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 $wgCommandLineMode;
  81
  82                 if ( self::$executeContext && self::$executeContext['stage'] >= $stage ) {
  83                         // This is a sub-DeferredUpdate; run it right after its parent update.
  84                         // Also, while post-send updates are running, push any "pre-send" jobs to the
  85                         // active post-send queue to make sure they get run this round (or at all).
  86                         self::$executeContext['subqueue'][] = $update;
  87
  88                         return;
  89                 }
  90
  91                 if ( $stage === self::PRESEND ) {
  92                         self::push( self::$preSendUpdates, $update );
  93                 } else {
  94                         self::push( self::$postSendUpdates, $update );
  95                 }
  96
  97                 // Try to run the updates now if in CLI mode and no transaction is active.
  98                 // This covers scripts that don't/barely use the DB but make updates to other stores.
  99                 if ( $wgCommandLineMode ) {
 100                         self::tryOpportunisticExecute( 'run' );
 101                 }
 102         }
 103
 104         /**
 105          * Add a callable update. In a lot of cases, we just need a callback/closure,
 106          * defining a new DeferrableUpdate object is not necessary
 107          *
 108          * @see MWCallableUpdate::__construct()
 109          *
 110          * @param callable $callable
 111          * @param int $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
 112          * @param IDatabase|IDatabase[]|null $dbw Abort if this DB is rolled back [optional] (since 1.28)
 113          */
 114         public static function addCallableUpdate(
 115                 $callable, $stage = self::POSTSEND, $dbw = null
 116         ) {
 117                 self::addUpdate( new MWCallableUpdate( $callable, wfGetCaller(), $dbw ), $stage );
 118         }
 119
 120         /**
 121          * Do any deferred updates and clear the list
 122          *
 123          * @param string $mode Use "enqueue" to use the job queue when possible [Default: "run"]
 124          * @param int $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL) (since 1.27)
 125          */
 126         public static function doUpdates( $mode = 'run', $stage = self::ALL ) {
 127                 $stageEffective = ( $stage === self::ALL ) ? self::POSTSEND : $stage;
 128
 129                 if ( $stage === self::ALL || $stage === self::PRESEND ) {
 130                         self::execute( self::$preSendUpdates, $mode, $stageEffective );
 131                 }
 132
 133                 if ( $stage === self::ALL || $stage == self::POSTSEND ) {
 134                         self::execute( self::$postSendUpdates, $mode, $stageEffective );
 135                 }
 136         }
 137
 138         /**
 139          * @param bool $value Whether to just immediately run updates in addUpdate()
 140          * @since 1.28
 141          * @deprecated 1.29 Causes issues in Web-executed jobs - see T165714 and T100085.
 142          */
 143         public static function setImmediateMode( $value ) {
 144                 wfDeprecated( __METHOD__, '1.29' );
 145         }
 146
 147         /**
 148          * @param DeferrableUpdate[] $queue
 149          * @param DeferrableUpdate $update
 150          */
 151         private static function push( array &$queue, DeferrableUpdate $update ) {
 152                 if ( $update instanceof MergeableUpdate ) {
 153                         $class = get_class( $update ); // fully-qualified class
 154                         if ( isset( $queue[$class] ) ) {
 155                                 /** @var MergeableUpdate $existingUpdate */
 156                                 $existingUpdate = $queue[$class];
 157                                 $existingUpdate->merge( $update );
 158                         } else {
 159                                 $queue[$class] = $update;
 160                         }
 161                 } else {
 162                         $queue[] = $update;
 163                 }
 164         }
 165
 166         /**
 167          * Immediately run/queue a list of updates
 168          *
 169          * @param DeferrableUpdate[] &$queue List of DeferrableUpdate objects
 170          * @param string $mode Use "enqueue" to use the job queue when possible
 171          * @param int $stage Class constant (PRESEND, POSTSEND) (since 1.28)
 172          * @throws ErrorPageError Happens on top-level calls
 173          * @throws Exception Happens on second-level calls
 174          */
 175         protected static function execute( array &$queue, $mode, $stage ) {
 176                 $services = MediaWikiServices::getInstance();
 177                 $stats = $services->getStatsdDataFactory();
 178                 $lbFactory = $services->getDBLoadBalancerFactory();
 179                 $method = RequestContext::getMain()->getRequest()->getMethod();
 180
 181                 $ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
 182
 183                 /** @var ErrorPageError $reportableError */
 184                 $reportableError = null;
 185                 /** @var DeferrableUpdate[] $updates Snapshot of queue */
 186                 $updates = $queue;
 187
 188                 // Keep doing rounds of updates until none get enqueued...
 189                 while ( $updates ) {
 190                         $queue = []; // clear the queue
 191
 192                         // Order will be DataUpdate followed by generic DeferrableUpdate tasks
 193                         $updatesByType = [ 'data' => [], 'generic' => [] ];
 194                         foreach ( $updates as $du ) {
 195                                 if ( $du instanceof DataUpdate ) {
 196                                         $du->setTransactionTicket( $ticket );
 197                                         $updatesByType['data'][] = $du;
 198                                 } else {
 199                                         $updatesByType['generic'][] = $du;
 200                                 }
 201
 202                                 $name = ( $du instanceof DeferrableCallback )
 203                                         ? get_class( $du ) . '-' . $du->getOrigin()
 204                                         : get_class( $du );
 205                                 $stats->increment( 'deferred_updates.' . $method . '.' . $name );
 206                         }
 207
 208                         // Execute all remaining tasks...
 209                         foreach ( $updatesByType as $updatesForType ) {
 210                                 foreach ( $updatesForType as $update ) {
 211                                         self::$executeContext = [ 'stage' => $stage, 'subqueue' => [] ];
 212                                         try {
 213                                                 /** @var DeferrableUpdate $update */
 214                                                 $guiError = self::runUpdate( $update, $lbFactory, $mode, $stage );
 215                                                 $reportableError = $reportableError ?: $guiError;
 216                                                 // Do the subqueue updates for $update until there are none
 217                                                 while ( self::$executeContext['subqueue'] ) {
 218                                                         $subUpdate = reset( self::$executeContext['subqueue'] );
 219                                                         $firstKey = key( self::$executeContext['subqueue'] );
 220                                                         unset( self::$executeContext['subqueue'][$firstKey] );
 221
 222                                                         if ( $subUpdate instanceof DataUpdate ) {
 223                                                                 $subUpdate->setTransactionTicket( $ticket );
 224                                                         }
 225
 226                                                         $guiError = self::runUpdate( $subUpdate, $lbFactory, $mode, $stage );
 227                                                         $reportableError = $reportableError ?: $guiError;
 228                                                 }
 229                                         } finally {
 230                                                 // Make sure we always clean up the context.
 231                                                 // Losing updates while rewinding the stack is acceptable,
 232                                                 // losing updates that are added later is not.
 233                                                 self::$executeContext = null;
 234                                         }
 235                                 }
 236                         }
 237
 238                         $updates = $queue; // new snapshot of queue (check for new entries)
 239                 }
 240
 241                 if ( $reportableError ) {
 242                         throw $reportableError; // throw the first of any GUI errors
 243                 }
 244         }
 245
 246         /**
 247          * @param DeferrableUpdate $update
 248          * @param LBFactory $lbFactory
 249          * @param string $mode
 250          * @param int $stage
 251          * @return ErrorPageError|null
 252          */
 253         private static function runUpdate(
 254                 DeferrableUpdate $update, LBFactory $lbFactory, $mode, $stage
 255         ) {
 256                 $guiError = null;
 257                 try {
 258                         if ( $mode === 'enqueue' && $update instanceof EnqueueableDataUpdate ) {
 259                                 // Run only the job enqueue logic to complete the update later
 260                                 $spec = $update->getAsJobSpecification();
 261                                 JobQueueGroup::singleton( $spec['wiki'] )->push( $spec['job'] );
 262                         } elseif ( $update instanceof TransactionRoundDefiningUpdate ) {
 263                                 $update->doUpdate();
 264                         } else {
 265                                 // Run the bulk of the update now
 266                                 $fnameTrxOwner = get_class( $update ) . '::doUpdate';
 267                                 $lbFactory->beginMasterChanges( $fnameTrxOwner );
 268                                 $update->doUpdate();
 269                                 $lbFactory->commitMasterChanges( $fnameTrxOwner );
 270                         }
 271                 } catch ( Exception $e ) {
 272                         // Reporting GUI exceptions does not work post-send
 273                         if ( $e instanceof ErrorPageError && $stage === self::PRESEND ) {
 274                                 $guiError = $e;
 275                         }
 276                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 277
 278                         // VW-style hack to work around T190178, so we can make sure
 279                         // PageMetaDataUpdater doesn't throw exceptions.
 280                         if ( defined( 'MW_PHPUNIT_TEST' ) ) {
 281                                 throw $e;
 282                         }
 283                 }
 284
 285                 return $guiError;
 286         }
 287
 288         /**
 289          * Run all deferred updates immediately if there are no DB writes active
 290          *
 291          * If there are many deferred updates pending, $mode is 'run', and there
 292          * are still busy LBFactory database handles, then any EnqueueableDataUpdate
 293          * tasks might be enqueued as jobs to be executed later.
 294          *
 295          * @param string $mode Use "enqueue" to use the job queue when possible
 296          * @return bool Whether updates were allowed to run
 297          * @since 1.28
 298          */
 299         public static function tryOpportunisticExecute( $mode = 'run' ) {
 300                 // execute() loop is already running
 301                 if ( self::$executeContext ) {
 302                         return false;
 303                 }
 304
 305                 // Avoiding running updates without them having outer scope
 306                 if ( !self::areDatabaseTransactionsActive() ) {
 307                         self::doUpdates( $mode );
 308                         return true;
 309                 }
 310
 311                 if ( self::pendingUpdatesCount() >= self::BIG_QUEUE_SIZE ) {
 312                         // If we cannot run the updates with outer transaction context, try to
 313                         // at least enqueue all the updates that support queueing to job queue
 314                         self::$preSendUpdates = self::enqueueUpdates( self::$preSendUpdates );
 315                         self::$postSendUpdates = self::enqueueUpdates( self::$postSendUpdates );
 316                 }
 317
 318                 return !self::pendingUpdatesCount();
 319         }
 320
 321         /**
 322          * Enqueue a job for each EnqueueableDataUpdate item and return the other items
 323          *
 324          * @param DeferrableUpdate[] $updates A list of deferred update instances
 325          * @return DeferrableUpdate[] Remaining updates that do not support being queued
 326          */
 327         private static function enqueueUpdates( array $updates ) {
 328                 $remaining = [];
 329
 330                 foreach ( $updates as $update ) {
 331                         if ( $update instanceof EnqueueableDataUpdate ) {
 332                                 $spec = $update->getAsJobSpecification();
 333                                 JobQueueGroup::singleton( $spec['wiki'] )->push( $spec['job'] );
 334                         } else {
 335                                 $remaining[] = $update;
 336                         }
 337                 }
 338
 339                 return $remaining;
 340         }
 341
 342         /**
 343          * @return int Number of enqueued updates
 344          * @since 1.28
 345          */
 346         public static function pendingUpdatesCount() {
 347                 return count( self::$preSendUpdates ) + count( self::$postSendUpdates );
 348         }
 349
 350         /**
 351          * @param int $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL)
 352          * @return DeferrableUpdate[]
 353          * @since 1.29
 354          */
 355         public static function getPendingUpdates( $stage = self::ALL ) {
 356                 $updates = [];
 357                 if ( $stage === self::ALL || $stage === self::PRESEND ) {
 358                         $updates = array_merge( $updates, self::$preSendUpdates );
 359                 }
 360                 if ( $stage === self::ALL || $stage === self::POSTSEND ) {
 361                         $updates = array_merge( $updates, self::$postSendUpdates );
 362                 }
 363                 return $updates;
 364         }
 365
 366         /**
 367          * Clear all pending updates without performing them. Generally, you don't
 368          * want or need to call this. Unit tests need it though.
 369          */
 370         public static function clearPendingUpdates() {
 371                 self::$preSendUpdates = [];
 372                 self::$postSendUpdates = [];
 373         }
 374
 375         /**
 376          * @return bool If a transaction round is active or connection is not ready for commit()
 377          */
 378         private static function areDatabaseTransactionsActive() {
 379                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 380                 if ( $lbFactory->hasTransactionRound() || !$lbFactory->isReadyForRoundOperations() ) {
 381                         return true;
 382                 }
 383
 384                 $connsBusy = false;
 385                 $lbFactory->forEachLB( function ( LoadBalancer $lb ) use ( &$connsBusy ) {
 386                         $lb->forEachOpenMasterConnection( function ( IDatabase $conn ) use ( &$connsBusy ) {
 387                                 if ( $conn->writesOrCallbacksPending() || $conn->explicitTrxActive() ) {
 388                                         $connsBusy = true;
 389                                 }
 390                         } );
 391                 } );
 392
 393                 return $connsBusy;
 394         }
 395 }