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