includes/libs/rdbms/lbfactory/LBFactory.php

   1 <?php
   2 /**
   3  * Generator and manager 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 namespace Wikimedia\Rdbms;
  25
  26 use Psr\Log\LoggerInterface;
  27 use Wikimedia\ScopedCallback;
  28 use BagOStuff;
  29 use EmptyBagOStuff;
  30 use WANObjectCache;
  31 use Exception;
  32 use RuntimeException;
  33 use DBTransactionError;
  34 use DBReplicationWaitError;
  35
  36 /**
  37  * An interface for generating database load balancers
  38  * @ingroup Database
  39  */
  40 abstract class LBFactory implements ILBFactory {
  41         /** @var ChronologyProtector */
  42         protected $chronProt;
  43         /** @var object|string Class name or object With profileIn/profileOut methods */
  44         protected $profiler;
  45         /** @var TransactionProfiler */
  46         protected $trxProfiler;
  47         /** @var LoggerInterface */
  48         protected $replLogger;
  49         /** @var LoggerInterface */
  50         protected $connLogger;
  51         /** @var LoggerInterface */
  52         protected $queryLogger;
  53         /** @var LoggerInterface */
  54         protected $perfLogger;
  55         /** @var callable Error logger */
  56         protected $errorLogger;
  57         /** @var BagOStuff */
  58         protected $srvCache;
  59         /** @var BagOStuff */
  60         protected $memCache;
  61         /** @var WANObjectCache */
  62         protected $wanCache;
  63
  64         /** @var DatabaseDomain Local domain */
  65         protected $localDomain;
  66         /** @var string Local hostname of the app server */
  67         protected $hostname;
  68         /** @var array Web request information about the client */
  69         protected $requestInfo;
  70
  71         /** @var mixed */
  72         protected $ticket;
  73         /** @var string|bool String if a requested DBO_TRX transaction round is active */
  74         protected $trxRoundId = false;
  75         /** @var string|bool Reason all LBs are read-only or false if not */
  76         protected $readOnlyReason = false;
  77         /** @var callable[] */
  78         protected $replicationWaitCallbacks = [];
  79
  80         /** @var bool Whether this PHP instance is for a CLI script */
  81         protected $cliMode;
  82         /** @var string Agent name for query profiling */
  83         protected $agent;
  84
  85         private static $loggerFields =
  86                 [ 'replLogger', 'connLogger', 'queryLogger', 'perfLogger' ];
  87
  88         public function __construct( array $conf ) {
  89                 $this->localDomain = isset( $conf['localDomain'] )
  90                         ? DatabaseDomain::newFromId( $conf['localDomain'] )
  91                         : DatabaseDomain::newUnspecified();
  92
  93                 if ( isset( $conf['readOnlyReason'] ) && is_string( $conf['readOnlyReason'] ) ) {
  94                         $this->readOnlyReason = $conf['readOnlyReason'];
  95                 }
  96
  97                 $this->srvCache = isset( $conf['srvCache'] ) ? $conf['srvCache'] : new EmptyBagOStuff();
  98                 $this->memCache = isset( $conf['memCache'] ) ? $conf['memCache'] : new EmptyBagOStuff();
  99                 $this->wanCache = isset( $conf['wanCache'] )
 100                         ? $conf['wanCache']
 101                         : WANObjectCache::newEmpty();
 102
 103                 foreach ( self::$loggerFields as $key ) {
 104                         $this->$key = isset( $conf[$key] ) ? $conf[$key] : new \Psr\Log\NullLogger();
 105                 }
 106                 $this->errorLogger = isset( $conf['errorLogger'] )
 107                         ? $conf['errorLogger']
 108                         : function ( Exception $e ) {
 109                                 trigger_error( E_USER_WARNING, get_class( $e ) . ': ' . $e->getMessage() );
 110                         };
 111
 112                 $this->profiler = isset( $conf['profiler'] ) ? $conf['profiler'] : null;
 113                 $this->trxProfiler = isset( $conf['trxProfiler'] )
 114                         ? $conf['trxProfiler']
 115                         : new TransactionProfiler();
 116
 117                 $this->requestInfo = [
 118                         'IPAddress' => isset( $_SERVER[ 'REMOTE_ADDR' ] ) ? $_SERVER[ 'REMOTE_ADDR' ] : '',
 119                         'UserAgent' => isset( $_SERVER['HTTP_USER_AGENT'] ) ? $_SERVER['HTTP_USER_AGENT'] : '',
 120                         'ChronologyProtection' => 'true'
 121                 ];
 122
 123                 $this->cliMode = isset( $conf['cliMode'] ) ? $conf['cliMode'] : PHP_SAPI === 'cli';
 124                 $this->hostname = isset( $conf['hostname'] ) ? $conf['hostname'] : gethostname();
 125                 $this->agent = isset( $conf['agent'] ) ? $conf['agent'] : '';
 126
 127                 $this->ticket = mt_rand();
 128         }
 129
 130         public function destroy() {
 131                 $this->shutdown( self::SHUTDOWN_NO_CHRONPROT );
 132                 $this->forEachLBCallMethod( 'disable' );
 133         }
 134
 135         public function shutdown(
 136                 $mode = self::SHUTDOWN_CHRONPROT_SYNC, callable $workCallback = null
 137         ) {
 138                 $chronProt = $this->getChronologyProtector();
 139                 if ( $mode === self::SHUTDOWN_CHRONPROT_SYNC ) {
 140                         $this->shutdownChronologyProtector( $chronProt, $workCallback, 'sync' );
 141                 } elseif ( $mode === self::SHUTDOWN_CHRONPROT_ASYNC ) {
 142                         $this->shutdownChronologyProtector( $chronProt, null, 'async' );
 143                 }
 144
 145                 $this->commitMasterChanges( __METHOD__ ); // sanity
 146         }
 147
 148         /**
 149          * @see ILBFactory::newMainLB()
 150          * @param bool $domain
 151          * @return LoadBalancer
 152          */
 153         abstract public function newMainLB( $domain = false );
 154
 155         /**
 156          * @see ILBFactory::getMainLB()
 157          * @param bool $domain
 158          * @return LoadBalancer
 159          */
 160         abstract public function getMainLB( $domain = false );
 161
 162         /**
 163          * @see ILBFactory::newExternalLB()
 164          * @param string $cluster
 165          * @return LoadBalancer
 166          */
 167         abstract public function newExternalLB( $cluster );
 168
 169         /**
 170          * @see ILBFactory::getExternalLB()
 171          * @param string $cluster
 172          * @return LoadBalancer
 173          */
 174         abstract public function getExternalLB( $cluster );
 175
 176         /**
 177          * Call a method of each tracked load balancer
 178          *
 179          * @param string $methodName
 180          * @param array $args
 181          */
 182         protected function forEachLBCallMethod( $methodName, array $args = [] ) {
 183                 $this->forEachLB(
 184                         function ( ILoadBalancer $loadBalancer, $methodName, array $args ) {
 185                                 call_user_func_array( [ $loadBalancer, $methodName ], $args );
 186                         },
 187                         [ $methodName, $args ]
 188                 );
 189         }
 190
 191         public function flushReplicaSnapshots( $fname = __METHOD__ ) {
 192                 $this->forEachLBCallMethod( 'flushReplicaSnapshots', [ $fname ] );
 193         }
 194
 195         public function commitAll( $fname = __METHOD__, array $options = [] ) {
 196                 $this->commitMasterChanges( $fname, $options );
 197                 $this->forEachLBCallMethod( 'commitAll', [ $fname ] );
 198         }
 199
 200         public function beginMasterChanges( $fname = __METHOD__ ) {
 201                 if ( $this->trxRoundId !== false ) {
 202                         throw new DBTransactionError(
 203                                 null,
 204                                 "$fname: transaction round '{$this->trxRoundId}' already started."
 205                         );
 206                 }
 207                 $this->trxRoundId = $fname;
 208                 // Set DBO_TRX flags on all appropriate DBs
 209                 $this->forEachLBCallMethod( 'beginMasterChanges', [ $fname ] );
 210         }
 211
 212         public function commitMasterChanges( $fname = __METHOD__, array $options = [] ) {
 213                 if ( $this->trxRoundId !== false && $this->trxRoundId !== $fname ) {
 214                         throw new DBTransactionError(
 215                                 null,
 216                                 "$fname: transaction round '{$this->trxRoundId}' still running."
 217                         );
 218                 }
 219                 /** @noinspection PhpUnusedLocalVariableInspection */
 220                 $scope = $this->getScopedPHPBehaviorForCommit(); // try to ignore client aborts
 221                 // Run pre-commit callbacks and suppress post-commit callbacks, aborting on failure
 222                 $this->forEachLBCallMethod( 'finalizeMasterChanges' );
 223                 $this->trxRoundId = false;
 224                 // Perform 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 and revert DBO_TRX
 229                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 230                 // Run all post-commit callbacks
 231                 /** @var Exception $e */
 232                 $e = null; // first callback exception
 233                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$e ) {
 234                         $ex = $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_COMMIT );
 235                         $e = $e ?: $ex;
 236                 } );
 237                 // Commit any dangling DBO_TRX transactions from callbacks on one DB to another DB
 238                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 239                 // Throw any last post-commit callback error
 240                 if ( $e instanceof Exception ) {
 241                         throw $e;
 242                 }
 243         }
 244
 245         public function rollbackMasterChanges( $fname = __METHOD__ ) {
 246                 $this->trxRoundId = false;
 247                 $this->forEachLBCallMethod( 'suppressTransactionEndCallbacks' );
 248                 $this->forEachLBCallMethod( 'rollbackMasterChanges', [ $fname ] );
 249                 // Run all post-rollback callbacks
 250                 $this->forEachLB( function ( ILoadBalancer $lb ) {
 251                         $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_ROLLBACK );
 252                 } );
 253         }
 254
 255         /**
 256          * Log query info if multi DB transactions are going to be committed now
 257          */
 258         private function logIfMultiDbTransaction() {
 259                 $callersByDB = [];
 260                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$callersByDB ) {
 261                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 262                         $callers = $lb->pendingMasterChangeCallers();
 263                         if ( $callers ) {
 264                                 $callersByDB[$masterName] = $callers;
 265                         }
 266                 } );
 267
 268                 if ( count( $callersByDB ) >= 2 ) {
 269                         $dbs = implode( ', ', array_keys( $callersByDB ) );
 270                         $msg = "Multi-DB transaction [{$dbs}]:\n";
 271                         foreach ( $callersByDB as $db => $callers ) {
 272                                 $msg .= "$db: " . implode( '; ', $callers ) . "\n";
 273                         }
 274                         $this->queryLogger->info( $msg );
 275                 }
 276         }
 277
 278         public function hasMasterChanges() {
 279                 $ret = false;
 280                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$ret ) {
 281                         $ret = $ret || $lb->hasMasterChanges();
 282                 } );
 283
 284                 return $ret;
 285         }
 286
 287         public function laggedReplicaUsed() {
 288                 $ret = false;
 289                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$ret ) {
 290                         $ret = $ret || $lb->laggedReplicaUsed();
 291                 } );
 292
 293                 return $ret;
 294         }
 295
 296         public function hasOrMadeRecentMasterChanges( $age = null ) {
 297                 $ret = false;
 298                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $age, &$ret ) {
 299                         $ret = $ret || $lb->hasOrMadeRecentMasterChanges( $age );
 300                 } );
 301                 return $ret;
 302         }
 303
 304         public function waitForReplication( array $opts = [] ) {
 305                 $opts += [
 306                         'domain' => false,
 307                         'cluster' => false,
 308                         'timeout' => 60,
 309                         'ifWritesSince' => null
 310                 ];
 311
 312                 if ( $opts['domain'] === false && isset( $opts['wiki'] ) ) {
 313                         $opts['domain'] = $opts['wiki']; // b/c
 314                 }
 315
 316                 // Figure out which clusters need to be checked
 317                 /** @var ILoadBalancer[] $lbs */
 318                 $lbs = [];
 319                 if ( $opts['cluster'] !== false ) {
 320                         $lbs[] = $this->getExternalLB( $opts['cluster'] );
 321                 } elseif ( $opts['domain'] !== false ) {
 322                         $lbs[] = $this->getMainLB( $opts['domain'] );
 323                 } else {
 324                         $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$lbs ) {
 325                                 $lbs[] = $lb;
 326                         } );
 327                         if ( !$lbs ) {
 328                                 return; // nothing actually used
 329                         }
 330                 }
 331
 332                 // Get all the master positions of applicable DBs right now.
 333                 // This can be faster since waiting on one cluster reduces the
 334                 // time needed to wait on the next clusters.
 335                 $masterPositions = array_fill( 0, count( $lbs ), false );
 336                 foreach ( $lbs as $i => $lb ) {
 337                         if ( $lb->getServerCount() <= 1 ) {
 338                                 // T29975 - Don't try to wait for replica DBs if there are none
 339                                 // Prevents permission error when getting master position
 340                                 continue;
 341                         } elseif ( $opts['ifWritesSince']
 342                                 && $lb->lastMasterChangeTimestamp() < $opts['ifWritesSince']
 343                         ) {
 344                                 continue; // no writes since the last wait
 345                         }
 346                         $masterPositions[$i] = $lb->getMasterPos();
 347                 }
 348
 349                 // Run any listener callbacks *after* getting the DB positions. The more
 350                 // time spent in the callbacks, the less time is spent in waitForAll().
 351                 foreach ( $this->replicationWaitCallbacks as $callback ) {
 352                         $callback();
 353                 }
 354
 355                 $failed = [];
 356                 foreach ( $lbs as $i => $lb ) {
 357                         if ( $masterPositions[$i] ) {
 358                                 // The DBMS may not support getMasterPos()
 359                                 if ( !$lb->waitForAll( $masterPositions[$i], $opts['timeout'] ) ) {
 360                                         $failed[] = $lb->getServerName( $lb->getWriterIndex() );
 361                                 }
 362                         }
 363                 }
 364
 365                 if ( $failed ) {
 366                         throw new DBReplicationWaitError(
 367                                 null,
 368                                 "Could not wait for replica DBs to catch up to " .
 369                                 implode( ', ', $failed )
 370                         );
 371                 }
 372         }
 373
 374         public function setWaitForReplicationListener( $name, callable $callback = null ) {
 375                 if ( $callback ) {
 376                         $this->replicationWaitCallbacks[$name] = $callback;
 377                 } else {
 378                         unset( $this->replicationWaitCallbacks[$name] );
 379                 }
 380         }
 381
 382         public function getEmptyTransactionTicket( $fname ) {
 383                 if ( $this->hasMasterChanges() ) {
 384                         $this->queryLogger->error( __METHOD__ . ": $fname does not have outer scope.\n" .
 385                                 ( new RuntimeException() )->getTraceAsString() );
 386
 387                         return null;
 388                 }
 389
 390                 return $this->ticket;
 391         }
 392
 393         public function commitAndWaitForReplication( $fname, $ticket, array $opts = [] ) {
 394                 if ( $ticket !== $this->ticket ) {
 395                         $this->perfLogger->error( __METHOD__ . ": $fname does not have outer scope.\n" .
 396                                 ( new RuntimeException() )->getTraceAsString() );
 397
 398                         return;
 399                 }
 400
 401                 // The transaction owner and any caller with the empty transaction ticket can commit
 402                 // so that getEmptyTransactionTicket() callers don't risk seeing DBTransactionError.
 403                 if ( $this->trxRoundId !== false && $fname !== $this->trxRoundId ) {
 404                         $this->queryLogger->info( "$fname: committing on behalf of {$this->trxRoundId}." );
 405                         $fnameEffective = $this->trxRoundId;
 406                 } else {
 407                         $fnameEffective = $fname;
 408                 }
 409
 410                 $this->commitMasterChanges( $fnameEffective );
 411                 $this->waitForReplication( $opts );
 412                 // If a nested caller committed on behalf of $fname, start another empty $fname
 413                 // transaction, leaving the caller with the same empty transaction state as before.
 414                 if ( $fnameEffective !== $fname ) {
 415                         $this->beginMasterChanges( $fnameEffective );
 416                 }
 417         }
 418
 419         public function getChronologyProtectorTouched( $dbName ) {
 420                 return $this->getChronologyProtector()->getTouched( $dbName );
 421         }
 422
 423         public function disableChronologyProtection() {
 424                 $this->getChronologyProtector()->setEnabled( false );
 425         }
 426
 427         /**
 428          * @return ChronologyProtector
 429          */
 430         protected function getChronologyProtector() {
 431                 if ( $this->chronProt ) {
 432                         return $this->chronProt;
 433                 }
 434
 435                 $this->chronProt = new ChronologyProtector(
 436                         $this->memCache,
 437                         [
 438                                 'ip' => $this->requestInfo['IPAddress'],
 439                                 'agent' => $this->requestInfo['UserAgent'],
 440                         ],
 441                         isset( $_GET['cpPosTime'] ) ? $_GET['cpPosTime'] : null
 442                 );
 443                 $this->chronProt->setLogger( $this->replLogger );
 444
 445                 if ( $this->cliMode ) {
 446                         $this->chronProt->setEnabled( false );
 447                 } elseif ( $this->requestInfo['ChronologyProtection'] === 'false' ) {
 448                         // Request opted out of using position wait logic. This is useful for requests
 449                         // done by the job queue or background ETL that do not have a meaningful session.
 450                         $this->chronProt->setWaitEnabled( false );
 451                 }
 452
 453                 $this->replLogger->debug( __METHOD__ . ': using request info ' .
 454                         json_encode( $this->requestInfo, JSON_PRETTY_PRINT ) );
 455
 456                 return $this->chronProt;
 457         }
 458
 459         /**
 460          * Get and record all of the staged DB positions into persistent memory storage
 461          *
 462          * @param ChronologyProtector $cp
 463          * @param callable|null $workCallback Work to do instead of waiting on syncing positions
 464          * @param string $mode One of (sync, async); whether to wait on remote datacenters
 465          */
 466         protected function shutdownChronologyProtector(
 467                 ChronologyProtector $cp, $workCallback, $mode
 468         ) {
 469                 // Record all the master positions needed
 470                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $cp ) {
 471                         $cp->shutdownLB( $lb );
 472                 } );
 473                 // Write them to the persistent stash. Try to do something useful by running $work
 474                 // while ChronologyProtector waits for the stash write to replicate to all DCs.
 475                 $unsavedPositions = $cp->shutdown( $workCallback, $mode );
 476                 if ( $unsavedPositions && $workCallback ) {
 477                         // Invoke callback in case it did not cache the result yet
 478                         $workCallback(); // work now to block for less time in waitForAll()
 479                 }
 480                 // If the positions failed to write to the stash, at least wait on local datacenter
 481                 // replica DBs to catch up before responding. Even if there are several DCs, this increases
 482                 // the chance that the user will see their own changes immediately afterwards. As long
 483                 // as the sticky DC cookie applies (same domain), this is not even an issue.
 484                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $unsavedPositions ) {
 485                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 486                         if ( isset( $unsavedPositions[$masterName] ) ) {
 487                                 $lb->waitForAll( $unsavedPositions[$masterName] );
 488                         }
 489                 } );
 490         }
 491
 492         /**
 493          * Base parameters to LoadBalancer::__construct()
 494          * @return array
 495          */
 496         final protected function baseLoadBalancerParams() {
 497                 return [
 498                         'localDomain' => $this->localDomain,
 499                         'readOnlyReason' => $this->readOnlyReason,
 500                         'srvCache' => $this->srvCache,
 501                         'wanCache' => $this->wanCache,
 502                         'profiler' => $this->profiler,
 503                         'trxProfiler' => $this->trxProfiler,
 504                         'queryLogger' => $this->queryLogger,
 505                         'connLogger' => $this->connLogger,
 506                         'replLogger' => $this->replLogger,
 507                         'errorLogger' => $this->errorLogger,
 508                         'hostname' => $this->hostname,
 509                         'cliMode' => $this->cliMode,
 510                         'agent' => $this->agent
 511                 ];
 512         }
 513
 514         /**
 515          * @param ILoadBalancer $lb
 516          */
 517         protected function initLoadBalancer( ILoadBalancer $lb ) {
 518                 if ( $this->trxRoundId !== false ) {
 519                         $lb->beginMasterChanges( $this->trxRoundId ); // set DBO_TRX
 520                 }
 521         }
 522
 523         public function setDomainPrefix( $prefix ) {
 524                 $this->localDomain = new DatabaseDomain(
 525                         $this->localDomain->getDatabase(),
 526                         null,
 527                         $prefix
 528                 );
 529
 530                 $this->forEachLB( function( ILoadBalancer $lb ) use ( $prefix ) {
 531                         $lb->setDomainPrefix( $prefix );
 532                 } );
 533         }
 534
 535         public function closeAll() {
 536                 $this->forEachLBCallMethod( 'closeAll', [] );
 537         }
 538
 539         public function setAgentName( $agent ) {
 540                 $this->agent = $agent;
 541         }
 542
 543         public function appendPreShutdownTimeAsQuery( $url, $time ) {
 544                 $usedCluster = 0;
 545                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$usedCluster ) {
 546                         $usedCluster |= ( $lb->getServerCount() > 1 );
 547                 } );
 548
 549                 if ( !$usedCluster ) {
 550                         return $url; // no master/replica clusters touched
 551                 }
 552
 553                 return strpos( $url, '?' ) === false ? "$url?cpPosTime=$time" : "$url&cpPosTime=$time";
 554         }
 555
 556         public function setRequestInfo( array $info ) {
 557                 $this->requestInfo = $info + $this->requestInfo;
 558         }
 559
 560         /**
 561          * Make PHP ignore user aborts/disconnects until the returned
 562          * value leaves scope. This returns null and does nothing in CLI mode.
 563          *
 564          * @return ScopedCallback|null
 565          */
 566         final protected function getScopedPHPBehaviorForCommit() {
 567                 if ( PHP_SAPI != 'cli' ) { // https://bugs.php.net/bug.php?id=47540
 568                         $old = ignore_user_abort( true ); // avoid half-finished operations
 569                         return new ScopedCallback( function () use ( $old ) {
 570                                 ignore_user_abort( $old );
 571                         } );
 572                 }
 573
 574                 return null;
 575         }
 576
 577         function __destruct() {
 578                 $this->destroy();
 579         }
 580 }