includes/deferred/DeferredUpdates.php

   1 <?php
   2 /**
   3  * Interface and manager for deferred updates.
   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  */
  22 use Wikimedia\Rdbms\IDatabase;
  23 use MediaWiki\MediaWikiServices;
  24 use Wikimedia\Rdbms\LBFactory;
  25 use Wikimedia\Rdbms\LoadBalancer;
  26
  27 /**
  28  * Class for managing the deferred updates
  29  *
  30  * In web request mode, deferred updates can be run at the end of the request, either before or
  31  * after the HTTP response has been sent. In either case, they run after the DB commit step. If
  32  * an update runs after the response is sent, it will not block clients. If sent before, it will
  33  * run synchronously. These two modes are defined via PRESEND and POSTSEND constants, the latter
  34  * being the default for addUpdate() and addCallableUpdate().
  35  *
  36  * Updates that work through this system will be more likely to complete by the time the client
  37  * makes their next request after this one than with the JobQueue system.
  38  *
  39  * In CLI mode, deferred updates will run:
  40  *   - a) During DeferredUpdates::addUpdate if no LBFactory DB handles have writes pending
  41  *   - b) On commit of an LBFactory DB handle if no other such handles have writes pending
  42  *   - c) During an LBFactory::waitForReplication call if no LBFactory DBs have writes pending
  43  *   - d) When the queue is large and an LBFactory DB handle commits (EnqueueableDataUpdate only)
  44  *   - e) At the completion of Maintenance::execute()
  45  *
  46  * @see Maintenance::setLBFactoryTriggers
  47  *
  48  * When updates are deferred, they go into one two FIFO "top-queues" (one for pre-send and one
  49  * for post-send). Updates enqueued *during* doUpdate() of a "top" update go into the "sub-queue"
  50  * for that update. After that method finishes, the sub-queue is run until drained. This continues
  51  * for each top-queue job until the entire top queue is drained. This happens for the pre-send
  52  * top-queue, and later on, the post-send top-queue, in execute().
  53  *
  54  * @since 1.19
  55  */
  56 class DeferredUpdates {
  57         /** @var DeferrableUpdate[] Updates to be deferred until before request end */
  58         private static $preSendUpdates = [];
  59         /** @var DeferrableUpdate[] Updates to be deferred until after request end */
  60         private static $postSendUpdates = [];
  61
  62         const ALL = 0; // all updates; in web requests, use only after flushing the output buffer
  63         const PRESEND = 1; // for updates that should run before flushing output buffer
  64         const POSTSEND = 2; // for updates that should run after flushing output buffer
  65
  66         const BIG_QUEUE_SIZE = 100;
  67
  68         /** @var array|null Information about the current execute() call or null if not running */
  69         private static $executeContext;
  70
  71         /**
  72          * Add an update to the deferred list to be run later by execute()
  73          *
  74          * In CLI mode, callback magic will also be used to run updates when safe
  75          *
  76          * @param DeferrableUpdate $update Some object that implements doUpdate()
  77          * @param int $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
  78          */
  79         public static function addUpdate( DeferrableUpdate $update, $stage = self::POSTSEND ) {
  80                 global $wgCommandLineMode;
  81
  82                 if (
  83                         self::$executeContext &&
  84                         self::$executeContext['stage'] >= $stage &&
  85                         !( $update instanceof MergeableUpdate )
  86                 ) {
  87                         // This is a sub-DeferredUpdate; run it right after its parent update.
  88                         // Also, while post-send updates are running, push any "pre-send" jobs to the
  89                         // active post-send queue to make sure they get run this round (or at all).
  90                         self::$executeContext['subqueue'][] = $update;
  91
  92                         return;
  93                 }
  94
  95                 if ( $stage === self::PRESEND ) {
  96                         self::push( self::$preSendUpdates, $update );
  97                 } else {
  98                         self::push( self::$postSendUpdates, $update );
  99                 }
 100
 101                 // Try to run the updates now if in CLI mode and no transaction is active.
 102                 // This covers scripts that don't/barely use the DB but make updates to other stores.
 103                 if ( $wgCommandLineMode ) {
 104                         self::tryOpportunisticExecute( 'run' );
 105                 }
 106         }
 107
 108         /**
 109          * Add a callable update. In a lot of cases, we just need a callback/closure,
 110          * defining a new DeferrableUpdate object is not necessary
 111          *
 112          * @see MWCallableUpdate::__construct()
 113          *
 114          * @param callable $callable
 115          * @param int $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
 116          * @param IDatabase|IDatabase[]|null $dbw Abort if this DB is rolled back [optional] (since 1.28)
 117          */
 118         public static function addCallableUpdate(
 119                 $callable, $stage = self::POSTSEND, $dbw = null
 120         ) {
 121                 self::addUpdate( new MWCallableUpdate( $callable, wfGetCaller(), $dbw ), $stage );
 122         }
 123
 124         /**
 125          * Do any deferred updates and clear the list
 126          *
 127          * If $stage is self::ALL then the queue of PRESEND updates will be resolved,
 128          * followed by the queue of POSTSEND updates
 129          *
 130          * @param string $mode Use "enqueue" to use the job queue when possible [Default: "run"]
 131          * @param int $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL) (since 1.27)
 132          */
 133         public static function doUpdates( $mode = 'run', $stage = self::ALL ) {
 134                 $stageEffective = ( $stage === self::ALL ) ? self::POSTSEND : $stage;
 135                 // For ALL mode, make sure that any PRESEND updates added along the way get run.
 136                 // Normally, these use the subqueue, but that isn't true for MergeableUpdate items.
 137                 do {
 138                         if ( $stage === self::ALL || $stage === self::PRESEND ) {
 139                                 self::execute( self::$preSendUpdates, $mode, $stageEffective );
 140                         }
 141
 142                         if ( $stage === self::ALL || $stage == self::POSTSEND ) {
 143                                 self::execute( self::$postSendUpdates, $mode, $stageEffective );
 144                         }
 145                 } while ( $stage === self::ALL && self::$preSendUpdates );
 146         }
 147
 148         /**
 149          * @param DeferrableUpdate[] $queue
 150          * @param DeferrableUpdate $update
 151          */
 152         private static function push( array &$queue, DeferrableUpdate $update ) {
 153                 if ( $update instanceof MergeableUpdate ) {
 154                         $class = get_class( $update ); // fully-qualified class
 155                         if ( isset( $queue[$class] ) ) {
 156                                 /** @var MergeableUpdate $existingUpdate */
 157                                 $existingUpdate = $queue[$class];
 158                                 $existingUpdate->merge( $update );
 159                                 // Move the update to the end to handle things like mergeable purge
 160                                 // updates that might depend on the prior updates in the queue running
 161                                 unset( $queue[$class] );
 162                                 $queue[$class] = $existingUpdate;
 163                         } else {
 164                                 $queue[$class] = $update;
 165                         }
 166                 } else {
 167                         $queue[] = $update;
 168                 }
 169         }
 170
 171         /**
 172          * Immediately run/queue a list of updates
 173          *
 174          * @param DeferrableUpdate[] &$queue List of DeferrableUpdate objects
 175          * @param string $mode Use "enqueue" to use the job queue when possible
 176          * @param int $stage Class constant (PRESEND, POSTSEND) (since 1.28)
 177          * @throws ErrorPageError Happens on top-level calls
 178          * @throws Exception Happens on second-level calls
 179          */
 180         protected static function execute( array &$queue, $mode, $stage ) {
 181                 $services = MediaWikiServices::getInstance();
 182                 $stats = $services->getStatsdDataFactory();
 183                 $lbFactory = $services->getDBLoadBalancerFactory();
 184                 $method = RequestContext::getMain()->getRequest()->getMethod();
 185
 186                 $ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
 187
 188                 /** @var ErrorPageError $reportableError */
 189                 $reportableError = null;
 190                 /** @var DeferrableUpdate[] $updates Snapshot of queue */
 191                 $updates = $queue;
 192
 193                 // Keep doing rounds of updates until none get enqueued...
 194                 while ( $updates ) {
 195                         $queue = []; // clear the queue
 196
 197                         // Order will be DataUpdate followed by generic DeferrableUpdate tasks
 198                         $updatesByType = [ 'data' => [], 'generic' => [] ];
 199                         foreach ( $updates as $du ) {
 200                                 if ( $du instanceof DataUpdate ) {
 201                                         $du->setTransactionTicket( $ticket );
 202                                         $updatesByType['data'][] = $du;
 203                                 } else {
 204                                         $updatesByType['generic'][] = $du;
 205                                 }
 206
 207                                 $name = ( $du instanceof DeferrableCallback )
 208                                         ? get_class( $du ) . '-' . $du->getOrigin()
 209                                         : get_class( $du );
 210                                 $stats->increment( 'deferred_updates.' . $method . '.' . $name );
 211                         }
 212
 213                         // Execute all remaining tasks...
 214                         foreach ( $updatesByType as $updatesForType ) {
 215                                 foreach ( $updatesForType as $update ) {
 216                                         self::$executeContext = [ 'stage' => $stage, 'subqueue' => [] ];
 217                                         try {
 218                                                 /** @var DeferrableUpdate $update */
 219                                                 $guiError = self::runUpdate( $update, $lbFactory, $mode, $stage );
 220                                                 $reportableError = $reportableError ?: $guiError;
 221                                                 // Do the subqueue updates for $update until there are none
 222                                                 while ( self::$executeContext['subqueue'] ) {
 223                                                         $subUpdate = reset( self::$executeContext['subqueue'] );
 224                                                         $firstKey = key( self::$executeContext['subqueue'] );
 225                                                         unset( self::$executeContext['subqueue'][$firstKey] );
 226
 227                                                         if ( $subUpdate instanceof DataUpdate ) {
 228                                                                 $subUpdate->setTransactionTicket( $ticket );
 229                                                         }
 230
 231                                                         $guiError = self::runUpdate( $subUpdate, $lbFactory, $mode, $stage );
 232                                                         $reportableError = $reportableError ?: $guiError;
 233                                                 }
 234                                         } finally {
 235                                                 // Make sure we always clean up the context.
 236                                                 // Losing updates while rewinding the stack is acceptable,
 237                                                 // losing updates that are added later is not.
 238                                                 self::$executeContext = null;
 239                                         }
 240                                 }
 241                         }
 242
 243                         $updates = $queue; // new snapshot of queue (check for new entries)
 244                 }
 245
 246                 if ( $reportableError ) {
 247                         throw $reportableError; // throw the first of any GUI errors
 248                 }
 249         }
 250
 251         /**
 252          * @param DeferrableUpdate $update
 253          * @param LBFactory $lbFactory
 254          * @param string $mode
 255          * @param int $stage
 256          * @return ErrorPageError|null
 257          */
 258         private static function runUpdate(
 259                 DeferrableUpdate $update, LBFactory $lbFactory, $mode, $stage
 260         ) {
 261                 $guiError = null;
 262                 try {
 263                         if ( $mode === 'enqueue' && $update instanceof EnqueueableDataUpdate ) {
 264                                 // Run only the job enqueue logic to complete the update later
 265                                 $spec = $update->getAsJobSpecification();
 266                                 $domain = $spec['domain'] ?? $spec['wiki'];
 267                                 JobQueueGroup::singleton( $domain )->push( $spec['job'] );
 268                         } elseif ( $update instanceof TransactionRoundDefiningUpdate ) {
 269                                 $update->doUpdate();
 270                         } else {
 271                                 // Run the bulk of the update now
 272                                 $fnameTrxOwner = get_class( $update ) . '::doUpdate';
 273                                 $lbFactory->beginMasterChanges( $fnameTrxOwner );
 274                                 $update->doUpdate();
 275                                 $lbFactory->commitMasterChanges( $fnameTrxOwner );
 276                         }
 277                 } catch ( Exception $e ) {
 278                         // Reporting GUI exceptions does not work post-send
 279                         if ( $e instanceof ErrorPageError && $stage === self::PRESEND ) {
 280                                 $guiError = $e;
 281                         }
 282                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 283
 284                         // VW-style hack to work around T190178, so we can make sure
 285                         // PageMetaDataUpdater doesn't throw exceptions.
 286                         if ( defined( 'MW_PHPUNIT_TEST' ) ) {
 287                                 throw $e;
 288                         }
 289                 }
 290
 291                 return $guiError;
 292         }
 293
 294         /**
 295          * Run all deferred updates immediately if there are no DB writes active
 296          *
 297          * If there are many deferred updates pending, $mode is 'run', and there
 298          * are still busy LBFactory database handles, then any EnqueueableDataUpdate
 299          * tasks might be enqueued as jobs to be executed later.
 300          *
 301          * @param string $mode Use "enqueue" to use the job queue when possible
 302          * @return bool Whether updates were allowed to run
 303          * @since 1.28
 304          */
 305         public static function tryOpportunisticExecute( $mode = 'run' ) {
 306                 // execute() loop is already running
 307                 if ( self::$executeContext ) {
 308                         return false;
 309                 }
 310
 311                 // Avoiding running updates without them having outer scope
 312                 if ( !self::areDatabaseTransactionsActive() ) {
 313                         self::doUpdates( $mode );
 314                         return true;
 315                 }
 316
 317                 if ( self::pendingUpdatesCount() >= self::BIG_QUEUE_SIZE ) {
 318                         // If we cannot run the updates with outer transaction context, try to
 319                         // at least enqueue all the updates that support queueing to job queue
 320                         self::$preSendUpdates = self::enqueueUpdates( self::$preSendUpdates );
 321                         self::$postSendUpdates = self::enqueueUpdates( self::$postSendUpdates );
 322                 }
 323
 324                 return !self::pendingUpdatesCount();
 325         }
 326
 327         /**
 328          * Enqueue a job for each EnqueueableDataUpdate item and return the other items
 329          *
 330          * @param DeferrableUpdate[] $updates A list of deferred update instances
 331          * @return DeferrableUpdate[] Remaining updates that do not support being queued
 332          */
 333         private static function enqueueUpdates( array $updates ) {
 334                 $remaining = [];
 335
 336                 foreach ( $updates as $update ) {
 337                         if ( $update instanceof EnqueueableDataUpdate ) {
 338                                 $spec = $update->getAsJobSpecification();
 339                                 $domain = $spec['domain'] ?? $spec['wiki'];
 340                                 JobQueueGroup::singleton( $domain )->push( $spec['job'] );
 341                         } else {
 342                                 $remaining[] = $update;
 343                         }
 344                 }
 345
 346                 return $remaining;
 347         }
 348
 349         /**
 350          * @return int Number of enqueued updates
 351          * @since 1.28
 352          */
 353         public static function pendingUpdatesCount() {
 354                 return count( self::$preSendUpdates ) + count( self::$postSendUpdates );
 355         }
 356
 357         /**
 358          * @param int $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL)
 359          * @return DeferrableUpdate[]
 360          * @since 1.29
 361          */
 362         public static function getPendingUpdates( $stage = self::ALL ) {
 363                 $updates = [];
 364                 if ( $stage === self::ALL || $stage === self::PRESEND ) {
 365                         $updates = array_merge( $updates, self::$preSendUpdates );
 366                 }
 367                 if ( $stage === self::ALL || $stage === self::POSTSEND ) {
 368                         $updates = array_merge( $updates, self::$postSendUpdates );
 369                 }
 370                 return $updates;
 371         }
 372
 373         /**
 374          * Clear all pending updates without performing them. Generally, you don't
 375          * want or need to call this. Unit tests need it though.
 376          */
 377         public static function clearPendingUpdates() {
 378                 self::$preSendUpdates = [];
 379                 self::$postSendUpdates = [];
 380         }
 381
 382         /**
 383          * @return bool If a transaction round is active or connection is not ready for commit()
 384          */
 385         private static function areDatabaseTransactionsActive() {
 386                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 387                 if ( $lbFactory->hasTransactionRound() || !$lbFactory->isReadyForRoundOperations() ) {
 388                         return true;
 389                 }
 390
 391                 $connsBusy = false;
 392                 $lbFactory->forEachLB( function ( LoadBalancer $lb ) use ( &$connsBusy ) {
 393                         $lb->forEachOpenMasterConnection( function ( IDatabase $conn ) use ( &$connsBusy ) {
 394                                 if ( $conn->writesOrCallbacksPending() || $conn->explicitTrxActive() ) {
 395                                         $connsBusy = true;
 396                                 }
 397                         } );
 398                 } );
 399
 400                 return $connsBusy;
 401         }
 402 }