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 Wikimedia\Rdbms\DBReplicationWaitError;
  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                 // Job to update all (or a range of) backlink pages for a page
  87                 if ( !empty( $this->params['recursive'] ) ) {
  88                         // When the base job branches, wait for the replica DBs to catch up to the master.
  89                         // From then on, we know that any template changes at the time the base job was
  90                         // enqueued will be reflected in backlink page parses when the leaf jobs run.
  91                         if ( !isset( $this->params['range'] ) ) {
  92                                 try {
  93                                         $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
  94                                         $lbFactory->waitForReplication( [
  95                                                 'wiki'    => wfWikiID(),
  96                                                 'timeout' => self::LAG_WAIT_TIMEOUT
  97                                         ] );
  98                                 } catch ( DBReplicationWaitError $e ) { // only try so hard
  99                                         $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
 100                                         $stats->increment( 'refreshlinks.lag_wait_failed' );
 101                                 }
 102                         }
 103                         // Carry over information for de-duplication
 104                         $extraParams = $this->getRootJobParams();
 105                         $extraParams['triggeredRecursive'] = true;
 106                         // Carry over cause information for logging
 107                         $extraParams['causeAction'] = $this->params['causeAction'];
 108                         $extraParams['causeAgent'] = $this->params['causeAgent'];
 109                         // Convert this into no more than $wgUpdateRowsPerJob RefreshLinks per-title
 110                         // jobs and possibly a recursive RefreshLinks job for the rest of the backlinks
 111                         $jobs = BacklinkJobUtils::partitionBacklinkJob(
 112                                 $this,
 113                                 $wgUpdateRowsPerJob,
 114                                 1, // job-per-title
 115                                 [ 'params' => $extraParams ]
 116                         );
 117                         JobQueueGroup::singleton()->push( $jobs );
 118                 // Job to update link tables for a set of titles
 119                 } elseif ( isset( $this->params['pages'] ) ) {
 120                         foreach ( $this->params['pages'] as $nsAndKey ) {
 121                                 list( $ns, $dbKey ) = $nsAndKey;
 122                                 $this->runForTitle( Title::makeTitleSafe( $ns, $dbKey ) );
 123                         }
 124                 // Job to update link tables for a given title
 125                 } else {
 126                         $this->runForTitle( $this->title );
 127                 }
 128
 129                 return true;
 130         }
 131
 132         /**
 133          * @param Title $title
 134          * @return bool
 135          */
 136         protected function runForTitle( Title $title ) {
 137                 $services = MediaWikiServices::getInstance();
 138                 $stats = $services->getStatsdDataFactory();
 139                 $lbFactory = $services->getDBLoadBalancerFactory();
 140                 $ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
 141
 142                 $page = WikiPage::factory( $title );
 143                 $page->loadPageData( WikiPage::READ_LATEST );
 144
 145                 // Serialize links updates by page ID so they see each others' changes
 146                 $dbw = $lbFactory->getMainLB()->getConnection( DB_MASTER );
 147                 /** @noinspection PhpUnusedLocalVariableInspection */
 148                 $scopedLock = LinksUpdate::acquirePageLock( $dbw, $page->getId(), 'job' );
 149                 if ( $scopedLock === null ) {
 150                         // Another job is already updating the page, likely for an older revision (T170596).
 151                         $this->setLastError( 'LinksUpdate already running for this page, try again later.' );
 152                         return false;
 153                 }
 154                 // Get the latest ID *after* acquirePageLock() flushed the transaction.
 155                 // This is used to detect edits/moves after loadPageData() but before the scope lock.
 156                 // The works around the chicken/egg problem of determining the scope lock key.
 157                 $latest = $title->getLatestRevID( Title::GAID_FOR_UPDATE );
 158
 159                 if ( !empty( $this->params['triggeringRevisionId'] ) ) {
 160                         // Fetch the specified revision; lockAndGetLatest() below detects if the page
 161                         // was edited since and aborts in order to avoid corrupting the link tables
 162                         $revision = Revision::newFromId(
 163                                 $this->params['triggeringRevisionId'],
 164                                 Revision::READ_LATEST
 165                         );
 166                 } else {
 167                         // Fetch current revision; READ_LATEST reduces lockAndGetLatest() check failures
 168                         $revision = Revision::newFromTitle( $title, false, Revision::READ_LATEST );
 169                 }
 170
 171                 if ( !$revision ) {
 172                         $stats->increment( 'refreshlinks.rev_not_found' );
 173                         $this->setLastError( "Revision not found for {$title->getPrefixedDBkey()}" );
 174                         return false; // just deleted?
 175                 } elseif ( $revision->getId() != $latest || $revision->getPage() !== $page->getId() ) {
 176                         // Do not clobber over newer updates with older ones. If all jobs where FIFO and
 177                         // serialized, it would be OK to update links based on older revisions since it
 178                         // would eventually get to the latest. Since that is not the case (by design),
 179                         // only update the link tables to a state matching the current revision's output.
 180                         $stats->increment( 'refreshlinks.rev_not_current' );
 181                         $this->setLastError( "Revision {$revision->getId()} is not current" );
 182                         return false;
 183                 }
 184
 185                 $content = $revision->getContent( Revision::RAW );
 186                 if ( !$content ) {
 187                         // If there is no content, pretend the content is empty
 188                         $content = $revision->getContentHandler()->makeEmptyContent();
 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                         // Revision ID must be passed to the parser output to get revision variables correct
 235                         $parserOutput = $content->getParserOutput(
 236                                 $title, $revision->getId(), $parserOptions, false );
 237                         $elapsed = microtime( true ) - $start;
 238                         // If it took a long time to render, then save this back to the cache to avoid
 239                         // wasted CPU by other apaches or job runners. We don't want to always save to
 240                         // cache as this can cause high cache I/O and LRU churn when a template changes.
 241                         if ( $elapsed >= self::PARSE_THRESHOLD_SEC
 242                                 && $page->shouldCheckParserCache( $parserOptions, $revision->getId() )
 243                                 && $parserOutput->isCacheable()
 244                         ) {
 245                                 $ctime = wfTimestamp( TS_MW, (int)$start ); // cache time
 246                                 $services->getParserCache()->save(
 247                                         $parserOutput, $page, $parserOptions, $ctime, $revision->getId()
 248                                 );
 249                         }
 250                         $stats->increment( 'refreshlinks.parser_uncached' );
 251                 }
 252
 253                 $updates = $content->getSecondaryDataUpdates(
 254                         $title,
 255                         null,
 256                         !empty( $this->params['useRecursiveLinksUpdate'] ),
 257                         $parserOutput
 258                 );
 259
 260                 // For legacy hook handlers doing updates via LinksUpdateConstructed, make sure
 261                 // any pending writes they made get flushed before the doUpdate() calls below.
 262                 // This avoids snapshot-clearing errors in LinksUpdate::acquirePageLock().
 263                 $lbFactory->commitAndWaitForReplication( __METHOD__, $ticket );
 264
 265                 foreach ( $updates as $update ) {
 266                         // Carry over cause in case so the update can do extra logging
 267                         $update->setCause( $this->params['causeAction'], $this->params['causeAgent'] );
 268                         // FIXME: This code probably shouldn't be here?
 269                         // Needed by things like Echo notifications which need
 270                         // to know which user caused the links update
 271                         if ( $update instanceof LinksUpdate ) {
 272                                 $update->setRevision( $revision );
 273                                 if ( !empty( $this->params['triggeringUser'] ) ) {
 274                                         $userInfo = $this->params['triggeringUser'];
 275                                         if ( $userInfo['userId'] ) {
 276                                                 $user = User::newFromId( $userInfo['userId'] );
 277                                         } else {
 278                                                 // Anonymous, use the username
 279                                                 $user = User::newFromName( $userInfo['userName'], false );
 280                                         }
 281                                         $update->setTriggeringUser( $user );
 282                                 }
 283                         }
 284                 }
 285
 286                 foreach ( $updates as $update ) {
 287                         $update->setTransactionTicket( $ticket );
 288                         $update->doUpdate();
 289                 }
 290
 291                 InfoAction::invalidateCache( $title );
 292
 293                 // Commit any writes here in case this method is called in a loop.
 294                 // In that case, the scoped lock will fail to be acquired.
 295                 $lbFactory->commitAndWaitForReplication( __METHOD__, $ticket );
 296
 297                 return true;
 298         }
 299
 300         public function getDeduplicationInfo() {
 301                 $info = parent::getDeduplicationInfo();
 302                 unset( $info['causeAction'] );
 303                 unset( $info['causeAgent'] );
 304                 if ( is_array( $info['params'] ) ) {
 305                         // For per-pages jobs, the job title is that of the template that changed
 306                         // (or similar), so remove that since it ruins duplicate detection
 307                         if ( isset( $info['params']['pages'] ) ) {
 308                                 unset( $info['namespace'] );
 309                                 unset( $info['title'] );
 310                         }
 311                 }
 312
 313                 return $info;
 314         }
 315
 316         public function workItemCount() {
 317                 if ( !empty( $this->params['recursive'] ) ) {
 318                         return 0; // nothing actually refreshed
 319                 } elseif ( isset( $this->params['pages'] ) ) {
 320                         return count( $this->params['pages'] );
 321                 }
 322
 323                 return 1; // one title
 324         }
 325 }