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          * STUB
 200          */
 201         public function shutdown( $flags = 0 ) {
 202         }
 203
 204         /**
 205          * Call a method of each tracked load balancer
 206          *
 207          * @param string $methodName
 208          * @param array $args
 209          */
 210         private function forEachLBCallMethod( $methodName, array $args = [] ) {
 211                 $this->forEachLB(
 212                         function ( LoadBalancer $loadBalancer, $methodName, array $args ) {
 213                                 call_user_func_array( [ $loadBalancer, $methodName ], $args );
 214                         },
 215                         [ $methodName, $args ]
 216                 );
 217         }
 218
 219         /**
 220          * Flush any master transaction snapshots and set DBO_TRX (if DBO_DEFAULT is set)
 221          *
 222          * The DBO_TRX setting will be reverted to the default in each of these methods:
 223          *   - commitMasterChanges()
 224          *   - rollbackMasterChanges()
 225          *   - commitAll()
 226          * This allows for custom transaction rounds from any outer transaction scope.
 227          *
 228          * @param string $fname
 229          * @since 1.28
 230          */
 231         public function beginMasterChanges( $fname = __METHOD__ ) {
 232                 $this->forEachLBCallMethod( 'beginMasterChanges', [ $fname ] );
 233         }
 234
 235         /**
 236          * Commit on all connections. Done for two reasons:
 237          * 1. To commit changes to the masters.
 238          * 2. To release the snapshot on all connections, master and slave.
 239          * @param string $fname Caller name
 240          * @param array $options Options map:
 241          *   - maxWriteDuration: abort if more than this much time was spent in write queries
 242          */
 243         public function commitAll( $fname = __METHOD__, array $options = [] ) {
 244                 $this->commitMasterChanges( $fname, $options );
 245                 $this->forEachLBCallMethod( 'commitAll', [ $fname ] );
 246         }
 247
 248         /**
 249          * Commit changes on all master connections
 250          * @param string $fname Caller name
 251          * @param array $options Options map:
 252          *   - maxWriteDuration: abort if more than this much time was spent in write queries
 253          * @throws Exception
 254          */
 255         public function commitMasterChanges( $fname = __METHOD__, array $options = [] ) {
 256                 // Perform all pre-commit callbacks, aborting on failure
 257                 $this->forEachLBCallMethod( 'runMasterPreCommitCallbacks' );
 258                 // Perform all pre-commit checks, aborting on failure
 259                 $this->forEachLBCallMethod( 'approveMasterChanges', [ $options ] );
 260                 // Log the DBs and methods involved in multi-DB transactions
 261                 $this->logIfMultiDbTransaction();
 262                 // Actually perform the commit on all master DB connections
 263                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 264                 // Run all post-commit callbacks
 265                 /** @var Exception $e */
 266                 $e = null; // first callback exception
 267                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$e ) {
 268                         $ex = $lb->runMasterPostCommitCallbacks();
 269                         $e = $e ?: $ex;
 270                 } );
 271                 // Commit any dangling DBO_TRX transactions from callbacks on one DB to another DB
 272                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 273                 // Throw any last post-commit callback error
 274                 if ( $e instanceof Exception ) {
 275                         throw $e;
 276                 }
 277         }
 278
 279         /**
 280          * Rollback changes on all master connections
 281          * @param string $fname Caller name
 282          * @since 1.23
 283          */
 284         public function rollbackMasterChanges( $fname = __METHOD__ ) {
 285                 $this->forEachLBCallMethod( 'rollbackMasterChanges', [ $fname ] );
 286         }
 287
 288         /**
 289          * Log query info if multi DB transactions are going to be committed now
 290          */
 291         private function logIfMultiDbTransaction() {
 292                 $callersByDB = [];
 293                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$callersByDB ) {
 294                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 295                         $callers = $lb->pendingMasterChangeCallers();
 296                         if ( $callers ) {
 297                                 $callersByDB[$masterName] = $callers;
 298                         }
 299                 } );
 300
 301                 if ( count( $callersByDB ) >= 2 ) {
 302                         $dbs = implode( ', ', array_keys( $callersByDB ) );
 303                         $msg = "Multi-DB transaction [{$dbs}]:\n";
 304                         foreach ( $callersByDB as $db => $callers ) {
 305                                 $msg .= "$db: " . implode( '; ', $callers ) . "\n";
 306                         }
 307                         $this->trxLogger->info( $msg );
 308                 }
 309         }
 310
 311         /**
 312          * Determine if any master connection has pending changes
 313          * @return bool
 314          * @since 1.23
 315          */
 316         public function hasMasterChanges() {
 317                 $ret = false;
 318                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 319                         $ret = $ret || $lb->hasMasterChanges();
 320                 } );
 321
 322                 return $ret;
 323         }
 324
 325         /**
 326          * Detemine if any lagged slave connection was used
 327          * @since 1.27
 328          * @return bool
 329          */
 330         public function laggedSlaveUsed() {
 331                 $ret = false;
 332                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 333                         $ret = $ret || $lb->laggedSlaveUsed();
 334                 } );
 335
 336                 return $ret;
 337         }
 338
 339         /**
 340          * Determine if any master connection has pending/written changes from this request
 341          * @return bool
 342          * @since 1.27
 343          */
 344         public function hasOrMadeRecentMasterChanges() {
 345                 $ret = false;
 346                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 347                         $ret = $ret || $lb->hasOrMadeRecentMasterChanges();
 348                 } );
 349                 return $ret;
 350         }
 351
 352         /**
 353          * Waits for the slave DBs to catch up to the current master position
 354          *
 355          * Use this when updating very large numbers of rows, as in maintenance scripts,
 356          * to avoid causing too much lag. Of course, this is a no-op if there are no slaves.
 357          *
 358          * By default this waits on all DB clusters actually used in this request.
 359          * This makes sense when lag being waiting on is caused by the code that does this check.
 360          * In that case, setting "ifWritesSince" can avoid the overhead of waiting for clusters
 361          * that were not changed since the last wait check. To forcefully wait on a specific cluster
 362          * for a given wiki, use the 'wiki' parameter. To forcefully wait on an "external" cluster,
 363          * use the "cluster" parameter.
 364          *
 365          * Never call this function after a large DB write that is *still* in a transaction.
 366          * It only makes sense to call this after the possible lag inducing changes were committed.
 367          *
 368          * @param array $opts Optional fields that include:
 369          *   - wiki : wait on the load balancer DBs that handles the given wiki
 370          *   - cluster : wait on the given external load balancer DBs
 371          *   - timeout : Max wait time. Default: ~60 seconds
 372          *   - ifWritesSince: Only wait if writes were done since this UNIX timestamp
 373          * @throws DBReplicationWaitError If a timeout or error occured waiting on a DB cluster
 374          * @since 1.27
 375          */
 376         public function waitForReplication( array $opts = [] ) {
 377                 $opts += [
 378                         'wiki' => false,
 379                         'cluster' => false,
 380                         'timeout' => 60,
 381                         'ifWritesSince' => null
 382                 ];
 383
 384                 // Figure out which clusters need to be checked
 385                 /** @var LoadBalancer[] $lbs */
 386                 $lbs = [];
 387                 if ( $opts['cluster'] !== false ) {
 388                         $lbs[] = $this->getExternalLB( $opts['cluster'] );
 389                 } elseif ( $opts['wiki'] !== false ) {
 390                         $lbs[] = $this->getMainLB( $opts['wiki'] );
 391                 } else {
 392                         $this->forEachLB( function ( LoadBalancer $lb ) use ( &$lbs ) {
 393                                 $lbs[] = $lb;
 394                         } );
 395                         if ( !$lbs ) {
 396                                 return; // nothing actually used
 397                         }
 398                 }
 399
 400                 // Get all the master positions of applicable DBs right now.
 401                 // This can be faster since waiting on one cluster reduces the
 402                 // time needed to wait on the next clusters.
 403                 $masterPositions = array_fill( 0, count( $lbs ), false );
 404                 foreach ( $lbs as $i => $lb ) {
 405                         if ( $lb->getServerCount() <= 1 ) {
 406                                 // Bug 27975 - Don't try to wait for slaves if there are none
 407                                 // Prevents permission error when getting master position
 408                                 continue;
 409                         } elseif ( $opts['ifWritesSince']
 410                                 && $lb->lastMasterChangeTimestamp() < $opts['ifWritesSince']
 411                         ) {
 412                                 continue; // no writes since the last wait
 413                         }
 414                         $masterPositions[$i] = $lb->getMasterPos();
 415                 }
 416
 417                 $failed = [];
 418                 foreach ( $lbs as $i => $lb ) {
 419                         if ( $masterPositions[$i] ) {
 420                                 // The DBMS may not support getMasterPos() or the whole
 421                                 // load balancer might be fake (e.g. $wgAllDBsAreLocalhost).
 422                                 if ( !$lb->waitForAll( $masterPositions[$i], $opts['timeout'] ) ) {
 423                                         $failed[] = $lb->getServerName( $lb->getWriterIndex() );
 424                                 }
 425                         }
 426                 }
 427
 428                 if ( $failed ) {
 429                         throw new DBReplicationWaitError(
 430                                 "Could not wait for slaves to catch up to " .
 431                                 implode( ', ', $failed )
 432                         );
 433                 }
 434         }
 435
 436         /**
 437          * Get a token asserting that no transaction writes are active
 438          *
 439          * @param string $fname Caller name (e.g. __METHOD__)
 440          * @return mixed A value to pass to commitAndWaitForReplication()
 441          * @since 1.28
 442          */
 443         public function getEmptyTransactionTicket( $fname ) {
 444                 if ( $this->hasMasterChanges() ) {
 445                         $this->trxLogger->error( __METHOD__ . ": $fname does not have outer scope." );
 446                         return null;
 447                 }
 448
 449                 return $this->ticket;
 450         }
 451
 452         /**
 453          * Convenience method for safely running commitMasterChanges()/waitForReplication()
 454          *
 455          * This will commit and wait unless $ticket indicates it is unsafe to do so
 456          *
 457          * @param string $fname Caller name (e.g. __METHOD__)
 458          * @param mixed $ticket Result of getOuterTransactionScopeTicket()
 459          * @param array $opts Options to waitForReplication()
 460          * @throws DBReplicationWaitError
 461          * @since 1.28
 462          */
 463         public function commitAndWaitForReplication( $fname, $ticket, array $opts = [] ) {
 464                 if ( $ticket !== $this->ticket ) {
 465                         $logger = LoggerFactory::getInstance( 'DBPerformance' );
 466                         $logger->error( __METHOD__ . ": cannot commit; $fname does not have outer scope." );
 467                         return;
 468                 }
 469
 470                 $this->commitMasterChanges( $fname );
 471                 $this->waitForReplication( $opts );
 472         }
 473
 474         /**
 475          * Disable the ChronologyProtector for all load balancers
 476          *
 477          * This can be called at the start of special API entry points
 478          *
 479          * @since 1.27
 480          */
 481         public function disableChronologyProtection() {
 482                 $this->chronProt->setEnabled( false );
 483         }
 484
 485         /**
 486          * @return ChronologyProtector
 487          */
 488         protected function newChronologyProtector() {
 489                 $request = RequestContext::getMain()->getRequest();
 490                 $chronProt = new ChronologyProtector(
 491                         ObjectCache::getMainStashInstance(),
 492                         [
 493                                 'ip' => $request->getIP(),
 494                                 'agent' => $request->getHeader( 'User-Agent' )
 495                         ]
 496                 );
 497                 if ( PHP_SAPI === 'cli' ) {
 498                         $chronProt->setEnabled( false );
 499                 } elseif ( $request->getHeader( 'ChronologyProtection' ) === 'false' ) {
 500                         // Request opted out of using position wait logic. This is useful for requests
 501                         // done by the job queue or background ETL that do not have a meaningful session.
 502                         $chronProt->setWaitEnabled( false );
 503                 }
 504
 505                 return $chronProt;
 506         }
 507
 508         /**
 509          * @param ChronologyProtector $cp
 510          */
 511         protected function shutdownChronologyProtector( ChronologyProtector $cp ) {
 512                 // Get all the master positions needed
 513                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $cp ) {
 514                         $cp->shutdownLB( $lb );
 515                 } );
 516                 // Write them to the stash
 517                 $unsavedPositions = $cp->shutdown();
 518                 // If the positions failed to write to the stash, at least wait on local datacenter
 519                 // slaves to catch up before responding. Even if there are several DCs, this increases
 520                 // the chance that the user will see their own changes immediately afterwards. As long
 521                 // as the sticky DC cookie applies (same domain), this is not even an issue.
 522                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $unsavedPositions ) {
 523                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 524                         if ( isset( $unsavedPositions[$masterName] ) ) {
 525                                 $lb->waitForAll( $unsavedPositions[$masterName] );
 526                         }
 527                 } );
 528         }
 529
 530         /**
 531          * Close all open database connections on all open load balancers.
 532          * @since 1.28
 533          */
 534         public function closeAll() {
 535                 $this->forEachLBCallMethod( 'closeAll', [] );
 536         }
 537
 538 }
 539
 540 /**
 541  * Exception class for attempted DB access
 542  */
 543 class DBAccessError extends MWException {
 544         public function __construct() {
 545                 parent::__construct( "Mediawiki tried to access the database via wfGetDB(). " .
 546                         "This is not allowed, because database access has been disabled." );
 547         }
 548 }
 549
 550 /**
 551  * Exception class for replica DB wait timeouts
 552  */
 553 class DBReplicationWaitError extends Exception {
 554 }