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
  35         /** @var ChronologyProtector */
  36         protected $chronProt;
  37
  38         /** @var TransactionProfiler */
  39         protected $trxProfiler;
  40
  41         /** @var LoggerInterface */
  42         protected $logger;
  43
  44         /** @var string|bool Reason all LBs are read-only or false if not */
  45         protected $readOnlyReason = false;
  46
  47         const SHUTDOWN_NO_CHRONPROT = 1; // don't save ChronologyProtector positions (for async code)
  48
  49         /**
  50          * Construct a factory based on a configuration array (typically from $wgLBFactoryConf)
  51          * @param array $conf
  52          */
  53         public function __construct( array $conf ) {
  54                 if ( isset( $conf['readOnlyReason'] ) && is_string( $conf['readOnlyReason'] ) ) {
  55                         $this->readOnlyReason = $conf['readOnlyReason'];
  56                 }
  57
  58                 $this->chronProt = $this->newChronologyProtector();
  59                 $this->trxProfiler = Profiler::instance()->getTransactionProfiler();
  60                 $this->logger = LoggerFactory::getInstance( 'DBTransaction' );
  61         }
  62
  63         /**
  64          * Disables all load balancers. All connections are closed, and any attempt to
  65          * open a new connection will result in a DBAccessError.
  66          * @see LoadBalancer::disable()
  67          */
  68         public function destroy() {
  69                 $this->shutdown();
  70                 $this->forEachLBCallMethod( 'disable' );
  71         }
  72
  73         /**
  74          * Disables all access to the load balancer, will cause all database access
  75          * to throw a DBAccessError
  76          */
  77         public static function disableBackend() {
  78                 MediaWikiServices::disableStorageBackend();
  79         }
  80
  81         /**
  82          * Get an LBFactory instance
  83          *
  84          * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancerFactory() instead.
  85          *
  86          * @return LBFactory
  87          */
  88         public static function singleton() {
  89                 return MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
  90         }
  91
  92         /**
  93          * Returns the LBFactory class to use and the load balancer configuration.
  94          *
  95          * @todo instead of this, use a ServiceContainer for managing the different implementations.
  96          *
  97          * @param array $config (e.g. $wgLBFactoryConf)
  98          * @return string Class name
  99          */
 100         public static function getLBFactoryClass( array $config ) {
 101                 // For configuration backward compatibility after removing
 102                 // underscores from class names in MediaWiki 1.23.
 103                 $bcClasses = [
 104                         'LBFactory_Simple' => 'LBFactorySimple',
 105                         'LBFactory_Single' => 'LBFactorySingle',
 106                         'LBFactory_Multi' => 'LBFactoryMulti',
 107                         'LBFactory_Fake' => 'LBFactoryFake',
 108                 ];
 109
 110                 $class = $config['class'];
 111
 112                 if ( isset( $bcClasses[$class] ) ) {
 113                         $class = $bcClasses[$class];
 114                         wfDeprecated(
 115                                 '$wgLBFactoryConf must be updated. See RELEASE-NOTES for details',
 116                                 '1.23'
 117                         );
 118                 }
 119
 120                 return $class;
 121         }
 122
 123         /**
 124          * Create a new load balancer object. The resulting object will be untracked,
 125          * not chronology-protected, and the caller is responsible for cleaning it up.
 126          *
 127          * @param bool|string $wiki Wiki ID, or false for the current wiki
 128          * @return LoadBalancer
 129          */
 130         abstract public function newMainLB( $wiki = false );
 131
 132         /**
 133          * Get a cached (tracked) load balancer object.
 134          *
 135          * @param bool|string $wiki Wiki ID, or false for the current wiki
 136          * @return LoadBalancer
 137          */
 138         abstract public function getMainLB( $wiki = false );
 139
 140         /**
 141          * Create a new load balancer for external storage. The resulting object will be
 142          * untracked, not chronology-protected, and the caller is responsible for
 143          * cleaning it up.
 144          *
 145          * @param string $cluster External storage cluster, or false for core
 146          * @param bool|string $wiki Wiki ID, or false for the current wiki
 147          * @return LoadBalancer
 148          */
 149         abstract protected function newExternalLB( $cluster, $wiki = false );
 150
 151         /**
 152          * Get a cached (tracked) load balancer for external storage
 153          *
 154          * @param string $cluster External storage cluster, or false for core
 155          * @param bool|string $wiki Wiki ID, or false for the current wiki
 156          * @return LoadBalancer
 157          */
 158         abstract public function &getExternalLB( $cluster, $wiki = false );
 159
 160         /**
 161          * Execute a function for each tracked load balancer
 162          * The callback is called with the load balancer as the first parameter,
 163          * and $params passed as the subsequent parameters.
 164          *
 165          * @param callable $callback
 166          * @param array $params
 167          */
 168         abstract public function forEachLB( $callback, array $params = [] );
 169
 170         /**
 171          * Prepare all tracked load balancers for shutdown
 172          * @param integer $flags Supports SHUTDOWN_* flags
 173          * STUB
 174          */
 175         public function shutdown( $flags = 0 ) {
 176         }
 177
 178         /**
 179          * Call a method of each tracked load balancer
 180          *
 181          * @param string $methodName
 182          * @param array $args
 183          */
 184         private function forEachLBCallMethod( $methodName, array $args = [] ) {
 185                 $this->forEachLB(
 186                         function ( LoadBalancer $loadBalancer, $methodName, array $args ) {
 187                                 call_user_func_array( [ $loadBalancer, $methodName ], $args );
 188                         },
 189                         [ $methodName, $args ]
 190                 );
 191         }
 192
 193         /**
 194          * Commit on all connections. Done for two reasons:
 195          * 1. To commit changes to the masters.
 196          * 2. To release the snapshot on all connections, master and slave.
 197          * @param string $fname Caller name
 198          */
 199         public function commitAll( $fname = __METHOD__ ) {
 200                 $this->logMultiDbTransaction();
 201
 202                 $start = microtime( true );
 203                 $this->forEachLBCallMethod( 'commitAll', [ $fname ] );
 204                 $timeMs = 1000 * ( microtime( true ) - $start );
 205
 206                 RequestContext::getMain()->getStats()->timing( "db.commit-all", $timeMs );
 207         }
 208
 209         /**
 210          * Commit changes on all master connections
 211          * @param string $fname Caller name
 212          * @param array $options Options map:
 213          *   - maxWriteDuration: abort if more than this much time was spent in write queries
 214          */
 215         public function commitMasterChanges( $fname = __METHOD__, array $options = [] ) {
 216                 $limit = isset( $options['maxWriteDuration'] ) ? $options['maxWriteDuration'] : 0;
 217
 218                 $this->logMultiDbTransaction();
 219                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $limit ) {
 220                         $lb->forEachOpenConnection( function ( IDatabase $db ) use ( $limit ) {
 221                                 $time = $db->pendingWriteQueryDuration();
 222                                 if ( $limit > 0 && $time > $limit ) {
 223                                         throw new DBTransactionError(
 224                                                 $db,
 225                                                 wfMessage( 'transaction-duration-limit-exceeded', $time, $limit )->text()
 226                                         );
 227                                 }
 228                         } );
 229                 } );
 230
 231                 $start = microtime( true );
 232                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 233                 $timeMs = 1000 * ( microtime( true ) - $start );
 234
 235                 RequestContext::getMain()->getStats()->timing( "db.commit-masters", $timeMs );
 236         }
 237
 238         /**
 239          * Rollback changes on all master connections
 240          * @param string $fname Caller name
 241          * @since 1.23
 242          */
 243         public function rollbackMasterChanges( $fname = __METHOD__ ) {
 244                 $this->forEachLBCallMethod( 'rollbackMasterChanges', [ $fname ] );
 245         }
 246
 247         /**
 248          * Log query info if multi DB transactions are going to be committed now
 249          */
 250         private function logMultiDbTransaction() {
 251                 $callersByDB = [];
 252                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$callersByDB ) {
 253                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 254                         $callers = $lb->pendingMasterChangeCallers();
 255                         if ( $callers ) {
 256                                 $callersByDB[$masterName] = $callers;
 257                         }
 258                 } );
 259
 260                 if ( count( $callersByDB ) >= 2 ) {
 261                         $dbs = implode( ', ', array_keys( $callersByDB ) );
 262                         $msg = "Multi-DB transaction [{$dbs}]:\n";
 263                         foreach ( $callersByDB as $db => $callers ) {
 264                                 $msg .= "$db: " . implode( '; ', $callers ) . "\n";
 265                         }
 266                         $this->logger->info( $msg );
 267                 }
 268         }
 269
 270         /**
 271          * Determine if any master connection has pending changes
 272          * @return bool
 273          * @since 1.23
 274          */
 275         public function hasMasterChanges() {
 276                 $ret = false;
 277                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 278                         $ret = $ret || $lb->hasMasterChanges();
 279                 } );
 280
 281                 return $ret;
 282         }
 283
 284         /**
 285          * Detemine if any lagged slave connection was used
 286          * @since 1.27
 287          * @return bool
 288          */
 289         public function laggedSlaveUsed() {
 290                 $ret = false;
 291                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 292                         $ret = $ret || $lb->laggedSlaveUsed();
 293                 } );
 294
 295                 return $ret;
 296         }
 297
 298         /**
 299          * Determine if any master connection has pending/written changes from this request
 300          * @return bool
 301          * @since 1.27
 302          */
 303         public function hasOrMadeRecentMasterChanges() {
 304                 $ret = false;
 305                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 306                         $ret = $ret || $lb->hasOrMadeRecentMasterChanges();
 307                 } );
 308                 return $ret;
 309         }
 310
 311         /**
 312          * Waits for the slave DBs to catch up to the current master position
 313          *
 314          * Use this when updating very large numbers of rows, as in maintenance scripts,
 315          * to avoid causing too much lag. Of course, this is a no-op if there are no slaves.
 316          *
 317          * By default this waits on all DB clusters actually used in this request.
 318          * This makes sense when lag being waiting on is caused by the code that does this check.
 319          * In that case, setting "ifWritesSince" can avoid the overhead of waiting for clusters
 320          * that were not changed since the last wait check. To forcefully wait on a specific cluster
 321          * for a given wiki, use the 'wiki' parameter. To forcefully wait on an "external" cluster,
 322          * use the "cluster" parameter.
 323          *
 324          * Never call this function after a large DB write that is *still* in a transaction.
 325          * It only makes sense to call this after the possible lag inducing changes were committed.
 326          *
 327          * @param array $opts Optional fields that include:
 328          *   - wiki : wait on the load balancer DBs that handles the given wiki
 329          *   - cluster : wait on the given external load balancer DBs
 330          *   - timeout : Max wait time. Default: ~60 seconds
 331          *   - ifWritesSince: Only wait if writes were done since this UNIX timestamp
 332          * @throws DBReplicationWaitError If a timeout or error occured waiting on a DB cluster
 333          * @since 1.27
 334          */
 335         public function waitForReplication( array $opts = [] ) {
 336                 $opts += [
 337                         'wiki' => false,
 338                         'cluster' => false,
 339                         'timeout' => 60,
 340                         'ifWritesSince' => null
 341                 ];
 342
 343                 // Figure out which clusters need to be checked
 344                 /** @var LoadBalancer[] $lbs */
 345                 $lbs = [];
 346                 if ( $opts['cluster'] !== false ) {
 347                         $lbs[] = $this->getExternalLB( $opts['cluster'] );
 348                 } elseif ( $opts['wiki'] !== false ) {
 349                         $lbs[] = $this->getMainLB( $opts['wiki'] );
 350                 } else {
 351                         $this->forEachLB( function ( LoadBalancer $lb ) use ( &$lbs ) {
 352                                 $lbs[] = $lb;
 353                         } );
 354                         if ( !$lbs ) {
 355                                 return; // nothing actually used
 356                         }
 357                 }
 358
 359                 // Get all the master positions of applicable DBs right now.
 360                 // This can be faster since waiting on one cluster reduces the
 361                 // time needed to wait on the next clusters.
 362                 $masterPositions = array_fill( 0, count( $lbs ), false );
 363                 foreach ( $lbs as $i => $lb ) {
 364                         if ( $lb->getServerCount() <= 1 ) {
 365                                 // Bug 27975 - Don't try to wait for slaves if there are none
 366                                 // Prevents permission error when getting master position
 367                                 continue;
 368                         } elseif ( $opts['ifWritesSince']
 369                                 && $lb->lastMasterChangeTimestamp() < $opts['ifWritesSince']
 370                         ) {
 371                                 continue; // no writes since the last wait
 372                         }
 373                         $masterPositions[$i] = $lb->getMasterPos();
 374                 }
 375
 376                 $failed = [];
 377                 foreach ( $lbs as $i => $lb ) {
 378                         if ( $masterPositions[$i] ) {
 379                                 // The DBMS may not support getMasterPos() or the whole
 380                                 // load balancer might be fake (e.g. $wgAllDBsAreLocalhost).
 381                                 if ( !$lb->waitForAll( $masterPositions[$i], $opts['timeout'] ) ) {
 382                                         $failed[] = $lb->getServerName( $lb->getWriterIndex() );
 383                                 }
 384                         }
 385                 }
 386
 387                 if ( $failed ) {
 388                         throw new DBReplicationWaitError(
 389                                 "Could not wait for slaves to catch up to " .
 390                                 implode( ', ', $failed )
 391                         );
 392                 }
 393         }
 394
 395         /**
 396          * Disable the ChronologyProtector for all load balancers
 397          *
 398          * This can be called at the start of special API entry points
 399          *
 400          * @since 1.27
 401          */
 402         public function disableChronologyProtection() {
 403                 $this->chronProt->setEnabled( false );
 404         }
 405
 406         /**
 407          * @return ChronologyProtector
 408          */
 409         protected function newChronologyProtector() {
 410                 $request = RequestContext::getMain()->getRequest();
 411                 $chronProt = new ChronologyProtector(
 412                         ObjectCache::getMainStashInstance(),
 413                         [
 414                                 'ip' => $request->getIP(),
 415                                 'agent' => $request->getHeader( 'User-Agent' )
 416                         ]
 417                 );
 418                 if ( PHP_SAPI === 'cli' ) {
 419                         $chronProt->setEnabled( false );
 420                 } elseif ( $request->getHeader( 'ChronologyProtection' ) === 'false' ) {
 421                         // Request opted out of using position wait logic. This is useful for requests
 422                         // done by the job queue or background ETL that do not have a meaningful session.
 423                         $chronProt->setWaitEnabled( false );
 424                 }
 425
 426                 return $chronProt;
 427         }
 428
 429         /**
 430          * @param ChronologyProtector $cp
 431          */
 432         protected function shutdownChronologyProtector( ChronologyProtector $cp ) {
 433                 // Get all the master positions needed
 434                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $cp ) {
 435                         $cp->shutdownLB( $lb );
 436                 } );
 437                 // Write them to the stash
 438                 $unsavedPositions = $cp->shutdown();
 439                 // If the positions failed to write to the stash, at least wait on local datacenter
 440                 // slaves to catch up before responding. Even if there are several DCs, this increases
 441                 // the chance that the user will see their own changes immediately afterwards. As long
 442                 // as the sticky DC cookie applies (same domain), this is not even an issue.
 443                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $unsavedPositions ) {
 444                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 445                         if ( isset( $unsavedPositions[$masterName] ) ) {
 446                                 $lb->waitForAll( $unsavedPositions[$masterName] );
 447                         }
 448                 } );
 449         }
 450 }
 451
 452 /**
 453  * Exception class for attempted DB access
 454  */
 455 class DBAccessError extends MWException {
 456         public function __construct() {
 457                 parent::__construct( 'The storage backend is disabled!' );
 458         }
 459 }
 460
 461 /**
 462  * Exception class for replica DB wait timeouts
 463  */
 464 class DBReplicationWaitError extends Exception {
 465 }