includes/db/loadbalancer/LBFactory.php

   1 <?php
   2 /**
   3  * Generator of database load balancing objects.
   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  * @ingroup Database
  22  */
  23
  24 use MediaWiki\MediaWikiServices;
  25 use MediaWiki\Services\DestructibleService;
  26 use Psr\Log\LoggerInterface;
  27 use MediaWiki\Logger\LoggerFactory;
  28
  29 /**
  30  * An interface for generating database load balancers
  31  * @ingroup Database
  32  */
  33 abstract class LBFactory implements DestructibleService {
  34         /** @var ChronologyProtector */
  35         protected $chronProt;
  36         /** @var TransactionProfiler */
  37         protected $trxProfiler;
  38         /** @var LoggerInterface */
  39         protected $trxLogger;
  40         /** @var BagOStuff */
  41         protected $srvCache;
  42         /** @var WANObjectCache */
  43         protected $wanCache;
  44
  45         /** @var mixed */
  46         protected $ticket;
  47         /** @var string|bool String if a requested DBO_TRX transaction round is active */
  48         protected $trxRoundId = false;
  49         /** @var string|bool Reason all LBs are read-only or false if not */
  50         protected $readOnlyReason = false;
  51         /** @var callable[] */
  52         protected $replicationWaitCallbacks = [];
  53
  54         const SHUTDOWN_NO_CHRONPROT = 1; // don't save ChronologyProtector positions (for async code)
  55
  56         /**
  57          * Construct a factory based on a configuration array (typically from $wgLBFactoryConf)
  58          * @param array $conf
  59          * @TODO: inject objects via dependency framework
  60          */
  61         public function __construct( array $conf ) {
  62                 if ( isset( $conf['readOnlyReason'] ) && is_string( $conf['readOnlyReason'] ) ) {
  63                         $this->readOnlyReason = $conf['readOnlyReason'];
  64                 }
  65                 $this->chronProt = $this->newChronologyProtector();
  66                 $this->trxProfiler = Profiler::instance()->getTransactionProfiler();
  67                 // Use APC/memcached style caching, but avoids loops with CACHE_DB (T141804)
  68                 $cache = ObjectCache::getLocalServerInstance();
  69                 if ( $cache->getQoS( $cache::ATTR_EMULATION ) > $cache::QOS_EMULATION_SQL ) {
  70                         $this->srvCache = $cache;
  71                 } else {
  72                         $this->srvCache = new EmptyBagOStuff();
  73                 }
  74                 $wCache = ObjectCache::getMainWANInstance();
  75                 if ( $wCache->getQoS( $wCache::ATTR_EMULATION ) > $wCache::QOS_EMULATION_SQL ) {
  76                         $this->wanCache = $wCache;
  77                 } else {
  78                         $this->wanCache = WANObjectCache::newEmpty();
  79                 }
  80                 $this->trxLogger = LoggerFactory::getInstance( 'DBTransaction' );
  81                 $this->ticket = mt_rand();
  82         }
  83
  84         /**
  85          * Disables all load balancers. All connections are closed, and any attempt to
  86          * open a new connection will result in a DBAccessError.
  87          * @see LoadBalancer::disable()
  88          */
  89         public function destroy() {
  90                 $this->shutdown();
  91                 $this->forEachLBCallMethod( 'disable' );
  92         }
  93
  94         /**
  95          * Disables all access to the load balancer, will cause all database access
  96          * to throw a DBAccessError
  97          */
  98         public static function disableBackend() {
  99                 MediaWikiServices::disableStorageBackend();
 100         }
 101
 102         /**
 103          * Get an LBFactory instance
 104          *
 105          * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancerFactory() instead.
 106          *
 107          * @return LBFactory
 108          */
 109         public static function singleton() {
 110                 return MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 111         }
 112
 113         /**
 114          * Returns the LBFactory class to use and the load balancer configuration.
 115          *
 116          * @todo instead of this, use a ServiceContainer for managing the different implementations.
 117          *
 118          * @param array $config (e.g. $wgLBFactoryConf)
 119          * @return string Class name
 120          */
 121         public static function getLBFactoryClass( array $config ) {
 122                 // For configuration backward compatibility after removing
 123                 // underscores from class names in MediaWiki 1.23.
 124                 $bcClasses = [
 125                         'LBFactory_Simple' => 'LBFactorySimple',
 126                         'LBFactory_Single' => 'LBFactorySingle',
 127                         'LBFactory_Multi' => 'LBFactoryMulti',
 128                         'LBFactory_Fake' => 'LBFactoryFake',
 129                 ];
 130
 131                 $class = $config['class'];
 132
 133                 if ( isset( $bcClasses[$class] ) ) {
 134                         $class = $bcClasses[$class];
 135                         wfDeprecated(
 136                                 '$wgLBFactoryConf must be updated. See RELEASE-NOTES for details',
 137                                 '1.23'
 138                         );
 139                 }
 140
 141                 return $class;
 142         }
 143
 144         /**
 145          * Shut down, close connections and destroy the cached instance.
 146          *
 147          * @deprecated since 1.27, use LBFactory::destroy()
 148          */
 149         public static function destroyInstance() {
 150                 self::singleton()->destroy();
 151         }
 152
 153         /**
 154          * Create a new load balancer object. The resulting object will be untracked,
 155          * not chronology-protected, and the caller is responsible for cleaning it up.
 156          *
 157          * @param bool|string $wiki Wiki ID, or false for the current wiki
 158          * @return LoadBalancer
 159          */
 160         abstract public function newMainLB( $wiki = false );
 161
 162         /**
 163          * Get a cached (tracked) load balancer object.
 164          *
 165          * @param bool|string $wiki Wiki ID, or false for the current wiki
 166          * @return LoadBalancer
 167          */
 168         abstract public function getMainLB( $wiki = false );
 169
 170         /**
 171          * Create a new load balancer for external storage. The resulting object will be
 172          * untracked, not chronology-protected, and the caller is responsible for
 173          * cleaning it up.
 174          *
 175          * @param string $cluster External storage cluster, or false for core
 176          * @param bool|string $wiki Wiki ID, or false for the current wiki
 177          * @return LoadBalancer
 178          */
 179         abstract protected function newExternalLB( $cluster, $wiki = false );
 180
 181         /**
 182          * Get a cached (tracked) load balancer for external storage
 183          *
 184          * @param string $cluster External storage cluster, or false for core
 185          * @param bool|string $wiki Wiki ID, or false for the current wiki
 186          * @return LoadBalancer
 187          */
 188         abstract public function &getExternalLB( $cluster, $wiki = false );
 189
 190         /**
 191          * Execute a function for each tracked load balancer
 192          * The callback is called with the load balancer as the first parameter,
 193          * and $params passed as the subsequent parameters.
 194          *
 195          * @param callable $callback
 196          * @param array $params
 197          */
 198         abstract public function forEachLB( $callback, array $params = [] );
 199
 200         /**
 201          * Prepare all tracked load balancers for shutdown
 202          * @param integer $flags Supports SHUTDOWN_* flags
 203          */
 204         public function shutdown( $flags = 0 ) {
 205                 if ( !( $flags & self::SHUTDOWN_NO_CHRONPROT ) ) {
 206                         $this->shutdownChronologyProtector( $this->chronProt );
 207                 }
 208                 $this->commitMasterChanges( __METHOD__ ); // sanity
 209         }
 210
 211         /**
 212          * Call a method of each tracked load balancer
 213          *
 214          * @param string $methodName
 215          * @param array $args
 216          */
 217         private function forEachLBCallMethod( $methodName, array $args = [] ) {
 218                 $this->forEachLB(
 219                         function ( LoadBalancer $loadBalancer, $methodName, array $args ) {
 220                                 call_user_func_array( [ $loadBalancer, $methodName ], $args );
 221                         },
 222                         [ $methodName, $args ]
 223                 );
 224         }
 225
 226         /**
 227          * Flush any master transaction snapshots and set DBO_TRX (if DBO_DEFAULT is set)
 228          *
 229          * The DBO_TRX setting will be reverted to the default in each of these methods:
 230          *   - commitMasterChanges()
 231          *   - rollbackMasterChanges()
 232          *   - commitAll()
 233          * This allows for custom transaction rounds from any outer transaction scope.
 234          *
 235          * @param string $fname
 236          * @throws DBTransactionError
 237          * @since 1.28
 238          */
 239         public function beginMasterChanges( $fname = __METHOD__ ) {
 240                 if ( $this->trxRoundId !== false ) {
 241                         throw new DBTransactionError(
 242                                 null,
 243                                 "Transaction round '{$this->trxRoundId}' already started."
 244                         );
 245                 }
 246                 $this->trxRoundId = $fname;
 247                 // Set DBO_TRX flags on all appropriate DBs
 248                 $this->forEachLBCallMethod( 'beginMasterChanges', [ $fname ] );
 249         }
 250
 251         /**
 252          * Commit all replica DB transactions so as to flush any REPEATABLE-READ or SSI snapshot
 253          *
 254          * @param string $fname Caller name
 255          * @since 1.28
 256          */
 257         public function flushReplicaSnapshots( $fname = __METHOD__ ) {
 258                 $this->forEachLBCallMethod( 'flushReplicaSnapshots', [ $fname ] );
 259         }
 260
 261         /**
 262          * Commit on all connections. Done for two reasons:
 263          * 1. To commit changes to the masters.
 264          * 2. To release the snapshot on all connections, master and replica DB.
 265          * @param string $fname Caller name
 266          * @param array $options Options map:
 267          *   - maxWriteDuration: abort if more than this much time was spent in write queries
 268          */
 269         public function commitAll( $fname = __METHOD__, array $options = [] ) {
 270                 $this->commitMasterChanges( $fname, $options );
 271                 $this->forEachLBCallMethod( 'commitAll', [ $fname ] );
 272         }
 273
 274         /**
 275          * Commit changes on all master connections
 276          * @param string $fname Caller name
 277          * @param array $options Options map:
 278          *   - maxWriteDuration: abort if more than this much time was spent in write queries
 279          * @throws Exception
 280          */
 281         public function commitMasterChanges( $fname = __METHOD__, array $options = [] ) {
 282                 // Run pre-commit callbacks and suppress post-commit callbacks, aborting on failure
 283                 $this->forEachLBCallMethod( 'finalizeMasterChanges' );
 284                 $this->trxRoundId = false;
 285                 // Perform pre-commit checks, aborting on failure
 286                 $this->forEachLBCallMethod( 'approveMasterChanges', [ $options ] );
 287                 // Log the DBs and methods involved in multi-DB transactions
 288                 $this->logIfMultiDbTransaction();
 289                 // Actually perform the commit on all master DB connections and revert DBO_TRX
 290                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 291                 // Run all post-commit callbacks
 292                 /** @var Exception $e */
 293                 $e = null; // first callback exception
 294                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$e ) {
 295                         $ex = $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_COMMIT );
 296                         $e = $e ?: $ex;
 297                 } );
 298                 // Commit any dangling DBO_TRX transactions from callbacks on one DB to another DB
 299                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 300                 // Throw any last post-commit callback error
 301                 if ( $e instanceof Exception ) {
 302                         throw $e;
 303                 }
 304         }
 305
 306         /**
 307          * Rollback changes on all master connections
 308          * @param string $fname Caller name
 309          * @since 1.23
 310          */
 311         public function rollbackMasterChanges( $fname = __METHOD__ ) {
 312                 $this->trxRoundId = false;
 313                 $this->forEachLBCallMethod( 'suppressTransactionEndCallbacks' );
 314                 $this->forEachLBCallMethod( 'rollbackMasterChanges', [ $fname ] );
 315                 // Run all post-rollback callbacks
 316                 $this->forEachLB( function ( LoadBalancer $lb ) {
 317                         $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_ROLLBACK );
 318                 } );
 319         }
 320
 321         /**
 322          * Log query info if multi DB transactions are going to be committed now
 323          */
 324         private function logIfMultiDbTransaction() {
 325                 $callersByDB = [];
 326                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$callersByDB ) {
 327                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 328                         $callers = $lb->pendingMasterChangeCallers();
 329                         if ( $callers ) {
 330                                 $callersByDB[$masterName] = $callers;
 331                         }
 332                 } );
 333
 334                 if ( count( $callersByDB ) >= 2 ) {
 335                         $dbs = implode( ', ', array_keys( $callersByDB ) );
 336                         $msg = "Multi-DB transaction [{$dbs}]:\n";
 337                         foreach ( $callersByDB as $db => $callers ) {
 338                                 $msg .= "$db: " . implode( '; ', $callers ) . "\n";
 339                         }
 340                         $this->trxLogger->info( $msg );
 341                 }
 342         }
 343
 344         /**
 345          * Determine if any master connection has pending changes
 346          * @return bool
 347          * @since 1.23
 348          */
 349         public function hasMasterChanges() {
 350                 $ret = false;
 351                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 352                         $ret = $ret || $lb->hasMasterChanges();
 353                 } );
 354
 355                 return $ret;
 356         }
 357
 358         /**
 359          * Detemine if any lagged replica DB connection was used
 360          * @since 1.27
 361          * @return bool
 362          */
 363         public function laggedSlaveUsed() {
 364                 $ret = false;
 365                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 366                         $ret = $ret || $lb->laggedSlaveUsed();
 367                 } );
 368
 369                 return $ret;
 370         }
 371
 372         /**
 373          * Determine if any master connection has pending/written changes from this request
 374          * @return bool
 375          * @since 1.27
 376          */
 377         public function hasOrMadeRecentMasterChanges() {
 378                 $ret = false;
 379                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 380                         $ret = $ret || $lb->hasOrMadeRecentMasterChanges();
 381                 } );
 382                 return $ret;
 383         }
 384
 385         /**
 386          * Waits for the replica DBs to catch up to the current master position
 387          *
 388          * Use this when updating very large numbers of rows, as in maintenance scripts,
 389          * to avoid causing too much lag. Of course, this is a no-op if there are no replica DBs.
 390          *
 391          * By default this waits on all DB clusters actually used in this request.
 392          * This makes sense when lag being waiting on is caused by the code that does this check.
 393          * In that case, setting "ifWritesSince" can avoid the overhead of waiting for clusters
 394          * that were not changed since the last wait check. To forcefully wait on a specific cluster
 395          * for a given wiki, use the 'wiki' parameter. To forcefully wait on an "external" cluster,
 396          * use the "cluster" parameter.
 397          *
 398          * Never call this function after a large DB write that is *still* in a transaction.
 399          * It only makes sense to call this after the possible lag inducing changes were committed.
 400          *
 401          * @param array $opts Optional fields that include:
 402          *   - wiki : wait on the load balancer DBs that handles the given wiki
 403          *   - cluster : wait on the given external load balancer DBs
 404          *   - timeout : Max wait time. Default: ~60 seconds
 405          *   - ifWritesSince: Only wait if writes were done since this UNIX timestamp
 406          * @throws DBReplicationWaitError If a timeout or error occured waiting on a DB cluster
 407          * @since 1.27
 408          */
 409         public function waitForReplication( array $opts = [] ) {
 410                 $opts += [
 411                         'wiki' => false,
 412                         'cluster' => false,
 413                         'timeout' => 60,
 414                         'ifWritesSince' => null
 415                 ];
 416
 417                 foreach ( $this->replicationWaitCallbacks as $callback ) {
 418                         $callback();
 419                 }
 420
 421                 // Figure out which clusters need to be checked
 422                 /** @var LoadBalancer[] $lbs */
 423                 $lbs = [];
 424                 if ( $opts['cluster'] !== false ) {
 425                         $lbs[] = $this->getExternalLB( $opts['cluster'] );
 426                 } elseif ( $opts['wiki'] !== false ) {
 427                         $lbs[] = $this->getMainLB( $opts['wiki'] );
 428                 } else {
 429                         $this->forEachLB( function ( LoadBalancer $lb ) use ( &$lbs ) {
 430                                 $lbs[] = $lb;
 431                         } );
 432                         if ( !$lbs ) {
 433                                 return; // nothing actually used
 434                         }
 435                 }
 436
 437                 // Get all the master positions of applicable DBs right now.
 438                 // This can be faster since waiting on one cluster reduces the
 439                 // time needed to wait on the next clusters.
 440                 $masterPositions = array_fill( 0, count( $lbs ), false );
 441                 foreach ( $lbs as $i => $lb ) {
 442                         if ( $lb->getServerCount() <= 1 ) {
 443                                 // Bug 27975 - Don't try to wait for replica DBs if there are none
 444                                 // Prevents permission error when getting master position
 445                                 continue;
 446                         } elseif ( $opts['ifWritesSince']
 447                                 && $lb->lastMasterChangeTimestamp() < $opts['ifWritesSince']
 448                         ) {
 449                                 continue; // no writes since the last wait
 450                         }
 451                         $masterPositions[$i] = $lb->getMasterPos();
 452                 }
 453
 454                 $failed = [];
 455                 foreach ( $lbs as $i => $lb ) {
 456                         if ( $masterPositions[$i] ) {
 457                                 // The DBMS may not support getMasterPos() or the whole
 458                                 // load balancer might be fake (e.g. $wgAllDBsAreLocalhost).
 459                                 if ( !$lb->waitForAll( $masterPositions[$i], $opts['timeout'] ) ) {
 460                                         $failed[] = $lb->getServerName( $lb->getWriterIndex() );
 461                                 }
 462                         }
 463                 }
 464
 465                 if ( $failed ) {
 466                         throw new DBReplicationWaitError(
 467                                 "Could not wait for replica DBs to catch up to " .
 468                                 implode( ', ', $failed )
 469                         );
 470                 }
 471         }
 472
 473         /**
 474          * Add a callback to be run in every call to waitForReplication() before waiting
 475          *
 476          * Callbacks must clear any transactions that they start
 477          *
 478          * @param string $name Callback name
 479          * @param callable|null $callback Use null to unset a callback
 480          * @since 1.28
 481          */
 482         public function setWaitForReplicationListener( $name, callable $callback = null ) {
 483                 if ( $callback ) {
 484                         $this->replicationWaitCallbacks[$name] = $callback;
 485                 } else {
 486                         unset( $this->replicationWaitCallbacks[$name] );
 487                 }
 488         }
 489
 490         /**
 491          * Get a token asserting that no transaction writes are active
 492          *
 493          * @param string $fname Caller name (e.g. __METHOD__)
 494          * @return mixed A value to pass to commitAndWaitForReplication()
 495          * @since 1.28
 496          */
 497         public function getEmptyTransactionTicket( $fname ) {
 498                 if ( $this->hasMasterChanges() ) {
 499                         $this->trxLogger->error( __METHOD__ . ": $fname does not have outer scope." );
 500                         return null;
 501                 }
 502
 503                 return $this->ticket;
 504         }
 505
 506         /**
 507          * Convenience method for safely running commitMasterChanges()/waitForReplication()
 508          *
 509          * This will commit and wait unless $ticket indicates it is unsafe to do so
 510          *
 511          * @param string $fname Caller name (e.g. __METHOD__)
 512          * @param mixed $ticket Result of getOuterTransactionScopeTicket()
 513          * @param array $opts Options to waitForReplication()
 514          * @throws DBReplicationWaitError
 515          * @since 1.28
 516          */
 517         public function commitAndWaitForReplication( $fname, $ticket, array $opts = [] ) {
 518                 if ( $ticket !== $this->ticket ) {
 519                         $logger = LoggerFactory::getInstance( 'DBPerformance' );
 520                         $logger->error( __METHOD__ . ": cannot commit; $fname does not have outer scope." );
 521                         return;
 522                 }
 523
 524                 $this->commitMasterChanges( $fname );
 525                 $this->waitForReplication( $opts );
 526         }
 527
 528         /**
 529          * Disable the ChronologyProtector for all load balancers
 530          *
 531          * This can be called at the start of special API entry points
 532          *
 533          * @since 1.27
 534          */
 535         public function disableChronologyProtection() {
 536                 $this->chronProt->setEnabled( false );
 537         }
 538
 539         /**
 540          * @return ChronologyProtector
 541          */
 542         protected function newChronologyProtector() {
 543                 $request = RequestContext::getMain()->getRequest();
 544                 $chronProt = new ChronologyProtector(
 545                         ObjectCache::getMainStashInstance(),
 546                         [
 547                                 'ip' => $request->getIP(),
 548                                 'agent' => $request->getHeader( 'User-Agent' )
 549                         ]
 550                 );
 551                 if ( PHP_SAPI === 'cli' ) {
 552                         $chronProt->setEnabled( false );
 553                 } elseif ( $request->getHeader( 'ChronologyProtection' ) === 'false' ) {
 554                         // Request opted out of using position wait logic. This is useful for requests
 555                         // done by the job queue or background ETL that do not have a meaningful session.
 556                         $chronProt->setWaitEnabled( false );
 557                 }
 558
 559                 return $chronProt;
 560         }
 561
 562         /**
 563          * @param ChronologyProtector $cp
 564          */
 565         protected function shutdownChronologyProtector( ChronologyProtector $cp ) {
 566                 // Get all the master positions needed
 567                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $cp ) {
 568                         $cp->shutdownLB( $lb );
 569                 } );
 570                 // Write them to the stash
 571                 $unsavedPositions = $cp->shutdown();
 572                 // If the positions failed to write to the stash, at least wait on local datacenter
 573                 // replica DBs to catch up before responding. Even if there are several DCs, this increases
 574                 // the chance that the user will see their own changes immediately afterwards. As long
 575                 // as the sticky DC cookie applies (same domain), this is not even an issue.
 576                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $unsavedPositions ) {
 577                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 578                         if ( isset( $unsavedPositions[$masterName] ) ) {
 579                                 $lb->waitForAll( $unsavedPositions[$masterName] );
 580                         }
 581                 } );
 582         }
 583
 584         /**
 585          * @param LoadBalancer $lb
 586          */
 587         protected function initLoadBalancer( LoadBalancer $lb ) {
 588                 if ( $this->trxRoundId !== false ) {
 589                         $lb->beginMasterChanges( $this->trxRoundId ); // set DBO_TRX
 590                 }
 591         }
 592
 593         /**
 594          * Close all open database connections on all open load balancers.
 595          * @since 1.28
 596          */
 597         public function closeAll() {
 598                 $this->forEachLBCallMethod( 'closeAll', [] );
 599         }
 600
 601 }