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
  24 /**
  25  * Class for managing the deferred updates
  26  *
  27  * In web request mode, deferred updates can be run at the end of the request, either before or
  28  * after the HTTP response has been sent. In either case, they run after the DB commit step. If
  29  * an update runs after the response is sent, it will not block clients. If sent before, it will
  30  * run synchronously. These two modes are defined via PRESEND and POSTSEND constants, the latter
  31  * being the default for addUpdate() and addCallableUpdate().
  32  *
  33  * Updates that work through this system will be more likely to complete by the time the client
  34  * makes their next request after this one than with the JobQueue system.
  35  *
  36  * In CLI mode, updates run immediately if no DB writes are pending. Otherwise, they run when:
  37  *   - a) Any waitForReplication() call if no writes are pending on any DB
  38  *   - b) A commit happens on Maintenance::getDB( DB_MASTER ) if no writes are pending on any DB
  39  *   - c) EnqueueableDataUpdate tasks may enqueue on commit of Maintenance::getDB( DB_MASTER )
  40  *   - d) At the completion of Maintenance::execute()
  41  *
  42  * When updates are deferred, they go into one two FIFO "top-queues" (one for pre-send and one
  43  * for post-send). Updates enqueued *during* doUpdate() of a "top" update go into the "sub-queue"
  44  * for that update. After that method finishes, the sub-queue is run until drained. This continues
  45  * for each top-queue job until the entire top queue is drained. This happens for the pre-send
  46  * top-queue, and later on, the post-send top-queue, in execute().
  47  *
  48  * @since 1.19
  49  */
  50 class DeferredUpdates {
  51         /** @var DeferrableUpdate[] Updates to be deferred until before request end */
  52         private static $preSendUpdates = [];
  53         /** @var DeferrableUpdate[] Updates to be deferred until after request end */
  54         private static $postSendUpdates = [];
  55
  56         const ALL = 0; // all updates; in web requests, use only after flushing the output buffer
  57         const PRESEND = 1; // for updates that should run before flushing output buffer
  58         const POSTSEND = 2; // for updates that should run after flushing output buffer
  59
  60         const BIG_QUEUE_SIZE = 100;
  61
  62         /** @var array|null Information about the current execute() call or null if not running */
  63         private static $executeContext;
  64
  65         /**
  66          * Add an update to the deferred list to be run later by execute()
  67          *
  68          * In CLI mode, callback magic will also be used to run updates when safe
  69          *
  70          * @param DeferrableUpdate $update Some object that implements doUpdate()
  71          * @param integer $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
  72          */
  73         public static function addUpdate( DeferrableUpdate $update, $stage = self::POSTSEND ) {
  74                 global $wgCommandLineMode;
  75
  76                 // This is a sub-DeferredUpdate, run it right after its parent update
  77                 if ( self::$executeContext && self::$executeContext['stage'] >= $stage ) {
  78                         self::$executeContext['subqueue'][] = $update;
  79                         return;
  80                 }
  81
  82                 if ( $stage === self::PRESEND ) {
  83                         self::push( self::$preSendUpdates, $update );
  84                 } else {
  85                         self::push( self::$postSendUpdates, $update );
  86                 }
  87
  88                 // Try to run the updates now if in CLI mode and no transaction is active.
  89                 // This covers scripts that don't/barely use the DB but make updates to other stores.
  90                 if ( $wgCommandLineMode ) {
  91                         self::tryOpportunisticExecute( 'run' );
  92                 }
  93         }
  94
  95         /**
  96          * Add a callable update. In a lot of cases, we just need a callback/closure,
  97          * defining a new DeferrableUpdate object is not necessary
  98          *
  99          * @see MWCallableUpdate::__construct()
 100          *
 101          * @param callable $callable
 102          * @param integer $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
 103          * @param IDatabase|null $dbw Abort if this DB is rolled back [optional] (since 1.28)
 104          */
 105         public static function addCallableUpdate(
 106                 $callable, $stage = self::POSTSEND, IDatabase $dbw = null
 107         ) {
 108                 self::addUpdate( new MWCallableUpdate( $callable, wfGetCaller(), $dbw ), $stage );
 109         }
 110
 111         /**
 112          * Do any deferred updates and clear the list
 113          *
 114          * @param string $mode Use "enqueue" to use the job queue when possible [Default: "run"]
 115          * @param integer $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL) (since 1.27)
 116          */
 117         public static function doUpdates( $mode = 'run', $stage = self::ALL ) {
 118                 $stageEffective = ( $stage === self::ALL ) ? self::POSTSEND : $stage;
 119
 120                 if ( $stage === self::ALL || $stage === self::PRESEND ) {
 121                         self::execute( self::$preSendUpdates, $mode, $stageEffective );
 122                 }
 123
 124                 if ( $stage === self::ALL || $stage == self::POSTSEND ) {
 125                         self::execute( self::$postSendUpdates, $mode, $stageEffective );
 126                 }
 127         }
 128
 129         /**
 130          * @param DeferrableUpdate[] $queue
 131          * @param DeferrableUpdate $update
 132          */
 133         private static function push( array &$queue, DeferrableUpdate $update ) {
 134                 if ( $update instanceof MergeableUpdate ) {
 135                         $class = get_class( $update ); // fully-qualified class
 136                         if ( isset( $queue[$class] ) ) {
 137                                 /** @var $existingUpdate MergeableUpdate */
 138                                 $existingUpdate = $queue[$class];
 139                                 $existingUpdate->merge( $update );
 140                         } else {
 141                                 $queue[$class] = $update;
 142                         }
 143                 } else {
 144                         $queue[] = $update;
 145                 }
 146         }
 147
 148         /**
 149          * Immediately run/queue a list of updates
 150          *
 151          * @param DeferrableUpdate[] &$queue List of DeferrableUpdate objects
 152          * @param string $mode Use "enqueue" to use the job queue when possible
 153          * @param integer $stage Class constant (PRESEND, POSTSEND) (since 1.28)
 154          * @throws ErrorPageError Happens on top-level calls
 155          * @throws Exception Happens on second-level calls
 156          */
 157         protected static function execute( array &$queue, $mode, $stage ) {
 158                 $services = MediaWikiServices::getInstance();
 159                 $stats = $services->getStatsdDataFactory();
 160                 $lbFactory = $services->getDBLoadBalancerFactory();
 161                 $method = RequestContext::getMain()->getRequest()->getMethod();
 162
 163                 /** @var ErrorPageError $reportableError */
 164                 $reportableError = null;
 165                 /** @var DeferrableUpdate[] $updates Snapshot of queue */
 166                 $updates = $queue;
 167
 168                 // Keep doing rounds of updates until none get enqueued...
 169                 while ( $updates ) {
 170                         $queue = []; // clear the queue
 171
 172                         if ( $mode === 'enqueue' ) {
 173                                 try {
 174                                         // Push enqueuable updates to the job queue and get the rest
 175                                         $updates = self::enqueueUpdates( $updates );
 176                                 } catch ( Exception $e ) {
 177                                         // Let other updates have a chance to run if this failed
 178                                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 179                                 }
 180                         }
 181
 182                         // Order will be DataUpdate followed by generic DeferrableUpdate tasks
 183                         $updatesByType = [ 'data' => [], 'generic' => [] ];
 184                         foreach ( $updates as $du ) {
 185                                 $updatesByType[$du instanceof DataUpdate ? 'data' : 'generic'][] = $du;
 186                                 $name = ( $du instanceof DeferrableCallback )
 187                                         ? get_class( $du ) . '-' . $du->getOrigin()
 188                                         : get_class( $du );
 189                                 $stats->increment( 'deferred_updates.' . $method . '.' . $name );
 190                         }
 191
 192                         // Execute all remaining tasks...
 193                         foreach ( $updatesByType as $updatesForType ) {
 194                                 foreach ( $updatesForType as $update ) {
 195                                         self::$executeContext = [
 196                                                 'update' => $update,
 197                                                 'stage' => $stage,
 198                                                 'subqueue' => []
 199                                         ];
 200                                         /** @var DeferrableUpdate $update */
 201                                         $guiError = self::runUpdate( $update, $lbFactory, $stage );
 202                                         $reportableError = $reportableError ?: $guiError;
 203                                         // Do the subqueue updates for $update until there are none
 204                                         while ( self::$executeContext['subqueue'] ) {
 205                                                 $subUpdate = reset( self::$executeContext['subqueue'] );
 206                                                 $firstKey = key( self::$executeContext['subqueue'] );
 207                                                 unset( self::$executeContext['subqueue'][$firstKey] );
 208
 209                                                 $guiError = self::runUpdate( $subUpdate, $lbFactory, $stage );
 210                                                 $reportableError = $reportableError ?: $guiError;
 211                                         }
 212                                         self::$executeContext = null;
 213                                 }
 214                         }
 215
 216                         $updates = $queue; // new snapshot of queue (check for new entries)
 217                 }
 218
 219                 if ( $reportableError ) {
 220                         throw $reportableError; // throw the first of any GUI errors
 221                 }
 222         }
 223
 224         /**
 225          * @param DeferrableUpdate $update
 226          * @param LBFactory $lbFactory
 227          * @param integer $stage
 228          * @return ErrorPageError|null
 229          */
 230         private static function runUpdate( DeferrableUpdate $update, LBFactory $lbFactory, $stage ) {
 231                 $guiError = null;
 232                 try {
 233                         $fnameTrxOwner = get_class( $update ) . '::doUpdate';
 234                         $lbFactory->beginMasterChanges( $fnameTrxOwner );
 235                         $update->doUpdate();
 236                         $lbFactory->commitMasterChanges( $fnameTrxOwner );
 237                 } catch ( Exception $e ) {
 238                         // Reporting GUI exceptions does not work post-send
 239                         if ( $e instanceof ErrorPageError && $stage === self::PRESEND ) {
 240                                 $guiError = $e;
 241                         }
 242                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 243                 }
 244
 245                 return $guiError;
 246         }
 247
 248         /**
 249          * Run all deferred updates immediately if there are no DB writes active
 250          *
 251          * If $mode is 'run' but there are busy databates, EnqueueableDataUpdate
 252          * tasks will be enqueued anyway for the sake of progress.
 253          *
 254          * @param string $mode Use "enqueue" to use the job queue when possible
 255          * @return bool Whether updates were allowed to run
 256          * @since 1.28
 257          */
 258         public static function tryOpportunisticExecute( $mode = 'run' ) {
 259                 // execute() loop is already running
 260                 if ( self::$executeContext ) {
 261                         return false;
 262                 }
 263
 264                 // Avoiding running updates without them having outer scope
 265                 if ( !self::getBusyDbConnections() ) {
 266                         self::doUpdates( $mode );
 267                         return true;
 268                 }
 269
 270                 if ( self::pendingUpdatesCount() >= self::BIG_QUEUE_SIZE ) {
 271                         // If we cannot run the updates with outer transaction context, try to
 272                         // at least enqueue all the updates that support queueing to job queue
 273                         self::$preSendUpdates = self::enqueueUpdates( self::$preSendUpdates );
 274                         self::$postSendUpdates = self::enqueueUpdates( self::$postSendUpdates );
 275                 }
 276
 277                 return !self::pendingUpdatesCount();
 278         }
 279
 280         /**
 281          * Enqueue a job for each EnqueueableDataUpdate item and return the other items
 282          *
 283          * @param DeferrableUpdate[] $updates A list of deferred update instances
 284          * @return DeferrableUpdate[] Remaining updates that do not support being queued
 285          */
 286         private static function enqueueUpdates( array $updates ) {
 287                 $remaining = [];
 288
 289                 foreach ( $updates as $update ) {
 290                         if ( $update instanceof EnqueueableDataUpdate ) {
 291                                 $spec = $update->getAsJobSpecification();
 292                                 JobQueueGroup::singleton( $spec['wiki'] )->push( $spec['job'] );
 293                         } else {
 294                                 $remaining[] = $update;
 295                         }
 296                 }
 297
 298                 return $remaining;
 299         }
 300
 301         /**
 302          * @return integer Number of enqueued updates
 303          * @since 1.28
 304          */
 305         public static function pendingUpdatesCount() {
 306                 return count( self::$preSendUpdates ) + count( self::$postSendUpdates );
 307         }
 308
 309         /**
 310          * Clear all pending updates without performing them. Generally, you don't
 311          * want or need to call this. Unit tests need it though.
 312          */
 313         public static function clearPendingUpdates() {
 314                 self::$preSendUpdates = [];
 315                 self::$postSendUpdates = [];
 316         }
 317
 318         /**
 319          * @return IDatabase[] Connection where commit() cannot be called yet
 320          */
 321         private static function getBusyDbConnections() {
 322                 $connsBusy = [];
 323
 324                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 325                 $lbFactory->forEachLB( function ( LoadBalancer $lb ) use ( &$connsBusy ) {
 326                         $lb->forEachOpenMasterConnection( function ( IDatabase $conn ) use ( &$connsBusy ) {
 327                                 if ( $conn->writesOrCallbacksPending() || $conn->explicitTrxActive() ) {
 328                                         $connsBusy[] = $conn;
 329                                 }
 330                         } );
 331                 } );
 332
 333                 return $connsBusy;
 334         }
 335 }