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          * Shut down, close connections and destroy the cached instance.
 125          *
 126          * @deprecated since 1.27, use LBFactory::destroy()
 127          */
 128         public static function destroyInstance() {
 129                 self::singleton()->destroy();
 130         }
 131
 132         /**
 133          * Create a new load balancer object. The resulting object will be untracked,
 134          * not chronology-protected, and the caller is responsible for cleaning it up.
 135          *
 136          * @param bool|string $wiki Wiki ID, or false for the current wiki
 137          * @return LoadBalancer
 138          */
 139         abstract public function newMainLB( $wiki = false );
 140
 141         /**
 142          * Get a cached (tracked) load balancer object.
 143          *
 144          * @param bool|string $wiki Wiki ID, or false for the current wiki
 145          * @return LoadBalancer
 146          */
 147         abstract public function getMainLB( $wiki = false );
 148
 149         /**
 150          * Create a new load balancer for external storage. The resulting object will be
 151          * untracked, not chronology-protected, and the caller is responsible for
 152          * cleaning it up.
 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 protected function newExternalLB( $cluster, $wiki = false );
 159
 160         /**
 161          * Get a cached (tracked) load balancer for external storage
 162          *
 163          * @param string $cluster External storage cluster, or false for core
 164          * @param bool|string $wiki Wiki ID, or false for the current wiki
 165          * @return LoadBalancer
 166          */
 167         abstract public function &getExternalLB( $cluster, $wiki = false );
 168
 169         /**
 170          * Execute a function for each tracked load balancer
 171          * The callback is called with the load balancer as the first parameter,
 172          * and $params passed as the subsequent parameters.
 173          *
 174          * @param callable $callback
 175          * @param array $params
 176          */
 177         abstract public function forEachLB( $callback, array $params = [] );
 178
 179         /**
 180          * Prepare all tracked load balancers for shutdown
 181          * @param integer $flags Supports SHUTDOWN_* flags
 182          * STUB
 183          */
 184         public function shutdown( $flags = 0 ) {
 185         }
 186
 187         /**
 188          * Call a method of each tracked load balancer
 189          *
 190          * @param string $methodName
 191          * @param array $args
 192          */
 193         private function forEachLBCallMethod( $methodName, array $args = [] ) {
 194                 $this->forEachLB(
 195                         function ( LoadBalancer $loadBalancer, $methodName, array $args ) {
 196                                 call_user_func_array( [ $loadBalancer, $methodName ], $args );
 197                         },
 198                         [ $methodName, $args ]
 199                 );
 200         }
 201
 202         /**
 203          * Commit on all connections. Done for two reasons:
 204          * 1. To commit changes to the masters.
 205          * 2. To release the snapshot on all connections, master and slave.
 206          * @param string $fname Caller name
 207          * @param array $options Options map:
 208          *   - maxWriteDuration: abort if more than this much time was spent in write queries
 209          */
 210         public function commitAll( $fname = __METHOD__, array $options = [] ) {
 211                 $this->commitMasterChanges( $fname, $options );
 212                 $this->forEachLBCallMethod( 'commitAll', [ $fname ] );
 213         }
 214
 215         /**
 216          * Commit changes on all master connections
 217          * @param string $fname Caller name
 218          * @param array $options Options map:
 219          *   - maxWriteDuration: abort if more than this much time was spent in write queries
 220          */
 221         public function commitMasterChanges( $fname = __METHOD__, array $options = [] ) {
 222                 // Perform all pre-commit callbacks, aborting on failure
 223                 $this->forEachLBCallMethod( 'runMasterPreCommitCallbacks' );
 224                 // Perform all pre-commit checks, aborting on failure
 225                 $this->forEachLBCallMethod( 'approveMasterChanges', [ $options ] );
 226                 // Log the DBs and methods involved in multi-DB transactions
 227                 $this->logIfMultiDbTransaction();
 228                 // Actually perform the commit on all master DB connections
 229                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 230                 // Run all post-commit callbacks
 231                 $this->forEachLBCallMethod( 'runMasterPostCommitCallbacks' );
 232                 // Commit any dangling DBO_TRX transactions from callbacks on one DB to another DB
 233                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 234         }
 235
 236         /**
 237          * Rollback changes on all master connections
 238          * @param string $fname Caller name
 239          * @since 1.23
 240          */
 241         public function rollbackMasterChanges( $fname = __METHOD__ ) {
 242                 $this->forEachLBCallMethod( 'rollbackMasterChanges', [ $fname ] );
 243         }
 244
 245         /**
 246          * Log query info if multi DB transactions are going to be committed now
 247          */
 248         private function logIfMultiDbTransaction() {
 249                 $callersByDB = [];
 250                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$callersByDB ) {
 251                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 252                         $callers = $lb->pendingMasterChangeCallers();
 253                         if ( $callers ) {
 254                                 $callersByDB[$masterName] = $callers;
 255                         }
 256                 } );
 257
 258                 if ( count( $callersByDB ) >= 2 ) {
 259                         $dbs = implode( ', ', array_keys( $callersByDB ) );
 260                         $msg = "Multi-DB transaction [{$dbs}]:\n";
 261                         foreach ( $callersByDB as $db => $callers ) {
 262                                 $msg .= "$db: " . implode( '; ', $callers ) . "\n";
 263                         }
 264                         $this->logger->info( $msg );
 265                 }
 266         }
 267
 268         /**
 269          * Determine if any master connection has pending changes
 270          * @return bool
 271          * @since 1.23
 272          */
 273         public function hasMasterChanges() {
 274                 $ret = false;
 275                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 276                         $ret = $ret || $lb->hasMasterChanges();
 277                 } );
 278
 279                 return $ret;
 280         }
 281
 282         /**
 283          * Detemine if any lagged slave connection was used
 284          * @since 1.27
 285          * @return bool
 286          */
 287         public function laggedSlaveUsed() {
 288                 $ret = false;
 289                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 290                         $ret = $ret || $lb->laggedSlaveUsed();
 291                 } );
 292
 293                 return $ret;
 294         }
 295
 296         /**
 297          * Determine if any master connection has pending/written changes from this request
 298          * @return bool
 299          * @since 1.27
 300          */
 301         public function hasOrMadeRecentMasterChanges() {
 302                 $ret = false;
 303                 $this->forEachLB( function ( LoadBalancer $lb ) use ( &$ret ) {
 304                         $ret = $ret || $lb->hasOrMadeRecentMasterChanges();
 305                 } );
 306                 return $ret;
 307         }
 308
 309         /**
 310          * Waits for the slave DBs to catch up to the current master position
 311          *
 312          * Use this when updating very large numbers of rows, as in maintenance scripts,
 313          * to avoid causing too much lag. Of course, this is a no-op if there are no slaves.
 314          *
 315          * By default this waits on all DB clusters actually used in this request.
 316          * This makes sense when lag being waiting on is caused by the code that does this check.
 317          * In that case, setting "ifWritesSince" can avoid the overhead of waiting for clusters
 318          * that were not changed since the last wait check. To forcefully wait on a specific cluster
 319          * for a given wiki, use the 'wiki' parameter. To forcefully wait on an "external" cluster,
 320          * use the "cluster" parameter.
 321          *
 322          * Never call this function after a large DB write that is *still* in a transaction.
 323          * It only makes sense to call this after the possible lag inducing changes were committed.
 324          *
 325          * @param array $opts Optional fields that include:
 326          *   - wiki : wait on the load balancer DBs that handles the given wiki
 327          *   - cluster : wait on the given external load balancer DBs
 328          *   - timeout : Max wait time. Default: ~60 seconds
 329          *   - ifWritesSince: Only wait if writes were done since this UNIX timestamp
 330          * @throws DBReplicationWaitError If a timeout or error occured waiting on a DB cluster
 331          * @since 1.27
 332          */
 333         public function waitForReplication( array $opts = [] ) {
 334                 $opts += [
 335                         'wiki' => false,
 336                         'cluster' => false,
 337                         'timeout' => 60,
 338                         'ifWritesSince' => null
 339                 ];
 340
 341                 // Figure out which clusters need to be checked
 342                 /** @var LoadBalancer[] $lbs */
 343                 $lbs = [];
 344                 if ( $opts['cluster'] !== false ) {
 345                         $lbs[] = $this->getExternalLB( $opts['cluster'] );
 346                 } elseif ( $opts['wiki'] !== false ) {
 347                         $lbs[] = $this->getMainLB( $opts['wiki'] );
 348                 } else {
 349                         $this->forEachLB( function ( LoadBalancer $lb ) use ( &$lbs ) {
 350                                 $lbs[] = $lb;
 351                         } );
 352                         if ( !$lbs ) {
 353                                 return; // nothing actually used
 354                         }
 355                 }
 356
 357                 // Get all the master positions of applicable DBs right now.
 358                 // This can be faster since waiting on one cluster reduces the
 359                 // time needed to wait on the next clusters.
 360                 $masterPositions = array_fill( 0, count( $lbs ), false );
 361                 foreach ( $lbs as $i => $lb ) {
 362                         if ( $lb->getServerCount() <= 1 ) {
 363                                 // Bug 27975 - Don't try to wait for slaves if there are none
 364                                 // Prevents permission error when getting master position
 365                                 continue;
 366                         } elseif ( $opts['ifWritesSince']
 367                                 && $lb->lastMasterChangeTimestamp() < $opts['ifWritesSince']
 368                         ) {
 369                                 continue; // no writes since the last wait
 370                         }
 371                         $masterPositions[$i] = $lb->getMasterPos();
 372                 }
 373
 374                 $failed = [];
 375                 foreach ( $lbs as $i => $lb ) {
 376                         if ( $masterPositions[$i] ) {
 377                                 // The DBMS may not support getMasterPos() or the whole
 378                                 // load balancer might be fake (e.g. $wgAllDBsAreLocalhost).
 379                                 if ( !$lb->waitForAll( $masterPositions[$i], $opts['timeout'] ) ) {
 380                                         $failed[] = $lb->getServerName( $lb->getWriterIndex() );
 381                                 }
 382                         }
 383                 }
 384
 385                 if ( $failed ) {
 386                         throw new DBReplicationWaitError(
 387                                 "Could not wait for slaves to catch up to " .
 388                                 implode( ', ', $failed )
 389                         );
 390                 }
 391         }
 392
 393         /**
 394          * Disable the ChronologyProtector for all load balancers
 395          *
 396          * This can be called at the start of special API entry points
 397          *
 398          * @since 1.27
 399          */
 400         public function disableChronologyProtection() {
 401                 $this->chronProt->setEnabled( false );
 402         }
 403
 404         /**
 405          * @return ChronologyProtector
 406          */
 407         protected function newChronologyProtector() {
 408                 $request = RequestContext::getMain()->getRequest();
 409                 $chronProt = new ChronologyProtector(
 410                         ObjectCache::getMainStashInstance(),
 411                         [
 412                                 'ip' => $request->getIP(),
 413                                 'agent' => $request->getHeader( 'User-Agent' )
 414                         ]
 415                 );
 416                 if ( PHP_SAPI === 'cli' ) {
 417                         $chronProt->setEnabled( false );
 418                 } elseif ( $request->getHeader( 'ChronologyProtection' ) === 'false' ) {
 419                         // Request opted out of using position wait logic. This is useful for requests
 420                         // done by the job queue or background ETL that do not have a meaningful session.
 421                         $chronProt->setWaitEnabled( false );
 422                 }
 423
 424                 return $chronProt;
 425         }
 426
 427         /**
 428          * @param ChronologyProtector $cp
 429          */
 430         protected function shutdownChronologyProtector( ChronologyProtector $cp ) {
 431                 // Get all the master positions needed
 432                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $cp ) {
 433                         $cp->shutdownLB( $lb );
 434                 } );
 435                 // Write them to the stash
 436                 $unsavedPositions = $cp->shutdown();
 437                 // If the positions failed to write to the stash, at least wait on local datacenter
 438                 // slaves to catch up before responding. Even if there are several DCs, this increases
 439                 // the chance that the user will see their own changes immediately afterwards. As long
 440                 // as the sticky DC cookie applies (same domain), this is not even an issue.
 441                 $this->forEachLB( function ( LoadBalancer $lb ) use ( $unsavedPositions ) {
 442                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 443                         if ( isset( $unsavedPositions[$masterName] ) ) {
 444                                 $lb->waitForAll( $unsavedPositions[$masterName] );
 445                         }
 446                 } );
 447         }
 448
 449         /**
 450          * Close all open database connections on all open load balancers.
 451          * @since 1.28
 452          */
 453         public function closeAll() {
 454                 $this->forEachLBCallMethod( 'closeAll', [] );
 455         }
 456
 457 }
 458
 459 /**
 460  * Exception class for attempted DB access
 461  */
 462 class DBAccessError extends MWException {
 463         public function __construct() {
 464                 parent::__construct( "Mediawiki tried to access the database via wfGetDB(). " .
 465                         "This is not allowed, because database access has been disabled." );
 466         }
 467 }
 468
 469 /**
 470  * Exception class for replica DB wait timeouts
 471  */
 472 class DBReplicationWaitError extends Exception {
 473 }