includes/libs/rdbms/lbfactory/LBFactoryMulti.php

   1 <?php
   2 /**
   3  * Advanced generator of database load balancing objects for database farms.
   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 InvalidArgumentException;
  27 use UnexpectedValueException;
  28
  29 /**
  30  * A multi-database, multi-master factory for Wikimedia and similar installations.
  31  * Ignores the old configuration globals.
  32  *
  33  * @ingroup Database
  34  */
  35 class LBFactoryMulti extends LBFactory {
  36         /** @var LoadBalancer[] */
  37         private $mainLBs = [];
  38         /** @var LoadBalancer[] */
  39         private $externalLBs = [];
  40
  41         /** @var string[] Map of (hostname => IP address) */
  42         private $hostsByName = [];
  43         /** @var string[] Map of (database name => section name) */
  44         private $sectionsByDB = [];
  45         /** @var int[][][] Map of (section => group => host => load ratio) */
  46         private $groupLoadsBySection = [];
  47         /** @var int[][][] Map of (database => group => host => load ratio) */
  48         private $groupLoadsByDB = [];
  49         /** @var int[][] Map of (cluster => host => load ratio) */
  50         private $externalLoads = [];
  51         /** @var array Server config map ("host", "hostName", "load", and "groupLoads" are ignored) */
  52         private $serverTemplate = [];
  53         /** @var array Server config map overriding "serverTemplate" for external storage */
  54         private $externalTemplateOverrides = [];
  55         /** @var array[] Map of (section => server config map overrides) */
  56         private $templateOverridesBySection = [];
  57         /** @var array[] Map of (cluster => server config map overrides) for external storage */
  58         private $templateOverridesByCluster = [];
  59         /** @var array Server config override map for all main and external master servers */
  60         private $masterTemplateOverrides = [];
  61         /** @var array[] Map of (host => server config map overrides) for main and external servers */
  62         private $templateOverridesByServer = [];
  63         /**  @var string[]|bool[] A map of section name to read-only message */
  64         private $readOnlyBySection = [];
  65
  66         /** @var string An ILoadMonitor class */
  67         private $loadMonitorClass;
  68
  69         /** @var string */
  70         private $lastDomain;
  71         /** @var string */
  72         private $lastSection;
  73
  74         /**
  75          * Template override precedence (highest => lowest):
  76          *   - templateOverridesByServer
  77          *   - masterTemplateOverrides
  78          *   - templateOverridesBySection/templateOverridesByCluster
  79          *   - externalTemplateOverrides
  80          *   - serverTemplate
  81          * Overrides only work on top level keys (so nested values will not be merged).
  82          *
  83          * Server config maps should be of the format Database::factory() requires.
  84          * Additionally, a 'max lag' key should also be set on server maps, indicating how stale the
  85          * data can be before the load balancer tries to avoid using it. The map can have 'is static'
  86          * set to disable blocking  replication sync checks (intended for archive servers with
  87          * unchanging data).
  88
  89          * @see LBFactory::__construct()
  90          * @param array $conf Additional parameters include:
  91          *   - hostsByName                 Optional (hostname => IP address) map.
  92          *   - sectionsByDB                Optional map of (database => section name).
  93          *                                 For example:
  94          *                                 [
  95          *                                     'DEFAULT' => 'section1',
  96          *                                     'database1' => 'section2'
  97          *                                 ]
  98          *   - sectionLoads                Optional map of (section => host => load ratio); the first
  99          *                                 host in each section is the master server for that section.
 100          *                                 For example:
 101          *                                 [
 102          *                                     'dbmaser'    => 0,
 103          *                                     'dbreplica1' => 100,
 104          *                                     'dbreplica2' => 100
 105          *                                 ]
 106          *   - groupLoadsBySection         Optional map of (section => group => host => load ratio);
 107          *                                 any ILoadBalancer::GROUP_GENERIC group will be ignored.
 108          *                                 For example:
 109          *                                 [
 110          *                                     'section1' => [
 111          *                                         'group1' => [
 112          *                                             'dbreplica3  => 100,
 113          *                                             'dbreplica4' => 100
 114          *                                         ]
 115          *                                     ]
 116          *                                 ]
 117          *   - groupLoadsByDB              Optional (database => group => host => load ratio) map.
 118          *   - externalLoads               Optional (cluster => host => load ratio) map.
 119          *   - serverTemplate              server config map for Database::factory().
 120          *                                 Note that "host", "hostName" and "load" entries will be
 121          *                                 overridden by "groupLoadsBySection" and "hostsByName".
 122          *   - externalTemplateOverrides   Optional server config map overrides for external
 123          *                                 stores; respects the override precedence described above.
 124          *   - templateOverridesBySection  Optional (section => server config map overrides) map;
 125          *                                 respects the override precedence described above.
 126          *   - templateOverridesByCluster  Optional (external cluster => server config map overrides)
 127          *                                 map; respects the override precedence described above.
 128          *   - masterTemplateOverrides     Optional server config map overrides for masters;
 129          *                                 respects the override precedence described above.
 130          *   - templateOverridesByServer   Optional (host => server config map overrides) map;
 131          *                                 respects the override precedence described above
 132          *                                 and applies to both core and external storage.
 133          *   - loadMonitorClass            Name of the LoadMonitor class to always use. [optional]
 134          *   - readOnlyBySection           Optional map of (section name => message text or false).
 135          *                                 String values make sections read only, whereas anything
 136          *                                 else does not restrict read/write mode.
 137          */
 138         public function __construct( array $conf ) {
 139                 parent::__construct( $conf );
 140
 141                 $this->hostsByName = $conf['hostsByName'] ?? [];
 142                 $this->sectionsByDB = $conf['sectionsByDB'];
 143                 $this->groupLoadsBySection = $conf['groupLoadsBySection'] ?? [];
 144                 foreach ( ( $conf['sectionLoads'] ?? [] ) as $section => $loadByHost ) {
 145                         $this->groupLoadsBySection[$section][ILoadBalancer::GROUP_GENERIC] = $loadByHost;
 146                 }
 147                 $this->groupLoadsByDB = $conf['groupLoadsByDB'] ?? [];
 148                 $this->externalLoads = $conf['externalLoads'] ?? [];
 149                 $this->serverTemplate = $conf['serverTemplate'] ?? [];
 150                 $this->externalTemplateOverrides = $conf['externalTemplateOverrides'] ?? [];
 151                 $this->templateOverridesBySection = $conf['templateOverridesBySection'] ?? [];
 152                 $this->templateOverridesByCluster = $conf['templateOverridesByCluster'] ?? [];
 153                 $this->masterTemplateOverrides = $conf['masterTemplateOverrides'] ?? [];
 154                 $this->templateOverridesByServer = $conf['templateOverridesByServer'] ?? [];
 155                 $this->readOnlyBySection = $conf['readOnlyBySection'] ?? [];
 156
 157                 $this->loadMonitorClass = $conf['loadMonitorClass'] ?? LoadMonitor::class;
 158         }
 159
 160         public function newMainLB( $domain = false ) {
 161                 $section = $this->getSectionForDomain( $domain );
 162                 if ( !isset( $this->groupLoadsBySection[$section][ILoadBalancer::GROUP_GENERIC] ) ) {
 163                         throw new UnexpectedValueException( "Section '$section' has no hosts defined." );
 164                 }
 165
 166                 $dbGroupLoads = $this->groupLoadsByDB[$this->getDomainDatabase( $domain )] ?? [];
 167                 unset( $dbGroupLoads[ILoadBalancer::GROUP_GENERIC] ); // cannot override
 168
 169                 return $this->newLoadBalancer(
 170                         array_merge(
 171                                 $this->serverTemplate,
 172                                 $this->templateOverridesBySection[$section] ?? []
 173                         ),
 174                         array_merge( $this->groupLoadsBySection[$section], $dbGroupLoads ),
 175                         // Use the LB-specific read-only reason if everything isn't already read-only
 176                         is_string( $this->readOnlyReason )
 177                                 ? $this->readOnlyReason
 178                                 : ( $this->readOnlyBySection[$section] ?? false )
 179                 );
 180         }
 181
 182         public function getMainLB( $domain = false ) {
 183                 $section = $this->getSectionForDomain( $domain );
 184
 185                 if ( !isset( $this->mainLBs[$section] ) ) {
 186                         $this->mainLBs[$section] = $this->newMainLB( $domain );
 187                 }
 188
 189                 return $this->mainLBs[$section];
 190         }
 191
 192         public function newExternalLB( $cluster ) {
 193                 if ( !isset( $this->externalLoads[$cluster] ) ) {
 194                         throw new InvalidArgumentException( "Unknown cluster '$cluster'" );
 195                 }
 196
 197                 return $this->newLoadBalancer(
 198                         array_merge(
 199                                 $this->serverTemplate,
 200                                 $this->externalTemplateOverrides,
 201                                 $this->templateOverridesByCluster[$cluster] ?? []
 202                         ),
 203                         [ ILoadBalancer::GROUP_GENERIC => $this->externalLoads[$cluster] ],
 204                         $this->readOnlyReason
 205                 );
 206         }
 207
 208         public function getExternalLB( $cluster ) {
 209                 if ( !isset( $this->externalLBs[$cluster] ) ) {
 210                         $this->externalLBs[$cluster] = $this->newExternalLB( $cluster );
 211                 }
 212
 213                 return $this->externalLBs[$cluster];
 214         }
 215
 216         public function getAllMainLBs() {
 217                 $lbs = [];
 218                 foreach ( $this->sectionsByDB as $db => $section ) {
 219                         if ( !isset( $lbs[$section] ) ) {
 220                                 $lbs[$section] = $this->getMainLB( $db );
 221                         }
 222                 }
 223
 224                 return $lbs;
 225         }
 226
 227         public function getAllExternalLBs() {
 228                 $lbs = [];
 229                 foreach ( $this->externalLoads as $cluster => $unused ) {
 230                         $lbs[$cluster] = $this->getExternalLB( $cluster );
 231                 }
 232
 233                 return $lbs;
 234         }
 235
 236         public function forEachLB( $callback, array $params = [] ) {
 237                 foreach ( $this->mainLBs as $lb ) {
 238                         $callback( $lb, ...$params );
 239                 }
 240                 foreach ( $this->externalLBs as $lb ) {
 241                         $callback( $lb, ...$params );
 242                 }
 243         }
 244
 245         /**
 246          * @param bool|string $domain
 247          * @return string
 248          */
 249         private function getSectionForDomain( $domain = false ) {
 250                 if ( $this->lastDomain === $domain ) {
 251                         return $this->lastSection;
 252                 }
 253
 254                 $database = $this->getDomainDatabase( $domain );
 255                 $section = $this->sectionsByDB[$database] ?? self::CLUSTER_MAIN_DEFAULT;
 256                 $this->lastSection = $section;
 257                 $this->lastDomain = $domain;
 258
 259                 return $section;
 260         }
 261
 262         /**
 263          * Make a new load balancer object based on template and load array
 264          *
 265          * @param array $serverTemplate Server config map
 266          * @param int[][] $groupLoads Map of (group => host => load)
 267          * @param string|bool $readOnlyReason
 268          * @return LoadBalancer
 269          */
 270         private function newLoadBalancer( $serverTemplate, $groupLoads, $readOnlyReason ) {
 271                 $lb = new LoadBalancer( array_merge(
 272                         $this->baseLoadBalancerParams(),
 273                         [
 274                                 'servers' => $this->makeServerArray( $serverTemplate, $groupLoads ),
 275                                 'loadMonitor' => [ 'class' => $this->loadMonitorClass ],
 276                                 'readOnlyReason' => $readOnlyReason
 277                         ]
 278                 ) );
 279                 $this->initLoadBalancer( $lb );
 280
 281                 return $lb;
 282         }
 283
 284         /**
 285          * Make a server array as expected by LoadBalancer::__construct()
 286          *
 287          * @param array $serverTemplate Server config map
 288          * @param int[][] $groupLoads Map of (group => host => load)
 289          * @return array[] List of server config maps
 290          */
 291         private function makeServerArray( array $serverTemplate, array $groupLoads ) {
 292                 // The master server is the first host explicitly listed in the generic load group
 293                 if ( !$groupLoads[ILoadBalancer::GROUP_GENERIC] ) {
 294                         throw new UnexpectedValueException( "Empty generic load array; no master defined." );
 295                 }
 296
 297                 $groupLoadsByHost = $this->reindexGroupLoads( $groupLoads );
 298                 // Get the ordered map of (host => load); the master server is first
 299                 $genericLoads = $groupLoads[ILoadBalancer::GROUP_GENERIC];
 300                 // Implictly append any hosts that only appear in custom load groups
 301                 $genericLoads += array_fill_keys( array_keys( $groupLoadsByHost ), 0 );
 302
 303                 $servers = [];
 304                 foreach ( $genericLoads as $host => $load ) {
 305                         $servers[] = array_merge(
 306                                 $serverTemplate,
 307                                 $servers ? [] : $this->masterTemplateOverrides,
 308                                 $this->templateOverridesByServer[$host] ?? [],
 309                                 [
 310                                         'host' => $this->hostsByName[$host] ?? $host,
 311                                         'hostName' => $host,
 312                                         'load' => $load,
 313                                         'groupLoads' => $groupLoadsByHost[$host] ?? []
 314                                 ]
 315                         );
 316                 }
 317
 318                 return $servers;
 319         }
 320
 321         /**
 322          * Take a group load array indexed by group then server, and reindex it by server then group
 323          * @param int[][] $groupLoads Map of (group => host => load)
 324          * @return int[][] Map of (host => group => load)
 325          */
 326         private function reindexGroupLoads( array $groupLoads ) {
 327                 $reindexed = [];
 328
 329                 foreach ( $groupLoads as $group => $loadByHost ) {
 330                         foreach ( $loadByHost as $host => $load ) {
 331                                 $reindexed[$host][$group] = $load;
 332                         }
 333                 }
 334
 335                 return $reindexed;
 336         }
 337
 338         /**
 339          * @param DatabaseDomain|string|bool $domain Domain ID, or false for the current domain
 340          * @return string
 341          */
 342         private function getDomainDatabase( $domain = false ) {
 343                 return ( $domain === false )
 344                         ? $this->localDomain->getDatabase()
 345                         : DatabaseDomain::newFromId( $domain )->getDatabase();
 346         }
 347 }