 287                 $page->doSecondaryDataUpdates( $options );
 288
 289                 InfoAction::invalidateCache( $title );
 290
 291                 // Commit any writes here in case this method is called in a loop.
 292                 // In that case, the scoped lock will fail to be acquired.
 293                 $lbFactory->commitAndWaitForReplication( __METHOD__, $ticket );
 294
 295                 return true;
 296         }
 297
 298         public function getDeduplicationInfo() {
 299                 $info = parent::getDeduplicationInfo();
 300                 unset( $info['causeAction'] );
 301                 unset( $info['causeAgent'] );
 302                 if ( is_array( $info['params'] ) ) {
 303                         // For per-pages jobs, the job title is that of the template that changed
 304                         // (or similar), so remove that since it ruins duplicate detection
 305                         if ( isset( $info['params']['pages'] ) ) {
 306                                 unset( $info['namespace'] );
 307                                 unset( $info['title'] );
 308                         }
 309                 }
 310
 311                 return $info;
 312         }
 313
 314         public function workItemCount() {
 315                 if ( !empty( $this->params['recursive'] ) ) {
 316                         return 0; // nothing actually refreshed
 317                 } elseif ( isset( $this->params['pages'] ) ) {
 318                         return count( $this->params['pages'] );
 319                 }
 320
 321                 return 1; // one title
 322         }
 323 }