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
  25 /**
  26  * Class for managing the deferred updates
  27  *
  28  * In web request mode, deferred updates can be run at the end of the request, either before or
  29  * after the HTTP response has been sent. In either case, they run after the DB commit step. If
  30  * an update runs after the response is sent, it will not block clients. If sent before, it will
  31  * run synchronously. These two modes are defined via PRESEND and POSTSEND constants, the latter
  32  * being the default for addUpdate() and addCallableUpdate().
  33  *
  34  * Updates that work through this system will be more likely to complete by the time the client
  35  * makes their next request after this one than with the JobQueue system.
  36  *
  37  * In CLI mode, updates run immediately if no DB writes are pending. Otherwise, they run when:
  38  *   - a) Any waitForReplication() call if no writes are pending on any DB
  39  *   - b) A commit happens on Maintenance::getDB( DB_MASTER ) if no writes are pending on any DB
  40  *   - c) EnqueueableDataUpdate tasks may enqueue on commit of Maintenance::getDB( DB_MASTER )
  41  *   - d) At the completion of Maintenance::execute()
  42  *
  43  * When updates are deferred, they go into one two FIFO "top-queues" (one for pre-send and one
  44  * for post-send). Updates enqueued *during* doUpdate() of a "top" update go into the "sub-queue"
  45  * for that update. After that method finishes, the sub-queue is run until drained. This continues
  46  * for each top-queue job until the entire top queue is drained. This happens for the pre-send
  47  * top-queue, and later on, the post-send top-queue, in execute().
  48  *
  49  * @since 1.19
  50  */
  51 class DeferredUpdates {
  52         /** @var DeferrableUpdate[] Updates to be deferred until before request end */
  53         private static $preSendUpdates = [];
  54         /** @var DeferrableUpdate[] Updates to be deferred until after request end */
  55         private static $postSendUpdates = [];
  56         /** @var bool Whether to just run updates in addUpdate() */
  57         private static $immediateMode = false;
  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, $update );
  89                 }
  90
  91                 if ( self::$immediateMode ) {
  92                         // No more explicit doUpdates() calls will happen, so run this now
  93                         self::doUpdates( 'run' );
  94                         return;
  95                 }
  96
  97                 // Try to run the updates now if in CLI mode and no transaction is active.
  98                 // This covers scripts that don't/barely use the DB but make updates to other stores.
  99                 if ( $wgCommandLineMode ) {
 100                         self::tryOpportunisticExecute( 'run' );
 101                 }
 102         }
 103
 104         /**
 105          * Add a callable update. In a lot of cases, we just need a callback/closure,
 106          * defining a new DeferrableUpdate object is not necessary
 107          *
 108          * @see MWCallableUpdate::__construct()
 109          *
 110          * @param callable $callable
 111          * @param integer $stage DeferredUpdates constant (PRESEND or POSTSEND) (since 1.27)
 112          * @param IDatabase|null $dbw Abort if this DB is rolled back [optional] (since 1.28)
 113          */
 114         public static function addCallableUpdate(
 115                 $callable, $stage = self::POSTSEND, IDatabase $dbw = null
 116         ) {
 117                 self::addUpdate( new MWCallableUpdate( $callable, wfGetCaller(), $dbw ), $stage );
 118         }
 119
 120         /**
 121          * Do any deferred updates and clear the list
 122          *
 123          * @param string $mode Use "enqueue" to use the job queue when possible [Default: "run"]
 124          * @param integer $stage DeferredUpdates constant (PRESEND, POSTSEND, or ALL) (since 1.27)
 125          */
 126         public static function doUpdates( $mode = 'run', $stage = self::ALL ) {
 127                 $stageEffective = ( $stage === self::ALL ) ? self::POSTSEND : $stage;
 128
 129                 if ( $stage === self::ALL || $stage === self::PRESEND ) {
 130                         self::execute( self::$preSendUpdates, $mode, $stageEffective );
 131                 }
 132
 133                 if ( $stage === self::ALL || $stage == self::POSTSEND ) {
 134                         self::execute( self::$postSendUpdates, $mode, $stageEffective );
 135                 }
 136         }
 137
 138         /**
 139          * @param bool $value Whether to just immediately run updates in addUpdate()
 140          * @since 1.28
 141          */
 142         public static function setImmediateMode( $value ) {
 143                 self::$immediateMode = (bool)$value;
 144         }
 145
 146         /**
 147          * @param DeferrableUpdate[] $queue
 148          * @param DeferrableUpdate $update
 149          */
 150         private static function push( array &$queue, DeferrableUpdate $update ) {
 151                 if ( $update instanceof MergeableUpdate ) {
 152                         $class = get_class( $update ); // fully-qualified class
 153                         if ( isset( $queue[$class] ) ) {
 154                                 /** @var $existingUpdate MergeableUpdate */
 155                                 $existingUpdate = $queue[$class];
 156                                 $existingUpdate->merge( $update );
 157                         } else {
 158                                 $queue[$class] = $update;
 159                         }
 160                 } else {
 161                         $queue[] = $update;
 162                 }
 163         }
 164
 165         /**
 166          * Immediately run/queue a list of updates
 167          *
 168          * @param DeferrableUpdate[] &$queue List of DeferrableUpdate objects
 169          * @param string $mode Use "enqueue" to use the job queue when possible
 170          * @param integer $stage Class constant (PRESEND, POSTSEND) (since 1.28)
 171          * @throws ErrorPageError Happens on top-level calls
 172          * @throws Exception Happens on second-level calls
 173          */
 174         protected static function execute( array &$queue, $mode, $stage ) {
 175                 $services = MediaWikiServices::getInstance();
 176                 $stats = $services->getStatsdDataFactory();
 177                 $lbFactory = $services->getDBLoadBalancerFactory();
 178                 $method = RequestContext::getMain()->getRequest()->getMethod();
 179
 180                 $ticket = $lbFactory->getEmptyTransactionTicket( __METHOD__ );
 181
 182                 /** @var ErrorPageError $reportableError */
 183                 $reportableError = null;
 184                 /** @var DeferrableUpdate[] $updates Snapshot of queue */
 185                 $updates = $queue;
 186
 187                 // Keep doing rounds of updates until none get enqueued...
 188                 while ( $updates ) {
 189                         $queue = []; // clear the queue
 190
 191                         if ( $mode === 'enqueue' ) {
 192                                 try {
 193                                         // Push enqueuable updates to the job queue and get the rest
 194                                         $updates = self::enqueueUpdates( $updates );
 195                                 } catch ( Exception $e ) {
 196                                         // Let other updates have a chance to run if this failed
 197                                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 198                                 }
 199                         }
 200
 201                         // Order will be DataUpdate followed by generic DeferrableUpdate tasks
 202                         $updatesByType = [ 'data' => [], 'generic' => [] ];
 203                         foreach ( $updates as $du ) {
 204                                 if ( $du instanceof DataUpdate ) {
 205                                         $du->setTransactionTicket( $ticket );
 206                                         $updatesByType['data'][] = $du;
 207                                 } else {
 208                                         $updatesByType['generic'][] = $du;
 209                                 }
 210
 211                                 $name = ( $du instanceof DeferrableCallback )
 212                                         ? get_class( $du ) . '-' . $du->getOrigin()
 213                                         : get_class( $du );
 214                                 $stats->increment( 'deferred_updates.' . $method . '.' . $name );
 215                         }
 216
 217                         // Execute all remaining tasks...
 218                         foreach ( $updatesByType as $updatesForType ) {
 219                                 foreach ( $updatesForType as $update ) {
 220                                         self::$executeContext = [
 221                                                 'update' => $update,
 222                                                 'stage' => $stage,
 223                                                 'subqueue' => []
 224                                         ];
 225                                         /** @var DeferrableUpdate $update */
 226                                         $guiError = self::runUpdate( $update, $lbFactory, $stage );
 227                                         $reportableError = $reportableError ?: $guiError;
 228                                         // Do the subqueue updates for $update until there are none
 229                                         while ( self::$executeContext['subqueue'] ) {
 230                                                 $subUpdate = reset( self::$executeContext['subqueue'] );
 231                                                 $firstKey = key( self::$executeContext['subqueue'] );
 232                                                 unset( self::$executeContext['subqueue'][$firstKey] );
 233
 234                                                 if ( $subUpdate instanceof DataUpdate ) {
 235                                                         $subUpdate->setTransactionTicket( $ticket );
 236                                                 }
 237
 238                                                 $guiError = self::runUpdate( $subUpdate, $lbFactory, $stage );
 239                                                 $reportableError = $reportableError ?: $guiError;
 240                                         }
 241                                         self::$executeContext = null;
 242                                 }
 243                         }
 244
 245                         $updates = $queue; // new snapshot of queue (check for new entries)
 246                 }
 247
 248                 if ( $reportableError ) {
 249                         throw $reportableError; // throw the first of any GUI errors
 250                 }
 251         }
 252
 253         /**
 254          * @param DeferrableUpdate $update
 255          * @param LBFactory $lbFactory
 256          * @param integer $stage
 257          * @return ErrorPageError|null
 258          */
 259         private static function runUpdate( DeferrableUpdate $update, LBFactory $lbFactory, $stage ) {
 260                 $guiError = null;
 261                 try {
 262                         $fnameTrxOwner = get_class( $update ) . '::doUpdate';
 263                         $lbFactory->beginMasterChanges( $fnameTrxOwner );
 264                         $update->doUpdate();
 265                         $lbFactory->commitMasterChanges( $fnameTrxOwner );
 266                 } catch ( Exception $e ) {
 267                         // Reporting GUI exceptions does not work post-send
 268                         if ( $e instanceof ErrorPageError && $stage === self::PRESEND ) {
 269                                 $guiError = $e;
 270                         }
 271                         MWExceptionHandler::rollbackMasterChangesAndLog( $e );
 272                 }
 273
 274                 return $guiError;
 275         }
 276
 277         /**
 278          * Run all deferred updates immediately if there are no DB writes active
 279          *
 280          * If $mode is 'run' but there are busy databates, EnqueueableDataUpdate
 281          * tasks will be enqueued anyway for the sake of progress.
 282          *
 283          * @param string $mode Use "enqueue" to use the job queue when possible
 284          * @return bool Whether updates were allowed to run
 285          * @since 1.28
 286          */
 287         public static function tryOpportunisticExecute( $mode = 'run' ) {
 288                 // execute() loop is already running
 289                 if ( self::$executeContext ) {
 290                         return false;
 291                 }
 292
 293                 // Avoiding running updates without them having outer scope
 294                 if ( !self::getBusyDbConnections() ) {
 295                         self::doUpdates( $mode );
 296                         return true;
 297                 }
 298
 299                 if ( self::pendingUpdatesCount() >= self::BIG_QUEUE_SIZE ) {
 300                         // If we cannot run the updates with outer transaction context, try to
 301                         // at least enqueue all the updates that support queueing to job queue
 302                         self::$preSendUpdates = self::enqueueUpdates( self::$preSendUpdates );
 303                         self::$postSendUpdates = self::enqueueUpdates( self::$postSendUpdates );
 304                 }
 305
 306                 return !self::pendingUpdatesCount();
 307         }
 308
 309         /**
 310          * Enqueue a job for each EnqueueableDataUpdate item and return the other items
 311          *
 312          * @param DeferrableUpdate[] $updates A list of deferred update instances
 313          * @return DeferrableUpdate[] Remaining updates that do not support being queued
 314          */
 315         private static function enqueueUpdates( array $updates ) {
 316                 $remaining = [];
 317
 318                 foreach ( $updates as $update ) {
 319                         if ( $update instanceof EnqueueableDataUpdate ) {
 320                                 $spec = $update->getAsJobSpecification();
 321                                 JobQueueGroup::singleton( $spec['wiki'] )->push( $spec['job'] );
 322                         } else {
 323                                 $remaining[] = $update;
 324                         }
 325                 }
 326
 327                 return $remaining;
 328         }
 329
 330         /**
 331          * @return integer Number of enqueued updates
 332          * @since 1.28
 333          */
 334         public static function pendingUpdatesCount() {
 335                 return count( self::$preSendUpdates ) + count( self::$postSendUpdates );
 336         }
 337
 338         /**
 339          * Clear all pending updates without performing them. Generally, you don't
 340          * want or need to call this. Unit tests need it though.
 341          */
 342         public static function clearPendingUpdates() {
 343                 self::$preSendUpdates = [];
 344                 self::$postSendUpdates = [];
 345         }
 346
 347         /**
 348          * @return IDatabase[] Connection where commit() cannot be called yet
 349          */
 350         private static function getBusyDbConnections() {
 351                 $connsBusy = [];
 352
 353                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 354                 $lbFactory->forEachLB( function ( LoadBalancer $lb ) use ( &$connsBusy ) {
 355                         $lb->forEachOpenMasterConnection( function ( IDatabase $conn ) use ( &$connsBusy ) {
 356                                 if ( $conn->writesOrCallbacksPending() || $conn->explicitTrxActive() ) {
 357                                         $connsBusy[] = $conn;
 358                                 }
 359                         } );
 360                 } );
 361
 362                 return $connsBusy;
 363         }
 364 }