includes/db/loadbalancer/LoadMonitorMySQL.php

   1 <?php
   2 /**
   3  * This program is free software; you can redistribute it and/or modify
   4  * it under the terms of the GNU General Public License as published by
   5  * the Free Software Foundation; either version 2 of the License, or
   6  * (at your option) any later version.
   7  *
   8  * This program is distributed in the hope that it will be useful,
   9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  11  * GNU General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public License along
  14  * with this program; if not, write to the Free Software Foundation, Inc.,
  15  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  16  * http://www.gnu.org/copyleft/gpl.html
  17  *
  18  * @file
  19  * @ingroup Database
  20  */
  21
  22 /**
  23  * Basic MySQL load monitor with no external dependencies
  24  * Uses memcached to cache the replication lag for a short time
  25  *
  26  * @ingroup Database
  27  */
  28 class LoadMonitorMySQL implements LoadMonitor {
  29         /** @var LoadBalancer */
  30         public $parent;
  31         /** @var BagOStuff */
  32         protected $srvCache;
  33         /** @var BagOStuff */
  34         protected $mainCache;
  35
  36         public function __construct( $parent ) {
  37                 $this->parent = $parent;
  38
  39                 $this->srvCache = ObjectCache::getLocalServerInstance( 'hash' );
  40                 $this->mainCache = ObjectCache::getLocalClusterInstance();
  41         }
  42
  43         public function scaleLoads( &$loads, $group = false, $wiki = false ) {
  44         }
  45
  46         public function getLagTimes( $serverIndexes, $wiki ) {
  47                 if ( count( $serverIndexes ) == 1 && reset( $serverIndexes ) == 0 ) {
  48                         # Single server only, just return zero without caching
  49                         return array( 0 => 0 );
  50                 }
  51
  52                 $key = $this->getLagTimeCacheKey();
  53                 # Randomize TTLs to reduce stampedes (4.0 - 5.0 sec)
  54                 $ttl = mt_rand( 4e6, 5e6 ) / 1e6;
  55                 # Keep keys around longer as fallbacks
  56                 $staleTTL = 60;
  57
  58                 # (a) Check the local APC cache
  59                 $value = $this->srvCache->get( $key );
  60                 if ( $value && $value['timestamp'] > ( microtime( true ) - $ttl ) ) {
  61                         wfDebugLog( 'replication', __METHOD__ . ": got lag times ($key) from local cache" );
  62                         return $value['lagTimes']; // cache hit
  63                 }
  64                 $staleValue = $value ?: false;
  65
  66                 # (b) Check the shared cache and backfill APC
  67                 $value = $this->mainCache->get( $key );
  68                 if ( $value && $value['timestamp'] > ( microtime( true ) - $ttl ) ) {
  69                         $this->srvCache->set( $key, $value, $staleTTL );
  70                         wfDebugLog( 'replication', __METHOD__ . ": got lag times ($key) from main cache" );
  71
  72                         return $value['lagTimes']; // cache hit
  73                 }
  74                 $staleValue = $value ?: $staleValue;
  75
  76                 # (c) Cache key missing or expired; regenerate and backfill
  77                 if ( $this->mainCache->lock( $key, 0, 10 ) ) {
  78                         # Let this process alone update the cache value
  79                         $cache = $this->mainCache;
  80                         /** @noinspection PhpUnusedLocalVariableInspection */
  81                         $unlocker = new ScopedCallback( function () use ( $cache, $key ) {
  82                                 $cache->unlock( $key );
  83                         } );
  84                 } elseif ( $staleValue ) {
  85                         # Could not acquire lock but an old cache exists, so use it
  86                         return $staleValue['lagTimes'];
  87                 }
  88
  89                 $lagTimes = array();
  90                 foreach ( $serverIndexes as $i ) {
  91                         if ( $i == $this->parent->getWriterIndex() ) {
  92                                 $lagTimes[$i] = 0; // master always has no lag
  93                                 continue;
  94                         }
  95
  96                         $conn = $this->parent->getAnyOpenConnection( $i );
  97                         if ( $conn ) {
  98                                 $close = false; // already open
  99                         } else {
 100                                 $conn = $this->parent->openConnection( $i, $wiki );
 101                                 $close = true; // new connection
 102                         }
 103
 104                         if ( !$conn ) {
 105                                 $lagTimes[$i] = false;
 106                                 $host = $this->parent->getServerName( $i );
 107                                 wfDebugLog( 'replication', __METHOD__ . ": host $host (#$i) is unreachable" );
 108                                 continue;
 109                         }
 110
 111                         $lagTimes[$i] = $conn->getLag();
 112                         if ( $lagTimes[$i] === false ) {
 113                                 $host = $this->parent->getServerName( $i );
 114                                 wfDebugLog( 'replication', __METHOD__ . ": host $host (#$i) is not replicating?" );
 115                         }
 116
 117                         if ( $close ) {
 118                                 # Close the connection to avoid sleeper connections piling up.
 119                                 # Note that the caller will pick one of these DBs and reconnect,
 120                                 # which is slightly inefficient, but this only matters for the lag
 121                                 # time cache miss cache, which is far less common that cache hits.
 122                                 $this->parent->closeConnection( $conn );
 123                         }
 124                 }
 125
 126                 # Add a timestamp key so we know when it was cached
 127                 $value = array( 'lagTimes' => $lagTimes, 'timestamp' => microtime( true ) );
 128                 $this->mainCache->set( $key, $value, $staleTTL );
 129                 $this->srvCache->set( $key, $value, $staleTTL );
 130                 wfDebugLog( 'replication', __METHOD__ . ": re-calculated lag times ($key)" );
 131
 132                 return $value['lagTimes'];
 133         }
 134
 135         public function clearCaches() {
 136                 $key = $this->getLagTimeCacheKey();
 137                 $this->srvCache->delete( $key );
 138                 $this->mainCache->delete( $key );
 139         }
 140
 141         private function getLagTimeCacheKey() {
 142                 $writerIndex = $this->parent->getWriterIndex();
 143                 // Lag is per-server, not per-DB, so key on the master DB name
 144                 return $this->srvCache->makeGlobalKey(
 145                         'lag-times', $this->parent->getServerName( $writerIndex )
 146                 );
 147         }
 148 }