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 MediaWiki\MediaWikiServices;
  23 use Wikimedia\Rdbms\LBFactory;
  24 use Wikimedia\Rdbms\LoadBalancer;
  25
  26 /**
  27  * Class for managing the deferred updates
  28  *
  29  * In web request mode, deferred updates can be run at the end of the request, either before or
  30  * after the HTTP response has been sent. In either case, they run after the DB commit step. If
  31  * an update runs after the response is sent, it will not block clients. If sent before, it will
  32  * run synchronously. These two modes are defined via PRESEND and POSTSEND constants, the latter
  33  * being the default for addUpdate() and addCallableUpdate().
  34  *
  35  * Updates that work through this system will be more likely to complete by the time the client
  36  * makes their next request after this one than with the JobQueue system.
  37  *
  38  * In CLI mode, updates run immediately if no DB writes are pending. Otherwise, they run when:
  39  *   - a) Any waitForReplication() call if no writes are pending on any DB
  40  *   - b) A commit happens on Maintenance::getDB( DB_MASTER ) if no writes are pending on any DB
  41  *   - c) EnqueueableDataUpdate tasks may enqueue on commit of Maintenance::getDB( DB_MASTER )
  42  *   - d) At the completion of Maintenance::execute()
  43  *
  44  * When updates are deferred, they go into one two FIFO "top-queues" (one for pre-send and one
  45  * for post-send). Updates enqueued *during* doUpdate() of a "top" update go into the "sub-queue"
  46  * for that update. After that method finishes, the sub-queue is run until drained. This continues
  47  * for each top-queue job until the entire top queue is drained. This happens for the pre-send
  48  * top-queue, and later on, the post-send top-queue, in execute().
  49  *
  50  * @since 1.19
  51  */
  52 class DeferredUpdates {
  53         /** @var DeferrableUpdate[] Updates to be deferred until before request end */
  54         private static $preSendUpdates = [];
  55         /** @var DeferrableUpdate[] Updates to be deferred until after request end */
  56         private static $postSendUpdates = [];
  57         /** @var bool Whether to just run updates in addUpdate() */
  58         private static $immediateMode = false;
  59
  60         const ALL = 0; // all updates; in web requests, use only after flushing the output buffer
  61         const PRESEND = 1; // for updates that should run before flushing output buffer
  62         const POSTSEND = 2; // for updates that should run after flushing output buffer
  63
  64         const BIG_QUEUE_SIZE = 100;
  65
  66         /** @var array|null Information about the current execute() call or null if not running */
  67         private static $executeContext;
  68
  69         /**
  70          * Add an update to the deferred list to be run later by execute()
  71          *
  72          * In CLI mode, callback magic will also be used to run updates when safe
  73          *
  74          * @param DeferrableUpdate $update Some object that implements doUpdate()
  75          * @param integer $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
  76          */
  77         public static function addUpdate( DeferrableUpdate $update, $stage = self::POSTSEND ) {
  78                 global $wgCommandLineMode;
  79
  80                 // This is a sub-DeferredUpdate, run it right after its parent update
  81                 if ( self::$executeContext && self::$executeContext['stage'] >= $stage ) {
  82                         self::$executeContext['subqueue'][] = $update;
  83                         return;
  84                 }
  85
  86                 if ( $stage === self::PRESEND ) {
  87                         self::push( self::$preSendUpdates, $update );
  88                 } else {
  89                         self::push( self::$postSendUpdates, $update );
  90                 }
  91
  92                 if ( self::$immediateMode ) {
  93                         // No more explicit doUpdates() calls will happen, so run this now
  94                         self::doUpdates( 'run' );
  95                         return;
  96                 }
  97
  98                 // Try to run the updates now if in CLI mode and no transaction is active.
  99                 // This covers scripts that don't/barely use the DB but make updates to other stores.
 100                 if ( $wgCommandLineMode ) {
 101                         self::tryOpportunisticExecute( 'run' );
 102                 }
 103         }
 104
 105         /**
 106          * Add a callable update. In a lot of cases, we just need a callback/closure,
 107          * defining a new DeferrableUpdate object is not necessary
 108          *
 109          * @see MWCallableUpdate::__construct()
 110          *
 111          * @param callable $callable
 112          * @param integer $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
 113          * @param IDatabase|null $dbw Abort if this DB is rolled back [optional] (since 1.28)
 114          */
 115         public static function addCallableUpdate(
 116                 $callable, $stage = self::POSTSEND, IDatabase $dbw = null
 117         ) {
 118                 self::addUpdate( new MWCallableUpdate( $callable, wfGetCaller(), $dbw ), $stage );
 119         }
 120
 121         /**
 122          * Do any deferred updates and clear the list
 123          *
 124          * @param string $mode Use "enqueue" to use the job queue when possible [Default: "run"]
 125          * @param integer $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL) (since 1.27)
 126          */
 127         public static function doUpdates( $mode = 'run', $stage = self::ALL ) {
 128                 $stageEffective = ( $stage === self::ALL ) ? self::POSTSEND : $stage;
 129
 130                 if ( $stage === self::ALL || $stage === self::PRESEND ) {
 131                         self::execute( self::$preSendUpdates, $mode, $stageEffective );
 132                 }
 133
 134                 if ( $stage === self::ALL || $stage == self::POSTSEND ) {
 135                         self::execute( self::$postSendUpdates, $mode, $stageEffective );
 136                 }
 137         }
 138
 139         /**
 140          * @param bool $value Whether to just immediately run updates in addUpdate()
 141          * @since 1.28
 142          */
 143         public static function setImmediateMode( $value ) {
 144                 self::$immediateMode = (bool)$value;
 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 $existingUpdate MergeableUpdate */
 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 integer $stage Class constant (PRESEND, POSTSEND) (since 1.28)
 172          * @throws ErrorPageError Happens on top-level calls
 173          * @throws Exception Happens on second-level calls
 174          */
 175         protected static function execute( array &$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                         if ( $mode === 'enqueue' ) {
 193                                 try {
 194                                         // Push enqueuable updates to the job queue and get the rest
 195                                         $updates = self::enqueueUpdates( $updates );
 196                                 } catch ( Exception $e ) {
 197                                         // Let other updates have a chance to run if this failed
 198                                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 199                                 }
 200                         }
 201
 202                         // Order will be DataUpdate followed by generic DeferrableUpdate tasks
 203                         $updatesByType = [ 'data' => [], 'generic' => [] ];
 204                         foreach ( $updates as $du ) {
 205                                 if ( $du instanceof DataUpdate ) {
 206                                         $du->setTransactionTicket( $ticket );
 207                                         $updatesByType['data'][] = $du;
 208                                 } else {
 209                                         $updatesByType['generic'][] = $du;
 210                                 }
 211
 212                                 $name = ( $du instanceof DeferrableCallback )
 213                                         ? get_class( $du ) . '-' . $du->getOrigin()
 214                                         : get_class( $du );
 215                                 $stats->increment( 'deferred_updates.' . $method . '.' . $name );
 216                         }
 217
 218                         // Execute all remaining tasks...
 219                         foreach ( $updatesByType as $updatesForType ) {
 220                                 foreach ( $updatesForType as $update ) {
 221                                         self::$executeContext = [
 222                                                 'update' => $update,
 223                                                 'stage' => $stage,
 224                                                 'subqueue' => []
 225                                         ];
 226                                         /** @var DeferrableUpdate $update */
 227                                         $guiError = self::runUpdate( $update, $lbFactory, $stage );
 228                                         $reportableError = $reportableError ?: $guiError;
 229                                         // Do the subqueue updates for $update until there are none
 230                                         while ( self::$executeContext['subqueue'] ) {
 231                                                 $subUpdate = reset( self::$executeContext['subqueue'] );
 232                                                 $firstKey = key( self::$executeContext['subqueue'] );
 233                                                 unset( self::$executeContext['subqueue'][$firstKey] );
 234
 235                                                 if ( $subUpdate instanceof DataUpdate ) {
 236                                                         $subUpdate->setTransactionTicket( $ticket );
 237                                                 }
 238
 239                                                 $guiError = self::runUpdate( $subUpdate, $lbFactory, $stage );
 240                                                 $reportableError = $reportableError ?: $guiError;
 241                                         }
 242                                         self::$executeContext = null;
 243                                 }
 244                         }
 245
 246                         $updates = $queue; // new snapshot of queue (check for new entries)
 247                 }
 248
 249                 if ( $reportableError ) {
 250                         throw $reportableError; // throw the first of any GUI errors
 251                 }
 252         }
 253
 254         /**
 255          * @param DeferrableUpdate $update
 256          * @param LBFactory $lbFactory
 257          * @param integer $stage
 258          * @return ErrorPageError|null
 259          */
 260         private static function runUpdate( DeferrableUpdate $update, LBFactory $lbFactory, $stage ) {
 261                 $guiError = null;
 262                 try {
 263                         $fnameTrxOwner = get_class( $update ) . '::doUpdate';
 264                         $lbFactory->beginMasterChanges( $fnameTrxOwner );
 265                         $update->doUpdate();
 266                         $lbFactory->commitMasterChanges( $fnameTrxOwner );
 267                 } catch ( Exception $e ) {
 268                         // Reporting GUI exceptions does not work post-send
 269                         if ( $e instanceof ErrorPageError && $stage === self::PRESEND ) {
 270                                 $guiError = $e;
 271                         }
 272                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 273                 }
 274
 275                 return $guiError;
 276         }
 277
 278         /**
 279          * Run all deferred updates immediately if there are no DB writes active
 280          *
 281          * If $mode is 'run' but there are busy databates, EnqueueableDataUpdate
 282          * tasks will be enqueued anyway for the sake of progress.
 283          *
 284          * @param string $mode Use "enqueue" to use the job queue when possible
 285          * @return bool Whether updates were allowed to run
 286          * @since 1.28
 287          */
 288         public static function tryOpportunisticExecute( $mode = 'run' ) {
 289                 // execute() loop is already running
 290                 if ( self::$executeContext ) {
 291                         return false;
 292                 }
 293
 294                 // Avoiding running updates without them having outer scope
 295                 if ( !self::getBusyDbConnections() ) {
 296                         self::doUpdates( $mode );
 297                         return true;
 298                 }
 299
 300                 if ( self::pendingUpdatesCount() >= self::BIG_QUEUE_SIZE ) {
 301                         // If we cannot run the updates with outer transaction context, try to
 302                         // at least enqueue all the updates that support queueing to job queue
 303                         self::$preSendUpdates = self::enqueueUpdates( self::$preSendUpdates );
 304                         self::$postSendUpdates = self::enqueueUpdates( self::$postSendUpdates );
 305                 }
 306
 307                 return !self::pendingUpdatesCount();
 308         }
 309
 310         /**
 311          * Enqueue a job for each EnqueueableDataUpdate item and return the other items
 312          *
 313          * @param DeferrableUpdate[] $updates A list of deferred update instances
 314          * @return DeferrableUpdate[] Remaining updates that do not support being queued
 315          */
 316         private static function enqueueUpdates( array $updates ) {
 317                 $remaining = [];
 318
 319                 foreach ( $updates as $update ) {
 320                         if ( $update instanceof EnqueueableDataUpdate ) {
 321                                 $spec = $update->getAsJobSpecification();
 322                                 JobQueueGroup::singleton( $spec['wiki'] )->push( $spec['job'] );
 323                         } else {
 324                                 $remaining[] = $update;
 325                         }
 326                 }
 327
 328                 return $remaining;
 329         }
 330
 331         /**
 332          * @return integer Number of enqueued updates
 333          * @since 1.28
 334          */
 335         public static function pendingUpdatesCount() {
 336                 return count( self::$preSendUpdates ) + count( self::$postSendUpdates );
 337         }
 338
 339         /**
 340          * @param integer $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL)
 341          * @since 1.29
 342          */
 343         public static function getPendingUpdates( $stage = self::ALL ) {
 344                 $updates = [];
 345                 if ( $stage === self::ALL || $stage === self::PRESEND ) {
 346                         $updates = array_merge( $updates, self::$preSendUpdates );
 347                 }
 348                 if ( $stage === self::ALL || $stage === self::POSTSEND ) {
 349                         $updates = array_merge( $updates, self::$postSendUpdates );
 350                 }
 351                 return $updates;
 352         }
 353
 354         /**
 355          * Clear all pending updates without performing them. Generally, you don't
 356          * want or need to call this. Unit tests need it though.
 357          */
 358         public static function clearPendingUpdates() {
 359                 self::$preSendUpdates = [];
 360                 self::$postSendUpdates = [];
 361         }
 362
 363         /**
 364          * @return IDatabase[] Connection where commit() cannot be called yet
 365          */
 366         private static function getBusyDbConnections() {
 367                 $connsBusy = [];
 368
 369                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 370                 $lbFactory->forEachLB( function ( LoadBalancer $lb ) use ( &$connsBusy ) {
 371                         $lb->forEachOpenMasterConnection( function ( IDatabase $conn ) use ( &$connsBusy ) {
 372                                 if ( $conn->writesOrCallbacksPending() || $conn->explicitTrxActive() ) {
 373                                         $connsBusy[] = $conn;
 374                                 }
 375                         } );
 376                 } );
 377
 378                 return $connsBusy;
 379         }
 380 }