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
  59         const ALL = 0; // all updates; in web requests, use only after flushing the output buffer
  60         const PRESEND = 1; // for updates that should run before flushing output buffer
  61         const POSTSEND = 2; // for updates that should run after flushing output buffer
  62
  63         const BIG_QUEUE_SIZE = 100;
  64
  65         /** @var array|null Information about the current execute() call or null if not running */
  66         private static $executeContext;
  67
  68         /**
  69          * Add an update to the deferred list to be run later by execute()
  70          *
  71          * In CLI mode, callback magic will also be used to run updates when safe
  72          *
  73          * @param DeferrableUpdate $update Some object that implements doUpdate()
  74          * @param integer $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
  75          */
  76         public static function addUpdate( DeferrableUpdate $update, $stage = self::POSTSEND ) {
  77                 global $wgCommandLineMode;
  78
  79                 // This is a sub-DeferredUpdate, run it right after its parent update
  80                 if ( self::$executeContext && self::$executeContext['stage'] >= $stage ) {
  81                         self::$executeContext['subqueue'][] = $update;
  82                         return;
  83                 }
  84
  85                 if ( $stage === self::PRESEND ) {
  86                         self::push( self::$preSendUpdates, $update );
  87                 } else {
  88                         self::push( self::$postSendUpdates, $update );
  89                 }
  90
  91                 // Try to run the updates now if in CLI mode and no transaction is active.
  92                 // This covers scripts that don't/barely use the DB but make updates to other stores.
  93                 if ( $wgCommandLineMode ) {
  94                         self::tryOpportunisticExecute( 'run' );
  95                 }
  96         }
  97
  98         /**
  99          * Add a callable update. In a lot of cases, we just need a callback/closure,
 100          * defining a new DeferrableUpdate object is not necessary
 101          *
 102          * @see MWCallableUpdate::__construct()
 103          *
 104          * @param callable $callable
 105          * @param integer $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
 106          * @param IDatabase|null $dbw Abort if this DB is rolled back [optional] (since 1.28)
 107          */
 108         public static function addCallableUpdate(
 109                 $callable, $stage = self::POSTSEND, IDatabase $dbw = null
 110         ) {
 111                 self::addUpdate( new MWCallableUpdate( $callable, wfGetCaller(), $dbw ), $stage );
 112         }
 113
 114         /**
 115          * Do any deferred updates and clear the list
 116          *
 117          * @param string $mode Use "enqueue" to use the job queue when possible [Default: "run"]
 118          * @param integer $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL) (since 1.27)
 119          */
 120         public static function doUpdates( $mode = 'run', $stage = self::ALL ) {
 121                 $stageEffective = ( $stage === self::ALL ) ? self::POSTSEND : $stage;
 122
 123                 if ( $stage === self::ALL || $stage === self::PRESEND ) {
 124                         self::execute( self::$preSendUpdates, $mode, $stageEffective );
 125                 }
 126
 127                 if ( $stage === self::ALL || $stage == self::POSTSEND ) {
 128                         self::execute( self::$postSendUpdates, $mode, $stageEffective );
 129                 }
 130         }
 131
 132         /**
 133          * @param bool $value Whether to just immediately run updates in addUpdate()
 134          * @since 1.28
 135          * @deprecated 1.29 Causes issues in Web-executed jobs - see T165714 and T100085.
 136          */
 137         public static function setImmediateMode( $value ) {
 138                 wfDeprecated( __METHOD__, '1.29' );
 139         }
 140
 141         /**
 142          * @param DeferrableUpdate[] $queue
 143          * @param DeferrableUpdate $update
 144          */
 145         private static function push( array &$queue, DeferrableUpdate $update ) {
 146                 if ( $update instanceof MergeableUpdate ) {
 147                         $class = get_class( $update ); // fully-qualified class
 148                         if ( isset( $queue[$class] ) ) {
 149                                 /** @var $existingUpdate MergeableUpdate */
 150                                 $existingUpdate = $queue[$class];
 151                                 $existingUpdate->merge( $update );
 152                         } else {
 153                                 $queue[$class] = $update;
 154                         }
 155                 } else {
 156                         $queue[] = $update;
 157                 }
 158         }
 159
 160         /**
 161          * Immediately run/queue a list of updates
 162          *
 163          * @param DeferrableUpdate[] &$queue List of DeferrableUpdate objects
 164          * @param string $mode Use "enqueue" to use the job queue when possible
 165          * @param integer $stage Class constant (PRESEND, POSTSEND) (since 1.28)
 166          * @throws ErrorPageError Happens on top-level calls
 167          * @throws Exception Happens on second-level calls
 168          */
 169         protected static function execute( array &$queue, $mode, $stage ) {
 170                 $services = MediaWikiServices::getInstance();
 171                 $stats = $services->getStatsdDataFactory();
 172                 $lbFactory = $services->getDBLoadBalancerFactory();
 173                 $method = RequestContext::getMain()->getRequest()->getMethod();
 174
 175                 $ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
 176
 177                 /** @var ErrorPageError $reportableError */
 178                 $reportableError = null;
 179                 /** @var DeferrableUpdate[] $updates Snapshot of queue */
 180                 $updates = $queue;
 181
 182                 // Keep doing rounds of updates until none get enqueued...
 183                 while ( $updates ) {
 184                         $queue = []; // clear the queue
 185
 186                         if ( $mode === 'enqueue' ) {
 187                                 try {
 188                                         // Push enqueuable updates to the job queue and get the rest
 189                                         $updates = self::enqueueUpdates( $updates );
 190                                 } catch ( Exception $e ) {
 191                                         // Let other updates have a chance to run if this failed
 192                                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 193                                 }
 194                         }
 195
 196                         // Order will be DataUpdate followed by generic DeferrableUpdate tasks
 197                         $updatesByType = [ 'data' => [], 'generic' => [] ];
 198                         foreach ( $updates as $du ) {
 199                                 if ( $du instanceof DataUpdate ) {
 200                                         $du->setTransactionTicket( $ticket );
 201                                         $updatesByType['data'][] = $du;
 202                                 } else {
 203                                         $updatesByType['generic'][] = $du;
 204                                 }
 205
 206                                 $name = ( $du instanceof DeferrableCallback )
 207                                         ? get_class( $du ) . '-' . $du->getOrigin()
 208                                         : get_class( $du );
 209                                 $stats->increment( 'deferred_updates.' . $method . '.' . $name );
 210                         }
 211
 212                         // Execute all remaining tasks...
 213                         foreach ( $updatesByType as $updatesForType ) {
 214                                 foreach ( $updatesForType as $update ) {
 215                                         self::$executeContext = [
 216                                                 'update' => $update,
 217                                                 'stage' => $stage,
 218                                                 'subqueue' => []
 219                                         ];
 220                                         /** @var DeferrableUpdate $update */
 221                                         $guiError = self::runUpdate( $update, $lbFactory, $stage );
 222                                         $reportableError = $reportableError ?: $guiError;
 223                                         // Do the subqueue updates for $update until there are none
 224                                         while ( self::$executeContext['subqueue'] ) {
 225                                                 $subUpdate = reset( self::$executeContext['subqueue'] );
 226                                                 $firstKey = key( self::$executeContext['subqueue'] );
 227                                                 unset( self::$executeContext['subqueue'][$firstKey] );
 228
 229                                                 if ( $subUpdate instanceof DataUpdate ) {
 230                                                         $subUpdate->setTransactionTicket( $ticket );
 231                                                 }
 232
 233                                                 $guiError = self::runUpdate( $subUpdate, $lbFactory, $stage );
 234                                                 $reportableError = $reportableError ?: $guiError;
 235                                         }
 236                                         self::$executeContext = null;
 237                                 }
 238                         }
 239
 240                         $updates = $queue; // new snapshot of queue (check for new entries)
 241                 }
 242
 243                 if ( $reportableError ) {
 244                         throw $reportableError; // throw the first of any GUI errors
 245                 }
 246         }
 247
 248         /**
 249          * @param DeferrableUpdate $update
 250          * @param LBFactory $lbFactory
 251          * @param integer $stage
 252          * @return ErrorPageError|null
 253          */
 254         private static function runUpdate( DeferrableUpdate $update, LBFactory $lbFactory, $stage ) {
 255                 $guiError = null;
 256                 try {
 257                         $fnameTrxOwner = get_class( $update ) . '::doUpdate';
 258                         $lbFactory->beginMasterChanges( $fnameTrxOwner );
 259                         $update->doUpdate();
 260                         $lbFactory->commitMasterChanges( $fnameTrxOwner );
 261                 } catch ( Exception $e ) {
 262                         // Reporting GUI exceptions does not work post-send
 263                         if ( $e instanceof ErrorPageError && $stage === self::PRESEND ) {
 264                                 $guiError = $e;
 265                         }
 266                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 267                 }
 268
 269                 return $guiError;
 270         }
 271
 272         /**
 273          * Run all deferred updates immediately if there are no DB writes active
 274          *
 275          * If $mode is 'run' but there are busy databates, EnqueueableDataUpdate
 276          * tasks will be enqueued anyway for the sake of progress.
 277          *
 278          * @param string $mode Use "enqueue" to use the job queue when possible
 279          * @return bool Whether updates were allowed to run
 280          * @since 1.28
 281          */
 282         public static function tryOpportunisticExecute( $mode = 'run' ) {
 283                 // execute() loop is already running
 284                 if ( self::$executeContext ) {
 285                         return false;
 286                 }
 287
 288                 // Avoiding running updates without them having outer scope
 289                 if ( !self::areDatabaseTransactionsActive() ) {
 290                         self::doUpdates( $mode );
 291                         return true;
 292                 }
 293
 294                 if ( self::pendingUpdatesCount() >= self::BIG_QUEUE_SIZE ) {
 295                         // If we cannot run the updates with outer transaction context, try to
 296                         // at least enqueue all the updates that support queueing to job queue
 297                         self::$preSendUpdates = self::enqueueUpdates( self::$preSendUpdates );
 298                         self::$postSendUpdates = self::enqueueUpdates( self::$postSendUpdates );
 299                 }
 300
 301                 return !self::pendingUpdatesCount();
 302         }
 303
 304         /**
 305          * Enqueue a job for each EnqueueableDataUpdate item and return the other items
 306          *
 307          * @param DeferrableUpdate[] $updates A list of deferred update instances
 308          * @return DeferrableUpdate[] Remaining updates that do not support being queued
 309          */
 310         private static function enqueueUpdates( array $updates ) {
 311                 $remaining = [];
 312
 313                 foreach ( $updates as $update ) {
 314                         if ( $update instanceof EnqueueableDataUpdate ) {
 315                                 $spec = $update->getAsJobSpecification();
 316                                 JobQueueGroup::singleton( $spec['wiki'] )->push( $spec['job'] );
 317                         } else {
 318                                 $remaining[] = $update;
 319                         }
 320                 }
 321
 322                 return $remaining;
 323         }
 324
 325         /**
 326          * @return integer Number of enqueued updates
 327          * @since 1.28
 328          */
 329         public static function pendingUpdatesCount() {
 330                 return count( self::$preSendUpdates ) + count( self::$postSendUpdates );
 331         }
 332
 333         /**
 334          * @param integer $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL)
 335          * @return DeferrableUpdate[]
 336          * @since 1.29
 337          */
 338         public static function getPendingUpdates( $stage = self::ALL ) {
 339                 $updates = [];
 340                 if ( $stage === self::ALL || $stage === self::PRESEND ) {
 341                         $updates = array_merge( $updates, self::$preSendUpdates );
 342                 }
 343                 if ( $stage === self::ALL || $stage === self::POSTSEND ) {
 344                         $updates = array_merge( $updates, self::$postSendUpdates );
 345                 }
 346                 return $updates;
 347         }
 348
 349         /**
 350          * Clear all pending updates without performing them. Generally, you don't
 351          * want or need to call this. Unit tests need it though.
 352          */
 353         public static function clearPendingUpdates() {
 354                 self::$preSendUpdates = [];
 355                 self::$postSendUpdates = [];
 356         }
 357
 358         /**
 359          * @return bool If a transaction round is active or connection is not ready for commit()
 360          */
 361         private static function areDatabaseTransactionsActive() {
 362                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 363                 if ( $lbFactory->hasTransactionRound() ) {
 364                         return true;
 365                 }
 366
 367                 $connsBusy = false;
 368                 $lbFactory->forEachLB( function ( LoadBalancer $lb ) use ( &$connsBusy ) {
 369                         $lb->forEachOpenMasterConnection( function ( IDatabase $conn ) use ( &$connsBusy ) {
 370                                 if ( $conn->writesOrCallbacksPending() || $conn->explicitTrxActive() ) {
 371                                         $connsBusy = true;
 372                                 }
 373                         } );
 374                 } );
 375
 376                 return $connsBusy;
 377         }
 378 }