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 Psr\Log\LoggerInterface;
  25 use MediaWiki\Logger\LoggerFactory;
  26
  27 /**
  28  * An interface for generating database load balancers
  29  * @ingroup Database
  30  */
  31 abstract class LBFactory {
  32         /** @var ChronologyProtector */
  33         protected $chronProt;
  34
  35         /** @var TransactionProfiler */
  36         protected $trxProfiler;
  37
  38         /** @var LoggerInterface */
  39         protected $logger;
  40
  41         /** @var LBFactory */
  42         private static $instance;
  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 access to the load balancer, will cause all database access
  65          * to throw a DBAccessError
  66          */
  67         public static function disableBackend() {
  68                 global $wgLBFactoryConf;
  69                 self::$instance = new LBFactoryFake( $wgLBFactoryConf );
  70         }
  71
  72         /**
  73          * Get an LBFactory instance
  74          *
  75          * @return LBFactory
  76          */
  77         public static function singleton() {
  78                 global $wgLBFactoryConf;
  79
  80                 if ( is_null( self::$instance ) ) {
  81                         $class = self::getLBFactoryClass( $wgLBFactoryConf );
  82                         $config = $wgLBFactoryConf;
  83                         if ( !isset( $config['readOnlyReason'] ) ) {
  84                                 $config['readOnlyReason'] = wfConfiguredReadOnlyReason();
  85                         }
  86                         self::$instance = new $class( $config );
  87                 }
  88
  89                 return self::$instance;
  90         }
  91
  92         /**
  93          * Returns the LBFactory class to use and the load balancer configuration.
  94          *
  95          * @param array $config (e.g. $wgLBFactoryConf)
  96          * @return string Class name
  97          */
  98         public static function getLBFactoryClass( array $config ) {
  99                 // For configuration backward compatibility after removing
 100                 // underscores from class names in MediaWiki 1.23.
 101                 $bcClasses = [
 102                         'LBFactory_Simple' => 'LBFactorySimple',
 103                         'LBFactory_Single' => 'LBFactorySingle',
 104                         'LBFactory_Multi' => 'LBFactoryMulti',
 105                         'LBFactory_Fake' => 'LBFactoryFake',
 106                 ];
 107
 108                 $class = $config['class'];
 109
 110                 if ( isset( $bcClasses[$class] ) ) {
 111                         $class = $bcClasses[$class];
 112                         wfDeprecated(
 113                                 '$wgLBFactoryConf must be updated. See RELEASE-NOTES for details',
 114                                 '1.23'
 115                         );
 116                 }
 117
 118                 return $class;
 119         }
 120
 121         /**
 122          * Shut down, close connections and destroy the cached instance.
 123          */
 124         public static function destroyInstance() {
 125                 if ( self::$instance ) {
 126                         self::$instance->shutdown();
 127                         self::$instance->forEachLBCallMethod( 'closeAll' );
 128                         self::$instance = null;
 129                 }
 130         }
 131
 132         /**
 133          * Set the instance to be the given object
 134          *
 135          * @param LBFactory $instance
 136          */
 137         public static function setInstance( $instance ) {
 138                 self::destroyInstance();
 139                 self::$instance = $instance;
 140         }
 141
 142         /**
 143          * Create a new load balancer object. The resulting object will be untracked,
 144          * not chronology-protected, and the caller is responsible for cleaning it up.
 145          *
 146          * @param bool|string $wiki Wiki ID, or false for the current wiki
 147          * @return LoadBalancer
 148          */
 149         abstract public function newMainLB( $wiki = false );
 150
 151         /**
 152          * Get a cached (tracked) load balancer object.
 153          *
 154          * @param bool|string $wiki Wiki ID, or false for the current wiki
 155          * @return LoadBalancer
 156          */
 157         abstract public function getMainLB( $wiki = false );
 158
 159         /**
 160          * Create a new load balancer for external storage. The resulting object will be
 161          * untracked, not chronology-protected, and the caller is responsible for
 162          * cleaning it up.
 163          *
 164          * @param string $cluster External storage cluster, or false for core
 165          * @param bool|string $wiki Wiki ID, or false for the current wiki
 166          * @return LoadBalancer
 167          */
 168         abstract protected function newExternalLB( $cluster, $wiki = false );
 169
 170         /**
 171          * Get a cached (tracked) load balancer for external storage
 172          *
 173          * @param string $cluster External storage cluster, or false for core
 174          * @param bool|string $wiki Wiki ID, or false for the current wiki
 175          * @return LoadBalancer
 176          */
 177         abstract public function &getExternalLB( $cluster, $wiki = false );
 178
 179         /**
 180          * Execute a function for each tracked load balancer
 181          * The callback is called with the load balancer as the first parameter,
 182          * and $params passed as the subsequent parameters.
 183          *
 184          * @param callable $callback
 185          * @param array $params
 186          */
 187         abstract public function forEachLB( $callback, array $params = [] );
 188
 189         /**
 190          * Prepare all tracked load balancers for shutdown
 191          * @param integer $flags Supports SHUTDOWN_* flags
 192          * STUB
 193          */
 194         public function shutdown( $flags = 0 ) {
 195         }
 196
 197         /**
 198          * Call a method of each tracked load balancer
 199          *
 200          * @param string $methodName
 201          * @param array $args
 202          */
 203         private function forEachLBCallMethod( $methodName, array $args = [] ) {
 204                 $this->forEachLB(
 205                         function ( LoadBalancer $loadBalancer, $methodName, array $args ) {
 206                                 call_user_func_array( [ $loadBalancer, $methodName ], $args );
 207                         },
 208                         [ $methodName, $args ]
 209                 );
 210         }
 211
 212         /**
 213          * Commit on all connections. Done for two reasons:
 214          * 1. To commit changes to the masters.
 215          * 2. To release the snapshot on all connections, master and slave.
 216          * @param string $fname Caller name
 217          */
 218         public function commitAll( $fname = __METHOD__ ) {
 219                 $this->logMultiDbTransaction();
 220
 221                 $start = microtime( true );
 222                 $this->forEachLBCallMethod( 'commitAll', [ $fname ] );
 223                 $timeMs = 1000 * ( microtime( true ) - $start );
 224
 225                 RequestContext::getMain()->getStats()->timing( "db.commit-all", $timeMs );
 226         }
 227
 228         /**
 229          * Commit changes on all master connections
 230          * @param string $fname Caller name
 231          * @param array $options Options map:
 232          *   - maxWriteDuration: abort if more than this much time was spent in write queries
 233          */
 234         public function commitMasterChanges( $fname = __METHOD__, array $options = [] ) {
 235                 $limit = isset( $options['maxWriteDuration'] ) ? $options['maxWriteDuration'] : 0;
 236
 237                 $this->logMultiDbTransaction();
 238                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $limit ) {
 239                         $lb->forEachOpenConnection( function ( IDatabase $db ) use ( $limit ) {
 240                                 $time = $db->pendingWriteQueryDuration();
 241                                 if ( $limit > 0 && $time > $limit ) {
 242                                         throw new DBTransactionError(
 243                                                 $db,
 244                                                 wfMessage( 'transaction-duration-limit-exceeded', $time, $limit )->text()
 245                                         );
 246                                 }
 247                         } );
 248                 } );
 249
 250                 $start = microtime( true );
 251                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 252                 $timeMs = 1000 * ( microtime( true ) - $start );
 253
 254                 RequestContext::getMain()->getStats()->timing( "db.commit-masters", $timeMs );
 255         }
 256
 257         /**
 258          * Rollback changes on all master connections
 259          * @param string $fname Caller name
 260          * @since 1.23
 261          */
 262         public function rollbackMasterChanges( $fname = __METHOD__ ) {
 263                 $this->forEachLBCallMethod( 'rollbackMasterChanges', [ $fname ] );
 264         }
 265
 266         /**
 267          * Log query info if multi DB transactions are going to be committed now
 268          */
 269         private function logMultiDbTransaction() {
 270                 $callersByDB = [];
 271                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$callersByDB ) {
 272                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 273                         $callers = $lb->pendingMasterChangeCallers();
 274                         if ( $callers ) {
 275                                 $callersByDB[$masterName] = $callers;
 276                         }
 277                 } );
 278
 279                 if ( count( $callersByDB ) >= 2 ) {
 280                         $dbs = implode( ', ', array_keys( $callersByDB ) );
 281                         $msg = "Multi-DB transaction [{$dbs}]:\n";
 282                         foreach ( $callersByDB as $db => $callers ) {
 283                                 $msg .= "$db: " . implode( '; ', $callers ) . "\n";
 284                         }
 285                         $this->logger->info( $msg );
 286                 }
 287         }
 288
 289         /**
 290          * Determine if any master connection has pending changes
 291          * @return bool
 292          * @since 1.23
 293          */
 294         public function hasMasterChanges() {
 295                 $ret = false;
 296                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 297                         $ret = $ret || $lb->hasMasterChanges();
 298                 } );
 299
 300                 return $ret;
 301         }
 302
 303         /**
 304          * Detemine if any lagged slave connection was used
 305          * @since 1.27
 306          * @return bool
 307          */
 308         public function laggedSlaveUsed() {
 309                 $ret = false;
 310                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 311                         $ret = $ret || $lb->laggedSlaveUsed();
 312                 } );
 313
 314                 return $ret;
 315         }
 316
 317         /**
 318          * Determine if any master connection has pending/written changes from this request
 319          * @return bool
 320          * @since 1.27
 321          */
 322         public function hasOrMadeRecentMasterChanges() {
 323                 $ret = false;
 324                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 325                         $ret = $ret || $lb->hasOrMadeRecentMasterChanges();
 326                 } );
 327                 return $ret;
 328         }
 329
 330         /**
 331          * Waits for the slave DBs to catch up to the current master position
 332          *
 333          * Use this when updating very large numbers of rows, as in maintenance scripts,
 334          * to avoid causing too much lag. Of course, this is a no-op if there are no slaves.
 335          *
 336          * By default this waits on all DB clusters actually used in this request.
 337          * This makes sense when lag being waiting on is caused by the code that does this check.
 338          * In that case, setting "ifWritesSince" can avoid the overhead of waiting for clusters
 339          * that were not changed since the last wait check. To forcefully wait on a specific cluster
 340          * for a given wiki, use the 'wiki' parameter. To forcefully wait on an "external" cluster,
 341          * use the "cluster" parameter.
 342          *
 343          * Never call this function after a large DB write that is *still* in a transaction.
 344          * It only makes sense to call this after the possible lag inducing changes were committed.
 345          *
 346          * @param array $opts Optional fields that include:
 347          *   - wiki : wait on the load balancer DBs that handles the given wiki
 348          *   - cluster : wait on the given external load balancer DBs
 349          *   - timeout : Max wait time. Default: ~60 seconds
 350          *   - ifWritesSince: Only wait if writes were done since this UNIX timestamp
 351          * @throws DBReplicationWaitError If a timeout or error occured waiting on a DB cluster
 352          * @since 1.27
 353          */
 354         public function waitForReplication( array $opts = [] ) {
 355                 $opts += [
 356                         'wiki' => false,
 357                         'cluster' => false,
 358                         'timeout' => 60,
 359                         'ifWritesSince' => null
 360                 ];
 361
 362                 // Figure out which clusters need to be checked
 363                 /** @var LoadBalancer[] $lbs */
 364                 $lbs = [];
 365                 if ( $opts['cluster'] !== false ) {
 366                         $lbs[] = $this->getExternalLB( $opts['cluster'] );
 367                 } elseif ( $opts['wiki'] !== false ) {
 368                         $lbs[] = $this->getMainLB( $opts['wiki'] );
 369                 } else {
 370                         $this->forEachLB( function ( LoadBalancer $lb ) use ( &$lbs ) {
 371                                 $lbs[] = $lb;
 372                         } );
 373                         if ( !$lbs ) {
 374                                 return; // nothing actually used
 375                         }
 376                 }
 377
 378                 // Get all the master positions of applicable DBs right now.
 379                 // This can be faster since waiting on one cluster reduces the
 380                 // time needed to wait on the next clusters.
 381                 $masterPositions = array_fill( 0, count( $lbs ), false );
 382                 foreach ( $lbs as $i => $lb ) {
 383                         if ( $lb->getServerCount() <= 1 ) {
 384                                 // Bug 27975 - Don't try to wait for slaves if there are none
 385                                 // Prevents permission error when getting master position
 386                                 continue;
 387                         } elseif ( $opts['ifWritesSince']
 388                                 && $lb->lastMasterChangeTimestamp() < $opts['ifWritesSince']
 389                         ) {
 390                                 continue; // no writes since the last wait
 391                         }
 392                         $masterPositions[$i] = $lb->getMasterPos();
 393                 }
 394
 395                 $failed = [];
 396                 foreach ( $lbs as $i => $lb ) {
 397                         if ( $masterPositions[$i] ) {
 398                                 // The DBMS may not support getMasterPos() or the whole
 399                                 // load balancer might be fake (e.g. $wgAllDBsAreLocalhost).
 400                                 if ( !$lb->waitForAll( $masterPositions[$i], $opts['timeout'] ) ) {
 401                                         $failed[] = $lb->getServerName( $lb->getWriterIndex() );
 402                                 }
 403                         }
 404                 }
 405
 406                 if ( $failed ) {
 407                         throw new DBReplicationWaitError(
 408                                 "Could not wait for slaves to catch up to " .
 409                                 implode( ', ', $failed )
 410                         );
 411                 }
 412         }
 413
 414         /**
 415          * Disable the ChronologyProtector for all load balancers
 416          *
 417          * This can be called at the start of special API entry points
 418          *
 419          * @since 1.27
 420          */
 421         public function disableChronologyProtection() {
 422                 $this->chronProt->setEnabled( false );
 423         }
 424
 425         /**
 426          * @return ChronologyProtector
 427          */
 428         protected function newChronologyProtector() {
 429                 $request = RequestContext::getMain()->getRequest();
 430                 $chronProt = new ChronologyProtector(
 431                         ObjectCache::getMainStashInstance(),
 432                         [
 433                                 'ip' => $request->getIP(),
 434                                 'agent' => $request->getHeader( 'User-Agent' )
 435                         ]
 436                 );
 437                 if ( PHP_SAPI === 'cli' ) {
 438                         $chronProt->setEnabled( false );
 439                 } elseif ( $request->getHeader( 'ChronologyProtection' ) === 'false' ) {
 440                         // Request opted out of using position wait logic. This is useful for requests
 441                         // done by the job queue or background ETL that do not have a meaningful session.
 442                         $chronProt->setWaitEnabled( false );
 443                 }
 444
 445                 return $chronProt;
 446         }
 447
 448         /**
 449          * @param ChronologyProtector $cp
 450          */
 451         protected function shutdownChronologyProtector( ChronologyProtector $cp ) {
 452                 // Get all the master positions needed
 453                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $cp ) {
 454                         $cp->shutdownLB( $lb );
 455                 } );
 456                 // Write them to the stash
 457                 $unsavedPositions = $cp->shutdown();
 458                 // If the positions failed to write to the stash, at least wait on local datacenter
 459                 // slaves to catch up before responding. Even if there are several DCs, this increases
 460                 // the chance that the user will see their own changes immediately afterwards. As long
 461                 // as the sticky DC cookie applies (same domain), this is not even an issue.
 462                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $unsavedPositions ) {
 463                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 464                         if ( isset( $unsavedPositions[$masterName] ) ) {
 465                                 $lb->waitForAll( $unsavedPositions[$masterName] );
 466                         }
 467                 } );
 468         }
 469 }
 470
 471 /**
 472  * Exception class for attempted DB access
 473  */
 474 class DBAccessError extends MWException {
 475         public function __construct() {
 476                 parent::__construct( "Mediawiki tried to access the database via wfGetDB(). " .
 477                         "This is not allowed." );
 478         }
 479 }
 480
 481 /**
 482  * Exception class for replica DB wait timeouts
 483  */
 484 class DBReplicationWaitError extends Exception {
 485 }