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, updates run immediately if no DB writes are pending. Otherwise, they run when:
  40  *   - a) Any waitForReplication() call if no writes are pending on any DB
  41  *   - b) A commit happens on Maintenance::getDB( DB_MASTER ) if no writes are pending on any DB
  42  *   - c) EnqueueableDataUpdate tasks may enqueue on commit of Maintenance::getDB( DB_MASTER )
  43  *   - d) At the completion of Maintenance::execute()
  44  *
  45  * When updates are deferred, they go into one two FIFO "top-queues" (one for pre-send and one
  46  * for post-send). Updates enqueued *during* doUpdate() of a "top" update go into the "sub-queue"
  47  * for that update. After that method finishes, the sub-queue is run until drained. This continues
  48  * for each top-queue job until the entire top queue is drained. This happens for the pre-send
  49  * top-queue, and later on, the post-send top-queue, in execute().
  50  *
  51  * @since 1.19
  52  */
  53 class DeferredUpdates {
  54         /** @var DeferrableUpdate[] Updates to be deferred until before request end */
  55         private static $preSendUpdates = [];
  56         /** @var DeferrableUpdate[] Updates to be deferred until after request end */
  57         private static $postSendUpdates = [];
  58         /** @var bool Whether to just run updates in addUpdate() */
  59         private static $immediateMode = false;
  60
  61         const ALL = 0; // all updates; in web requests, use only after flushing the output buffer
  62         const PRESEND = 1; // for updates that should run before flushing output buffer
  63         const POSTSEND = 2; // for updates that should run after flushing output buffer
  64
  65         const BIG_QUEUE_SIZE = 100;
  66
  67         /** @var array|null Information about the current execute() call or null if not running */
  68         private static $executeContext;
  69
  70         /**
  71          * Add an update to the deferred list to be run later by execute()
  72          *
  73          * In CLI mode, callback magic will also be used to run updates when safe
  74          *
  75          * @param DeferrableUpdate $update Some object that implements doUpdate()
  76          * @param integer $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
  77          */
  78         public static function addUpdate( DeferrableUpdate $update, $stage = self::POSTSEND ) {
  79                 global $wgCommandLineMode;
  80
  81                 // This is a sub-DeferredUpdate, run it right after its parent update
  82                 if ( self::$executeContext && self::$executeContext['stage'] >= $stage ) {
  83                         self::$executeContext['subqueue'][] = $update;
  84                         return;
  85                 }
  86
  87                 if ( $stage === self::PRESEND ) {
  88                         self::push( self::$preSendUpdates, $update );
  89                 } else {
  90                         self::push( self::$postSendUpdates, $update );
  91                 }
  92
  93                 if ( self::$immediateMode ) {
  94                         // No more explicit doUpdates() calls will happen, so run this now
  95                         self::doUpdates( 'run' );
  96                         return;
  97                 }
  98
  99                 // Try to run the updates now if in CLI mode and no transaction is active.
 100                 // This covers scripts that don't/barely use the DB but make updates to other stores.
 101                 if ( $wgCommandLineMode ) {
 102                         self::tryOpportunisticExecute( 'run' );
 103                 }
 104         }
 105
 106         /**
 107          * Add a callable update. In a lot of cases, we just need a callback/closure,
 108          * defining a new DeferrableUpdate object is not necessary
 109          *
 110          * @see MWCallableUpdate::__construct()
 111          *
 112          * @param callable $callable
 113          * @param integer $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
 114          * @param IDatabase|null $dbw Abort if this DB is rolled back [optional] (since 1.28)
 115          */
 116         public static function addCallableUpdate(
 117                 $callable, $stage = self::POSTSEND, IDatabase $dbw = null
 118         ) {
 119                 self::addUpdate( new MWCallableUpdate( $callable, wfGetCaller(), $dbw ), $stage );
 120         }
 121
 122         /**
 123          * Do any deferred updates and clear the list
 124          *
 125          * @param string $mode Use "enqueue" to use the job queue when possible [Default: "run"]
 126          * @param integer $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL) (since 1.27)
 127          */
 128         public static function doUpdates( $mode = 'run', $stage = self::ALL ) {
 129                 $stageEffective = ( $stage === self::ALL ) ? self::POSTSEND : $stage;
 130
 131                 if ( $stage === self::ALL || $stage === self::PRESEND ) {
 132                         self::execute( self::$preSendUpdates, $mode, $stageEffective );
 133                 }
 134
 135                 if ( $stage === self::ALL || $stage == self::POSTSEND ) {
 136                         self::execute( self::$postSendUpdates, $mode, $stageEffective );
 137                 }
 138         }
 139
 140         /**
 141          * @param bool $value Whether to just immediately run updates in addUpdate()
 142          * @since 1.28
 143          */
 144         public static function setImmediateMode( $value ) {
 145                 self::$immediateMode = (bool)$value;
 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 $existingUpdate MergeableUpdate */
 157                                 $existingUpdate = $queue[$class];
 158                                 $existingUpdate->merge( $update );
 159                         } else {
 160                                 $queue[$class] = $update;
 161                         }
 162                 } else {
 163                         $queue[] = $update;
 164                 }
 165         }
 166
 167         /**
 168          * Immediately run/queue a list of updates
 169          *
 170          * @param DeferrableUpdate[] &$queue List of DeferrableUpdate objects
 171          * @param string $mode Use "enqueue" to use the job queue when possible
 172          * @param integer $stage Class constant (PRESEND, POSTSEND) (since 1.28)
 173          * @throws ErrorPageError Happens on top-level calls
 174          * @throws Exception Happens on second-level calls
 175          */
 176         protected static function execute( array &$queue, $mode, $stage ) {
 177                 $services = MediaWikiServices::getInstance();
 178                 $stats = $services->getStatsdDataFactory();
 179                 $lbFactory = $services->getDBLoadBalancerFactory();
 180                 $method = RequestContext::getMain()->getRequest()->getMethod();
 181
 182                 $ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
 183
 184                 /** @var ErrorPageError $reportableError */
 185                 $reportableError = null;
 186                 /** @var DeferrableUpdate[] $updates Snapshot of queue */
 187                 $updates = $queue;
 188
 189                 // Keep doing rounds of updates until none get enqueued...
 190                 while ( $updates ) {
 191                         $queue = []; // clear the queue
 192
 193                         if ( $mode === 'enqueue' ) {
 194                                 try {
 195                                         // Push enqueuable updates to the job queue and get the rest
 196                                         $updates = self::enqueueUpdates( $updates );
 197                                 } catch ( Exception $e ) {
 198                                         // Let other updates have a chance to run if this failed
 199                                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 200                                 }
 201                         }
 202
 203                         // Order will be DataUpdate followed by generic DeferrableUpdate tasks
 204                         $updatesByType = [ 'data' => [], 'generic' => [] ];
 205                         foreach ( $updates as $du ) {
 206                                 if ( $du instanceof DataUpdate ) {
 207                                         $du->setTransactionTicket( $ticket );
 208                                         $updatesByType['data'][] = $du;
 209                                 } else {
 210                                         $updatesByType['generic'][] = $du;
 211                                 }
 212
 213                                 $name = ( $du instanceof DeferrableCallback )
 214                                         ? get_class( $du ) . '-' . $du->getOrigin()
 215                                         : get_class( $du );
 216                                 $stats->increment( 'deferred_updates.' . $method . '.' . $name );
 217                         }
 218
 219                         // Execute all remaining tasks...
 220                         foreach ( $updatesByType as $updatesForType ) {
 221                                 foreach ( $updatesForType as $update ) {
 222                                         self::$executeContext = [
 223                                                 'update' => $update,
 224                                                 'stage' => $stage,
 225                                                 'subqueue' => []
 226                                         ];
 227                                         /** @var DeferrableUpdate $update */
 228                                         $guiError = self::runUpdate( $update, $lbFactory, $stage );
 229                                         $reportableError = $reportableError ?: $guiError;
 230                                         // Do the subqueue updates for $update until there are none
 231                                         while ( self::$executeContext['subqueue'] ) {
 232                                                 $subUpdate = reset( self::$executeContext['subqueue'] );
 233                                                 $firstKey = key( self::$executeContext['subqueue'] );
 234                                                 unset( self::$executeContext['subqueue'][$firstKey] );
 235
 236                                                 if ( $subUpdate instanceof DataUpdate ) {
 237                                                         $subUpdate->setTransactionTicket( $ticket );
 238                                                 }
 239
 240                                                 $guiError = self::runUpdate( $subUpdate, $lbFactory, $stage );
 241                                                 $reportableError = $reportableError ?: $guiError;
 242                                         }
 243                                         self::$executeContext = null;
 244                                 }
 245                         }
 246
 247                         $updates = $queue; // new snapshot of queue (check for new entries)
 248                 }
 249
 250                 if ( $reportableError ) {
 251                         throw $reportableError; // throw the first of any GUI errors
 252                 }
 253         }
 254
 255         /**
 256          * @param DeferrableUpdate $update
 257          * @param LBFactory $lbFactory
 258          * @param integer $stage
 259          * @return ErrorPageError|null
 260          */
 261         private static function runUpdate( DeferrableUpdate $update, LBFactory $lbFactory, $stage ) {
 262                 $guiError = null;
 263                 try {
 264                         $fnameTrxOwner = get_class( $update ) . '::doUpdate';
 265                         $lbFactory->beginMasterChanges( $fnameTrxOwner );
 266                         $update->doUpdate();
 267                         $lbFactory->commitMasterChanges( $fnameTrxOwner );
 268                 } catch ( Exception $e ) {
 269                         // Reporting GUI exceptions does not work post-send
 270                         if ( $e instanceof ErrorPageError && $stage === self::PRESEND ) {
 271                                 $guiError = $e;
 272                         }
 273                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 274                 }
 275
 276                 return $guiError;
 277         }
 278
 279         /**
 280          * Run all deferred updates immediately if there are no DB writes active
 281          *
 282          * If $mode is 'run' but there are busy databates, EnqueueableDataUpdate
 283          * tasks will be enqueued anyway for the sake of progress.
 284          *
 285          * @param string $mode Use "enqueue" to use the job queue when possible
 286          * @return bool Whether updates were allowed to run
 287          * @since 1.28
 288          */
 289         public static function tryOpportunisticExecute( $mode = 'run' ) {
 290                 // execute() loop is already running
 291                 if ( self::$executeContext ) {
 292                         return false;
 293                 }
 294
 295                 // Avoiding running updates without them having outer scope
 296                 if ( !self::getBusyDbConnections() ) {
 297                         self::doUpdates( $mode );
 298                         return true;
 299                 }
 300
 301                 if ( self::pendingUpdatesCount() >= self::BIG_QUEUE_SIZE ) {
 302                         // If we cannot run the updates with outer transaction context, try to
 303                         // at least enqueue all the updates that support queueing to job queue
 304                         self::$preSendUpdates = self::enqueueUpdates( self::$preSendUpdates );
 305                         self::$postSendUpdates = self::enqueueUpdates( self::$postSendUpdates );
 306                 }
 307
 308                 return !self::pendingUpdatesCount();
 309         }
 310
 311         /**
 312          * Enqueue a job for each EnqueueableDataUpdate item and return the other items
 313          *
 314          * @param DeferrableUpdate[] $updates A list of deferred update instances
 315          * @return DeferrableUpdate[] Remaining updates that do not support being queued
 316          */
 317         private static function enqueueUpdates( array $updates ) {
 318                 $remaining = [];
 319
 320                 foreach ( $updates as $update ) {
 321                         if ( $update instanceof EnqueueableDataUpdate ) {
 322                                 $spec = $update->getAsJobSpecification();
 323                                 JobQueueGroup::singleton( $spec['wiki'] )->push( $spec['job'] );
 324                         } else {
 325                                 $remaining[] = $update;
 326                         }
 327                 }
 328
 329                 return $remaining;
 330         }
 331
 332         /**
 333          * @return integer Number of enqueued updates
 334          * @since 1.28
 335          */
 336         public static function pendingUpdatesCount() {
 337                 return count( self::$preSendUpdates ) + count( self::$postSendUpdates );
 338         }
 339
 340         /**
 341          * @param integer $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL)
 342          * @since 1.29
 343          */
 344         public static function getPendingUpdates( $stage = self::ALL ) {
 345                 $updates = [];
 346                 if ( $stage === self::ALL || $stage === self::PRESEND ) {
 347                         $updates = array_merge( $updates, self::$preSendUpdates );
 348                 }
 349                 if ( $stage === self::ALL || $stage === self::POSTSEND ) {
 350                         $updates = array_merge( $updates, self::$postSendUpdates );
 351                 }
 352                 return $updates;
 353         }
 354
 355         /**
 356          * Clear all pending updates without performing them. Generally, you don't
 357          * want or need to call this. Unit tests need it though.
 358          */
 359         public static function clearPendingUpdates() {
 360                 self::$preSendUpdates = [];
 361                 self::$postSendUpdates = [];
 362         }
 363
 364         /**
 365          * @return IDatabase[] Connection where commit() cannot be called yet
 366          */
 367         private static function getBusyDbConnections() {
 368                 $connsBusy = [];
 369
 370                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 371                 $lbFactory->forEachLB( function ( LoadBalancer $lb ) use ( &$connsBusy ) {
 372                         $lb->forEachOpenMasterConnection( function ( IDatabase $conn ) use ( &$connsBusy ) {
 373                                 if ( $conn->writesOrCallbacksPending() || $conn->explicitTrxActive() ) {
 374                                         $connsBusy[] = $conn;
 375                                 }
 376                         } );
 377                 } );
 378
 379                 return $connsBusy;
 380         }
 381 }