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