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 IDatabase;
  34 use DBTransactionError;
  35 use DBReplicationWaitError;
  36
  37 /**
  38  * An interface for generating database load balancers
  39  * @ingroup Database
  40  */
  41 abstract class LBFactory implements ILBFactory {
  42         /** @var ChronologyProtector */
  43         protected $chronProt;
  44         /** @var object|string Class name or object With profileIn/profileOut methods */
  45         protected $profiler;
  46         /** @var TransactionProfiler */
  47         protected $trxProfiler;
  48         /** @var LoggerInterface */
  49         protected $replLogger;
  50         /** @var LoggerInterface */
  51         protected $connLogger;
  52         /** @var LoggerInterface */
  53         protected $queryLogger;
  54         /** @var LoggerInterface */
  55         protected $perfLogger;
  56         /** @var callable Error logger */
  57         protected $errorLogger;
  58         /** @var BagOStuff */
  59         protected $srvCache;
  60         /** @var BagOStuff */
  61         protected $memCache;
  62         /** @var WANObjectCache */
  63         protected $wanCache;
  64
  65         /** @var DatabaseDomain Local domain */
  66         protected $localDomain;
  67         /** @var string Local hostname of the app server */
  68         protected $hostname;
  69         /** @var array Web request information about the client */
  70         protected $requestInfo;
  71
  72         /** @var mixed */
  73         protected $ticket;
  74         /** @var string|bool String if a requested DBO_TRX transaction round is active */
  75         protected $trxRoundId = false;
  76         /** @var string|bool Reason all LBs are read-only or false if not */
  77         protected $readOnlyReason = false;
  78         /** @var callable[] */
  79         protected $replicationWaitCallbacks = [];
  80
  81         /** @var bool Whether this PHP instance is for a CLI script */
  82         protected $cliMode;
  83         /** @var string Agent name for query profiling */
  84         protected $agent;
  85
  86         private static $loggerFields =
  87                 [ 'replLogger', 'connLogger', 'queryLogger', 'perfLogger' ];
  88
  89         public function __construct( array $conf ) {
  90                 $this->localDomain = isset( $conf['localDomain'] )
  91                         ? DatabaseDomain::newFromId( $conf['localDomain'] )
  92                         : DatabaseDomain::newUnspecified();
  93
  94                 if ( isset( $conf['readOnlyReason'] ) && is_string( $conf['readOnlyReason'] ) ) {
  95                         $this->readOnlyReason = $conf['readOnlyReason'];
  96                 }
  97
  98                 $this->srvCache = isset( $conf['srvCache'] ) ? $conf['srvCache'] : new EmptyBagOStuff();
  99                 $this->memCache = isset( $conf['memCache'] ) ? $conf['memCache'] : new EmptyBagOStuff();
 100                 $this->wanCache = isset( $conf['wanCache'] )
 101                         ? $conf['wanCache']
 102                         : WANObjectCache::newEmpty();
 103
 104                 foreach ( self::$loggerFields as $key ) {
 105                         $this->$key = isset( $conf[$key] ) ? $conf[$key] : new \Psr\Log\NullLogger();
 106                 }
 107                 $this->errorLogger = isset( $conf['errorLogger'] )
 108                         ? $conf['errorLogger']
 109                         : function ( Exception $e ) {
 110                                 trigger_error( E_USER_WARNING, get_class( $e ) . ': ' . $e->getMessage() );
 111                         };
 112
 113                 $this->profiler = isset( $conf['profiler'] ) ? $conf['profiler'] : null;
 114                 $this->trxProfiler = isset( $conf['trxProfiler'] )
 115                         ? $conf['trxProfiler']
 116                         : new TransactionProfiler();
 117
 118                 $this->requestInfo = [
 119                         'IPAddress' => isset( $_SERVER[ 'REMOTE_ADDR' ] ) ? $_SERVER[ 'REMOTE_ADDR' ] : '',
 120                         'UserAgent' => isset( $_SERVER['HTTP_USER_AGENT'] ) ? $_SERVER['HTTP_USER_AGENT'] : '',
 121                         'ChronologyProtection' => 'true'
 122                 ];
 123
 124                 $this->cliMode = isset( $conf['cliMode'] ) ? $conf['cliMode'] : PHP_SAPI === 'cli';
 125                 $this->hostname = isset( $conf['hostname'] ) ? $conf['hostname'] : gethostname();
 126                 $this->agent = isset( $conf['agent'] ) ? $conf['agent'] : '';
 127
 128                 $this->ticket = mt_rand();
 129         }
 130
 131         public function destroy() {
 132                 $this->shutdown( self::SHUTDOWN_NO_CHRONPROT );
 133                 $this->forEachLBCallMethod( 'disable' );
 134         }
 135
 136         public function shutdown(
 137                 $mode = self::SHUTDOWN_CHRONPROT_SYNC, callable $workCallback = null
 138         ) {
 139                 $chronProt = $this->getChronologyProtector();
 140                 if ( $mode === self::SHUTDOWN_CHRONPROT_SYNC ) {
 141                         $this->shutdownChronologyProtector( $chronProt, $workCallback, 'sync' );
 142                 } elseif ( $mode === self::SHUTDOWN_CHRONPROT_ASYNC ) {
 143                         $this->shutdownChronologyProtector( $chronProt, null, 'async' );
 144                 }
 145
 146                 $this->commitMasterChanges( __METHOD__ ); // sanity
 147         }
 148
 149         /**
 150          * @see ILBFactory::newMainLB()
 151          * @param bool $domain
 152          * @return LoadBalancer
 153          */
 154         abstract public function newMainLB( $domain = false );
 155
 156         /**
 157          * @see ILBFactory::getMainLB()
 158          * @param bool $domain
 159          * @return LoadBalancer
 160          */
 161         abstract public function getMainLB( $domain = false );
 162
 163         /**
 164          * @see ILBFactory::newExternalLB()
 165          * @param string $cluster
 166          * @return LoadBalancer
 167          */
 168         abstract public function newExternalLB( $cluster );
 169
 170         /**
 171          * @see ILBFactory::getExternalLB()
 172          * @param string $cluster
 173          * @return LoadBalancer
 174          */
 175         abstract public function getExternalLB( $cluster );
 176
 177         /**
 178          * Call a method of each tracked load balancer
 179          *
 180          * @param string $methodName
 181          * @param array $args
 182          */
 183         protected function forEachLBCallMethod( $methodName, array $args = [] ) {
 184                 $this->forEachLB(
 185                         function ( ILoadBalancer $loadBalancer, $methodName, array $args ) {
 186                                 call_user_func_array( [ $loadBalancer, $methodName ], $args );
 187                         },
 188                         [ $methodName, $args ]
 189                 );
 190         }
 191
 192         public function flushReplicaSnapshots( $fname = __METHOD__ ) {
 193                 $this->forEachLBCallMethod( 'flushReplicaSnapshots', [ $fname ] );
 194         }
 195
 196         public function commitAll( $fname = __METHOD__, array $options = [] ) {
 197                 $this->commitMasterChanges( $fname, $options );
 198                 $this->forEachLBCallMethod( 'commitAll', [ $fname ] );
 199         }
 200
 201         public function beginMasterChanges( $fname = __METHOD__ ) {
 202                 if ( $this->trxRoundId !== false ) {
 203                         throw new DBTransactionError(
 204                                 null,
 205                                 "$fname: transaction round '{$this->trxRoundId}' already started."
 206                         );
 207                 }
 208                 $this->trxRoundId = $fname;
 209                 // Set DBO_TRX flags on all appropriate DBs
 210                 $this->forEachLBCallMethod( 'beginMasterChanges', [ $fname ] );
 211         }
 212
 213         public function commitMasterChanges( $fname = __METHOD__, array $options = [] ) {
 214                 if ( $this->trxRoundId !== false && $this->trxRoundId !== $fname ) {
 215                         throw new DBTransactionError(
 216                                 null,
 217                                 "$fname: transaction round '{$this->trxRoundId}' still running."
 218                         );
 219                 }
 220                 /** @noinspection PhpUnusedLocalVariableInspection */
 221                 $scope = $this->getScopedPHPBehaviorForCommit(); // try to ignore client aborts
 222                 // Run pre-commit callbacks and suppress post-commit callbacks, aborting on failure
 223                 $this->forEachLBCallMethod( 'finalizeMasterChanges' );
 224                 $this->trxRoundId = false;
 225                 // Perform pre-commit checks, aborting on failure
 226                 $this->forEachLBCallMethod( 'approveMasterChanges', [ $options ] );
 227                 // Log the DBs and methods involved in multi-DB transactions
 228                 $this->logIfMultiDbTransaction();
 229                 // Actually perform the commit on all master DB connections and revert DBO_TRX
 230                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 231                 // Run all post-commit callbacks
 232                 /** @var Exception $e */
 233                 $e = null; // first callback exception
 234                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$e ) {
 235                         $ex = $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_COMMIT );
 236                         $e = $e ?: $ex;
 237                 } );
 238                 // Commit any dangling DBO_TRX transactions from callbacks on one DB to another DB
 239                 $this->forEachLBCallMethod( 'commitMasterChanges', [ $fname ] );
 240                 // Throw any last post-commit callback error
 241                 if ( $e instanceof Exception ) {
 242                         throw $e;
 243                 }
 244         }
 245
 246         public function rollbackMasterChanges( $fname = __METHOD__ ) {
 247                 $this->trxRoundId = false;
 248                 $this->forEachLBCallMethod( 'suppressTransactionEndCallbacks' );
 249                 $this->forEachLBCallMethod( 'rollbackMasterChanges', [ $fname ] );
 250                 // Run all post-rollback callbacks
 251                 $this->forEachLB( function ( ILoadBalancer $lb ) {
 252                         $lb->runMasterPostTrxCallbacks( IDatabase::TRIGGER_ROLLBACK );
 253                 } );
 254         }
 255
 256         /**
 257          * Log query info if multi DB transactions are going to be committed now
 258          */
 259         private function logIfMultiDbTransaction() {
 260                 $callersByDB = [];
 261                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$callersByDB ) {
 262                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 263                         $callers = $lb->pendingMasterChangeCallers();
 264                         if ( $callers ) {
 265                                 $callersByDB[$masterName] = $callers;
 266                         }
 267                 } );
 268
 269                 if ( count( $callersByDB ) >= 2 ) {
 270                         $dbs = implode( ', ', array_keys( $callersByDB ) );
 271                         $msg = "Multi-DB transaction [{$dbs}]:\n";
 272                         foreach ( $callersByDB as $db => $callers ) {
 273                                 $msg .= "$db: " . implode( '; ', $callers ) . "\n";
 274                         }
 275                         $this->queryLogger->info( $msg );
 276                 }
 277         }
 278
 279         public function hasMasterChanges() {
 280                 $ret = false;
 281                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$ret ) {
 282                         $ret = $ret || $lb->hasMasterChanges();
 283                 } );
 284
 285                 return $ret;
 286         }
 287
 288         public function laggedReplicaUsed() {
 289                 $ret = false;
 290                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$ret ) {
 291                         $ret = $ret || $lb->laggedReplicaUsed();
 292                 } );
 293
 294                 return $ret;
 295         }
 296
 297         public function hasOrMadeRecentMasterChanges( $age = null ) {
 298                 $ret = false;
 299                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $age, &$ret ) {
 300                         $ret = $ret || $lb->hasOrMadeRecentMasterChanges( $age );
 301                 } );
 302                 return $ret;
 303         }
 304
 305         public function waitForReplication( array $opts = [] ) {
 306                 $opts += [
 307                         'domain' => false,
 308                         'cluster' => false,
 309                         'timeout' => 60,
 310                         'ifWritesSince' => null
 311                 ];
 312
 313                 if ( $opts['domain'] === false && isset( $opts['wiki'] ) ) {
 314                         $opts['domain'] = $opts['wiki']; // b/c
 315                 }
 316
 317                 // Figure out which clusters need to be checked
 318                 /** @var ILoadBalancer[] $lbs */
 319                 $lbs = [];
 320                 if ( $opts['cluster'] !== false ) {
 321                         $lbs[] = $this->getExternalLB( $opts['cluster'] );
 322                 } elseif ( $opts['domain'] !== false ) {
 323                         $lbs[] = $this->getMainLB( $opts['domain'] );
 324                 } else {
 325                         $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$lbs ) {
 326                                 $lbs[] = $lb;
 327                         } );
 328                         if ( !$lbs ) {
 329                                 return; // nothing actually used
 330                         }
 331                 }
 332
 333                 // Get all the master positions of applicable DBs right now.
 334                 // This can be faster since waiting on one cluster reduces the
 335                 // time needed to wait on the next clusters.
 336                 $masterPositions = array_fill( 0, count( $lbs ), false );
 337                 foreach ( $lbs as $i => $lb ) {
 338                         if ( $lb->getServerCount() <= 1 ) {
 339                                 // T29975 - Don't try to wait for replica DBs if there are none
 340                                 // Prevents permission error when getting master position
 341                                 continue;
 342                         } elseif ( $opts['ifWritesSince']
 343                                 && $lb->lastMasterChangeTimestamp() < $opts['ifWritesSince']
 344                         ) {
 345                                 continue; // no writes since the last wait
 346                         }
 347                         $masterPositions[$i] = $lb->getMasterPos();
 348                 }
 349
 350                 // Run any listener callbacks *after* getting the DB positions. The more
 351                 // time spent in the callbacks, the less time is spent in waitForAll().
 352                 foreach ( $this->replicationWaitCallbacks as $callback ) {
 353                         $callback();
 354                 }
 355
 356                 $failed = [];
 357                 foreach ( $lbs as $i => $lb ) {
 358                         if ( $masterPositions[$i] ) {
 359                                 // The DBMS may not support getMasterPos()
 360                                 if ( !$lb->waitForAll( $masterPositions[$i], $opts['timeout'] ) ) {
 361                                         $failed[] = $lb->getServerName( $lb->getWriterIndex() );
 362                                 }
 363                         }
 364                 }
 365
 366                 if ( $failed ) {
 367                         throw new DBReplicationWaitError(
 368                                 null,
 369                                 "Could not wait for replica DBs to catch up to " .
 370                                 implode( ', ', $failed )
 371                         );
 372                 }
 373         }
 374
 375         public function setWaitForReplicationListener( $name, callable $callback = null ) {
 376                 if ( $callback ) {
 377                         $this->replicationWaitCallbacks[$name] = $callback;
 378                 } else {
 379                         unset( $this->replicationWaitCallbacks[$name] );
 380                 }
 381         }
 382
 383         public function getEmptyTransactionTicket( $fname ) {
 384                 if ( $this->hasMasterChanges() ) {
 385                         $this->queryLogger->error( __METHOD__ . ": $fname does not have outer scope.\n" .
 386                                 ( new RuntimeException() )->getTraceAsString() );
 387
 388                         return null;
 389                 }
 390
 391                 return $this->ticket;
 392         }
 393
 394         public function commitAndWaitForReplication( $fname, $ticket, array $opts = [] ) {
 395                 if ( $ticket !== $this->ticket ) {
 396                         $this->perfLogger->error( __METHOD__ . ": $fname does not have outer scope.\n" .
 397                                 ( new RuntimeException() )->getTraceAsString() );
 398
 399                         return;
 400                 }
 401
 402                 // The transaction owner and any caller with the empty transaction ticket can commit
 403                 // so that getEmptyTransactionTicket() callers don't risk seeing DBTransactionError.
 404                 if ( $this->trxRoundId !== false && $fname !== $this->trxRoundId ) {
 405                         $this->queryLogger->info( "$fname: committing on behalf of {$this->trxRoundId}." );
 406                         $fnameEffective = $this->trxRoundId;
 407                 } else {
 408                         $fnameEffective = $fname;
 409                 }
 410
 411                 $this->commitMasterChanges( $fnameEffective );
 412                 $this->waitForReplication( $opts );
 413                 // If a nested caller committed on behalf of $fname, start another empty $fname
 414                 // transaction, leaving the caller with the same empty transaction state as before.
 415                 if ( $fnameEffective !== $fname ) {
 416                         $this->beginMasterChanges( $fnameEffective );
 417                 }
 418         }
 419
 420         public function getChronologyProtectorTouched( $dbName ) {
 421                 return $this->getChronologyProtector()->getTouched( $dbName );
 422         }
 423
 424         public function disableChronologyProtection() {
 425                 $this->getChronologyProtector()->setEnabled( false );
 426         }
 427
 428         /**
 429          * @return ChronologyProtector
 430          */
 431         protected function getChronologyProtector() {
 432                 if ( $this->chronProt ) {
 433                         return $this->chronProt;
 434                 }
 435
 436                 $this->chronProt = new ChronologyProtector(
 437                         $this->memCache,
 438                         [
 439                                 'ip' => $this->requestInfo['IPAddress'],
 440                                 'agent' => $this->requestInfo['UserAgent'],
 441                         ],
 442                         isset( $_GET['cpPosTime'] ) ? $_GET['cpPosTime'] : null
 443                 );
 444                 $this->chronProt->setLogger( $this->replLogger );
 445
 446                 if ( $this->cliMode ) {
 447                         $this->chronProt->setEnabled( false );
 448                 } elseif ( $this->requestInfo['ChronologyProtection'] === 'false' ) {
 449                         // Request opted out of using position wait logic. This is useful for requests
 450                         // done by the job queue or background ETL that do not have a meaningful session.
 451                         $this->chronProt->setWaitEnabled( false );
 452                 }
 453
 454                 $this->replLogger->debug( __METHOD__ . ': using request info ' .
 455                         json_encode( $this->requestInfo, JSON_PRETTY_PRINT ) );
 456
 457                 return $this->chronProt;
 458         }
 459
 460         /**
 461          * Get and record all of the staged DB positions into persistent memory storage
 462          *
 463          * @param ChronologyProtector $cp
 464          * @param callable|null $workCallback Work to do instead of waiting on syncing positions
 465          * @param string $mode One of (sync, async); whether to wait on remote datacenters
 466          */
 467         protected function shutdownChronologyProtector(
 468                 ChronologyProtector $cp, $workCallback, $mode
 469         ) {
 470                 // Record all the master positions needed
 471                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $cp ) {
 472                         $cp->shutdownLB( $lb );
 473                 } );
 474                 // Write them to the persistent stash. Try to do something useful by running $work
 475                 // while ChronologyProtector waits for the stash write to replicate to all DCs.
 476                 $unsavedPositions = $cp->shutdown( $workCallback, $mode );
 477                 if ( $unsavedPositions && $workCallback ) {
 478                         // Invoke callback in case it did not cache the result yet
 479                         $workCallback(); // work now to block for less time in waitForAll()
 480                 }
 481                 // If the positions failed to write to the stash, at least wait on local datacenter
 482                 // replica DBs to catch up before responding. Even if there are several DCs, this increases
 483                 // the chance that the user will see their own changes immediately afterwards. As long
 484                 // as the sticky DC cookie applies (same domain), this is not even an issue.
 485                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( $unsavedPositions ) {
 486                         $masterName = $lb->getServerName( $lb->getWriterIndex() );
 487                         if ( isset( $unsavedPositions[$masterName] ) ) {
 488                                 $lb->waitForAll( $unsavedPositions[$masterName] );
 489                         }
 490                 } );
 491         }
 492
 493         /**
 494          * Base parameters to LoadBalancer::__construct()
 495          * @return array
 496          */
 497         final protected function baseLoadBalancerParams() {
 498                 return [
 499                         'localDomain' => $this->localDomain,
 500                         'readOnlyReason' => $this->readOnlyReason,
 501                         'srvCache' => $this->srvCache,
 502                         'wanCache' => $this->wanCache,
 503                         'profiler' => $this->profiler,
 504                         'trxProfiler' => $this->trxProfiler,
 505                         'queryLogger' => $this->queryLogger,
 506                         'connLogger' => $this->connLogger,
 507                         'replLogger' => $this->replLogger,
 508                         'errorLogger' => $this->errorLogger,
 509                         'hostname' => $this->hostname,
 510                         'cliMode' => $this->cliMode,
 511                         'agent' => $this->agent
 512                 ];
 513         }
 514
 515         /**
 516          * @param ILoadBalancer $lb
 517          */
 518         protected function initLoadBalancer( ILoadBalancer $lb ) {
 519                 if ( $this->trxRoundId !== false ) {
 520                         $lb->beginMasterChanges( $this->trxRoundId ); // set DBO_TRX
 521                 }
 522         }
 523
 524         public function setDomainPrefix( $prefix ) {
 525                 $this->localDomain = new DatabaseDomain(
 526                         $this->localDomain->getDatabase(),
 527                         null,
 528                         $prefix
 529                 );
 530
 531                 $this->forEachLB( function( ILoadBalancer $lb ) use ( $prefix ) {
 532                         $lb->setDomainPrefix( $prefix );
 533                 } );
 534         }
 535
 536         public function closeAll() {
 537                 $this->forEachLBCallMethod( 'closeAll', [] );
 538         }
 539
 540         public function setAgentName( $agent ) {
 541                 $this->agent = $agent;
 542         }
 543
 544         public function appendPreShutdownTimeAsQuery( $url, $time ) {
 545                 $usedCluster = 0;
 546                 $this->forEachLB( function ( ILoadBalancer $lb ) use ( &$usedCluster ) {
 547                         $usedCluster |= ( $lb->getServerCount() > 1 );
 548                 } );
 549
 550                 if ( !$usedCluster ) {
 551                         return $url; // no master/replica clusters touched
 552                 }
 553
 554                 return strpos( $url, '?' ) === false ? "$url?cpPosTime=$time" : "$url&cpPosTime=$time";
 555         }
 556
 557         public function setRequestInfo( array $info ) {
 558                 $this->requestInfo = $info + $this->requestInfo;
 559         }
 560
 561         /**
 562          * Make PHP ignore user aborts/disconnects until the returned
 563          * value leaves scope. This returns null and does nothing in CLI mode.
 564          *
 565          * @return ScopedCallback|null
 566          */
 567         final protected function getScopedPHPBehaviorForCommit() {
 568                 if ( PHP_SAPI != 'cli' ) { // https://bugs.php.net/bug.php?id=47540
 569                         $old = ignore_user_abort( true ); // avoid half-finished operations
 570                         return new ScopedCallback( function () use ( $old ) {
 571                                 ignore_user_abort( $old );
 572                         } );
 573                 }
 574
 575                 return null;
 576         }
 577
 578         function __destruct() {
 579                 $this->destroy();
 580         }
 581 }