includes/libs/rdbms/ChronologyProtector.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 namespace Wikimedia\Rdbms;
  25
  26 use Psr\Log\LoggerAwareInterface;
  27 use Psr\Log\LoggerInterface;
  28 use Psr\Log\NullLogger;
  29 use Wikimedia\WaitConditionLoop;
  30 use BagOStuff;
  31
  32 /**
  33  * Class for ensuring a consistent ordering of events as seen by the user, despite replication.
  34  * Kind of like Hawking's [[Chronology Protection Agency]].
  35  */
  36 class ChronologyProtector implements LoggerAwareInterface {
  37         /** @var BagOStuff */
  38         protected $store;
  39         /** @var LoggerInterface */
  40         protected $logger;
  41
  42         /** @var string Storage key name */
  43         protected $key;
  44         /** @var string Hash of client parameters */
  45         protected $clientId;
  46         /** @var float|null Minimum UNIX timestamp of 1+ expected startup positions */
  47         protected $waitForPosTime;
  48         /** @var int Max seconds to wait on positions to appear */
  49         protected $waitForPosTimeout = self::POS_WAIT_TIMEOUT;
  50         /** @var bool Whether to no-op all method calls */
  51         protected $enabled = true;
  52         /** @var bool Whether to check and wait on positions */
  53         protected $wait = true;
  54
  55         /** @var bool Whether the client data was loaded */
  56         protected $initialized = false;
  57         /** @var DBMasterPos[] Map of (DB master name => position) */
  58         protected $startupPositions = [];
  59         /** @var DBMasterPos[] Map of (DB master name => position) */
  60         protected $shutdownPositions = [];
  61         /** @var float[] Map of (DB master name => 1) */
  62         protected $shutdownTouchDBs = [];
  63
  64         /** @var integer Seconds to store positions */
  65         const POSITION_TTL = 60;
  66         /** @var integer Max time to wait for positions to appear */
  67         const POS_WAIT_TIMEOUT = 5;
  68
  69         /**
  70          * @param BagOStuff $store
  71          * @param array $client Map of (ip: <IP>, agent: <user-agent>)
  72          * @param float $posTime UNIX timestamp
  73          * @since 1.27
  74          */
  75         public function __construct( BagOStuff $store, array $client, $posTime = null ) {
  76                 $this->store = $store;
  77                 $this->clientId = md5( $client['ip'] . "\n" . $client['agent'] );
  78                 $this->key = $store->makeGlobalKey( __CLASS__, $this->clientId );
  79                 $this->waitForPosTime = $posTime;
  80                 $this->logger = new NullLogger();
  81         }
  82
  83         public function setLogger( LoggerInterface $logger ) {
  84                 $this->logger = $logger;
  85         }
  86
  87         /**
  88          * @param bool $enabled Whether to no-op all method calls
  89          * @since 1.27
  90          */
  91         public function setEnabled( $enabled ) {
  92                 $this->enabled = $enabled;
  93         }
  94
  95         /**
  96          * @param bool $enabled Whether to check and wait on positions
  97          * @since 1.27
  98          */
  99         public function setWaitEnabled( $enabled ) {
 100                 $this->wait = $enabled;
 101         }
 102
 103         /**
 104          * Initialise a ILoadBalancer to give it appropriate chronology protection.
 105          *
 106          * If the stash has a previous master position recorded, this will try to
 107          * make sure that the next query to a replica DB of that master will see changes up
 108          * to that position by delaying execution. The delay may timeout and allow stale
 109          * data if no non-lagged replica DBs are available.
 110          *
 111          * @param ILoadBalancer $lb
 112          * @return void
 113          */
 114         public function initLB( ILoadBalancer $lb ) {
 115                 if ( !$this->enabled || $lb->getServerCount() <= 1 ) {
 116                         return; // non-replicated setup or disabled
 117                 }
 118
 119                 $this->initPositions();
 120
 121                 $masterName = $lb->getServerName( $lb->getWriterIndex() );
 122                 if ( !empty( $this->startupPositions[$masterName] ) ) {
 123                         $pos = $this->startupPositions[$masterName];
 124                         $this->logger->info( __METHOD__ . ": LB for '$masterName' set to pos $pos\n" );
 125                         $lb->waitFor( $pos );
 126                 }
 127         }
 128
 129         /**
 130          * Notify the ChronologyProtector that the ILoadBalancer is about to shut
 131          * down. Saves replication positions.
 132          *
 133          * @param ILoadBalancer $lb
 134          * @return void
 135          */
 136         public function shutdownLB( ILoadBalancer $lb ) {
 137                 if ( !$this->enabled ) {
 138                         return; // not enabled
 139                 } elseif ( !$lb->hasOrMadeRecentMasterChanges( INF ) ) {
 140                         // Only save the position if writes have been done on the connection
 141                         return;
 142                 }
 143
 144                 $masterName = $lb->getServerName( $lb->getWriterIndex() );
 145                 if ( $lb->getServerCount() > 1 ) {
 146                         $pos = $lb->getMasterPos();
 147                         $this->logger->info( __METHOD__ . ": LB for '$masterName' has pos $pos\n" );
 148                         $this->shutdownPositions[$masterName] = $pos;
 149                 } else {
 150                         $this->logger->info( __METHOD__ . ": DB '$masterName' touched\n" );
 151                 }
 152                 $this->shutdownTouchDBs[$masterName] = 1;
 153         }
 154
 155         /**
 156          * Notify the ChronologyProtector that the LBFactory is done calling shutdownLB() for now.
 157          * May commit chronology data to persistent storage.
 158          *
 159          * @param callable|null $workCallback Work to do instead of waiting on syncing positions
 160          * @param string $mode One of (sync, async); whether to wait on remote datacenters
 161          * @return DBMasterPos[] Empty on success; returns the (db name => position) map on failure
 162          */
 163         public function shutdown( callable $workCallback = null, $mode = 'sync' ) {
 164                 if ( !$this->enabled ) {
 165                         return [];
 166                 }
 167
 168                 $store = $this->store;
 169                 // Some callers might want to know if a user recently touched a DB.
 170                 // These writes do not need to block on all datacenters receiving them.
 171                 foreach ( $this->shutdownTouchDBs as $dbName => $unused ) {
 172                         $store->set(
 173                                 $this->getTouchedKey( $this->store, $dbName ),
 174                                 microtime( true ),
 175                                 $store::TTL_DAY
 176                         );
 177                 }
 178
 179                 if ( !count( $this->shutdownPositions ) ) {
 180                         return []; // nothing to save
 181                 }
 182
 183                 $this->logger->info( __METHOD__ . ": saving master pos for " .
 184                         implode( ', ', array_keys( $this->shutdownPositions ) ) . "\n"
 185                 );
 186
 187                 // CP-protected writes should overwhemingly go to the master datacenter, so get DC-local
 188                 // lock to merge the values. Use a DC-local get() and a synchronous all-DC set(). This
 189                 // makes it possible for the BagOStuff class to write in parallel to all DCs with one RTT.
 190                 if ( $store->lock( $this->key, 3 ) ) {
 191                         if ( $workCallback ) {
 192                                 // Let the store run the work before blocking on a replication sync barrier. By the
 193                                 // time it's done with the work, the barrier should be fast if replication caught up.
 194                                 $store->addBusyCallback( $workCallback );
 195                         }
 196                         $ok = $store->set(
 197                                 $this->key,
 198                                 self::mergePositions( $store->get( $this->key ), $this->shutdownPositions ),
 199                                 self::POSITION_TTL,
 200                                 ( $mode === 'sync' ) ? $store::WRITE_SYNC : 0
 201                         );
 202                         $store->unlock( $this->key );
 203                 } else {
 204                         $ok = false;
 205                 }
 206
 207                 if ( !$ok ) {
 208                         $bouncedPositions = $this->shutdownPositions;
 209                         // Raced out too many times or stash is down
 210                         $this->logger->warning( __METHOD__ . ": failed to save master pos for " .
 211                                 implode( ', ', array_keys( $this->shutdownPositions ) ) . "\n"
 212                         );
 213                 } elseif ( $mode === 'sync' &&
 214                         $store->getQoS( $store::ATTR_SYNCWRITES ) < $store::QOS_SYNCWRITES_BE
 215                 ) {
 216                         // Positions may not be in all datacenters, force LBFactory to play it safe
 217                         $this->logger->info( __METHOD__ . ": store may not support synchronous writes." );
 218                         $bouncedPositions = $this->shutdownPositions;
 219                 } else {
 220                         $bouncedPositions = [];
 221                 }
 222
 223                 return $bouncedPositions;
 224         }
 225
 226         /**
 227          * @param string $dbName DB master name (e.g. "db1052")
 228          * @return float|bool UNIX timestamp when client last touched the DB; false if not on record
 229          * @since 1.28
 230          */
 231         public function getTouched( $dbName ) {
 232                 return $this->store->get( $this->getTouchedKey( $this->store, $dbName ) );
 233         }
 234
 235         /**
 236          * @param BagOStuff $store
 237          * @param string $dbName
 238          * @return string
 239          */
 240         private function getTouchedKey( BagOStuff $store, $dbName ) {
 241                 return $store->makeGlobalKey( __CLASS__, 'mtime', $this->clientId, $dbName );
 242         }
 243
 244         /**
 245          * Load in previous master positions for the client
 246          */
 247         protected function initPositions() {
 248                 if ( $this->initialized ) {
 249                         return;
 250                 }
 251
 252                 $this->initialized = true;
 253                 if ( $this->wait ) {
 254                         // If there is an expectation to see master positions with a certain min
 255                         // timestamp, then block until they appear, or until a timeout is reached.
 256                         if ( $this->waitForPosTime > 0.0 ) {
 257                                 $data = null;
 258                                 $loop = new WaitConditionLoop(
 259                                         function () use ( &$data ) {
 260                                                 $data = $this->store->get( $this->key );
 261
 262                                                 return ( self::minPosTime( $data ) >= $this->waitForPosTime )
 263                                                         ? WaitConditionLoop::CONDITION_REACHED
 264                                                         : WaitConditionLoop::CONDITION_CONTINUE;
 265                                         },
 266                                         $this->waitForPosTimeout
 267                                 );
 268                                 $result = $loop->invoke();
 269                                 $waitedMs = $loop->getLastWaitTime() * 1e3;
 270
 271                                 if ( $result == $loop::CONDITION_REACHED ) {
 272                                         $msg = "expected and found pos time {$this->waitForPosTime} ({$waitedMs}ms)";
 273                                         $this->logger->debug( $msg );
 274                                 } else {
 275                                         $msg = "expected but missed pos time {$this->waitForPosTime} ({$waitedMs}ms)";
 276                                         $this->logger->info( $msg );
 277                                 }
 278                         } else {
 279                                 $data = $this->store->get( $this->key );
 280                         }
 281
 282                         $this->startupPositions = $data ? $data['positions'] : [];
 283                         $this->logger->info( __METHOD__ . ": key is {$this->key} (read)\n" );
 284                 } else {
 285                         $this->startupPositions = [];
 286                         $this->logger->info( __METHOD__ . ": key is {$this->key} (unread)\n" );
 287                 }
 288         }
 289
 290         /**
 291          * @param array|bool $data
 292          * @return float|null
 293          */
 294         private static function minPosTime( $data ) {
 295                 if ( !isset( $data['positions'] ) ) {
 296                         return null;
 297                 }
 298
 299                 $min = null;
 300                 foreach ( $data['positions'] as $pos ) {
 301                         /** @var DBMasterPos $pos */
 302                         $min = $min ? min( $pos->asOfTime(), $min ) : $pos->asOfTime();
 303                 }
 304
 305                 return $min;
 306         }
 307
 308         /**
 309          * @param array|bool $curValue
 310          * @param DBMasterPos[] $shutdownPositions
 311          * @return array
 312          */
 313         private static function mergePositions( $curValue, array $shutdownPositions ) {
 314                 /** @var $curPositions DBMasterPos[] */
 315                 if ( $curValue === false ) {
 316                         $curPositions = $shutdownPositions;
 317                 } else {
 318                         $curPositions = $curValue['positions'];
 319                         // Use the newest positions for each DB master
 320                         foreach ( $shutdownPositions as $db => $pos ) {
 321                                 if ( !isset( $curPositions[$db] )
 322                                         || $pos->asOfTime() > $curPositions[$db]->asOfTime()
 323                                 ) {
 324                                         $curPositions[$db] = $pos;
 325                                 }
 326                         }
 327                 }
 328
 329                 return [ 'positions' => $curPositions ];
 330         }
 331 }