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