Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
69.40% |
93 / 134 |
|
38.46% |
5 / 13 |
CRAP | |
0.00% |
0 / 1 |
| LBFactoryMulti | |
69.40% |
93 / 134 |
|
38.46% |
5 / 13 |
82.57 | |
0.00% |
0 / 1 |
| __construct | |
75.00% |
18 / 24 |
|
0.00% |
0 / 1 |
5.39 | |||
| newMainLB | |
87.50% |
14 / 16 |
|
0.00% |
0 / 1 |
3.02 | |||
| resolveDomainInstance | |
69.23% |
9 / 13 |
|
0.00% |
0 / 1 |
7.05 | |||
| getMainLB | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
2 | |||
| newExternalLB | |
91.67% |
11 / 12 |
|
0.00% |
0 / 1 |
2.00 | |||
| getExternalLB | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
2 | |||
| getAllMainLBs | |
0.00% |
0 / 5 |
|
0.00% |
0 / 1 |
12 | |||
| getAllExternalLBs | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
| getLBsForOwner | |
75.00% |
3 / 4 |
|
0.00% |
0 / 1 |
3.14 | |||
| newLoadBalancer | |
100.00% |
11 / 11 |
|
100.00% |
1 / 1 |
1 | |||
| makeServerConfigArrays | |
100.00% |
13 / 13 |
|
100.00% |
1 / 1 |
3 | |||
| getSectionFromDatabase | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
3 | |||
| reconfigure | |
0.00% |
0 / 18 |
|
0.00% |
0 / 1 |
20 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * @license GPL-2.0-or-later |
| 4 | * @file |
| 5 | */ |
| 6 | namespace Wikimedia\Rdbms; |
| 7 | |
| 8 | use InvalidArgumentException; |
| 9 | use LogicException; |
| 10 | use UnexpectedValueException; |
| 11 | |
| 12 | /** |
| 13 | * LoadBalancer manager for sites with several "main" database clusters |
| 14 | * |
| 15 | * Each database cluster consists of a "primary" server and any number of replica servers, |
| 16 | * all of which converge, as soon as possible, to contain the same tables and rows. If |
| 17 | * a replication topology has multiple primaries, then the "primary" is merely the preferred |
| 18 | * co-primary in the current local datacenter. |
| 19 | * |
| 20 | * For single-primary topologies, the tables and rows of the primary define the "dataset". |
| 21 | * For multiple-primary topologies, the "dataset" is the convergent result of applying/merging |
| 22 | * all committed transactions (regardless of which co-primary they originate on); it possible that |
| 23 | * no co-primary has yet converged upon this state at any given time (especially when there are |
| 24 | * frequent writes and co-primaries are geographically distant). |
| 25 | * |
| 26 | * There are two kinds of database clusters: |
| 27 | * |
| 28 | * - "main" sections, which hold the "core" databases for one or more web sites and are generally |
| 29 | * compact and highly relational (e.g. reads may join with other tables). |
| 30 | * |
| 31 | * Section names are purely for internal runtime state and should only be referred to in |
| 32 | * site configuration, not in source code or datasets. Connect via LBFactory::getReplicaDatabase |
| 33 | * or LoadBalancer::getConnection, which takes a database name. This is automatically resolved |
| 34 | * to a "main" section via the `sectionsByDB` option. Open connections are shared within a given |
| 35 | * section even across different database names, because they share same physical database cluster. |
| 36 | * |
| 37 | * - "external" clusters, which typically hold data that is non-relational (e.g. key/value pairs), |
| 38 | * self-contained (e.g. read queries don't join tables, and transactions don't write to a "main" |
| 39 | * section at the same time), or too bulky to reside in a "main" section (e.g. large binary blobs). |
| 40 | * |
| 41 | * Depending on the use case, an external cluster name may be considered long-term stable and |
| 42 | * referenced inside a primary dataset. External clusters are defined via the `externalLoads` |
| 43 | * option and may be referenced in the `virtualDomainsMapping` option ($wgVirtualDomainsMapping |
| 44 | * in MediaWiki). |
| 45 | * |
| 46 | * If MediaWiki is configured with the $wgExternalStores "DB option, then $wgDefaultExternalStore |
| 47 | * and the immutable rows stored in the `content` or `text` table (via SqlBlobStore and |
| 48 | * ExternalStoreDB) refer to external clusters by name. |
| 49 | * See also <https://www.mediawiki.org/wiki/Manual:External_storage>. |
| 50 | * |
| 51 | * The class allows for large site farms to split up their data in the following ways: |
| 52 | * |
| 53 | * - Vertically shard compact site-specific data by web site (e.g. page metadata) |
| 54 | * - Vertically shard compact global data by module (e.g. account data) |
| 55 | * - Horizontally shard any bulk data by blob key (e.g. page content blobs) |
| 56 | * |
| 57 | * @ingroup Database |
| 58 | */ |
| 59 | class LBFactoryMulti extends LBFactory { |
| 60 | /** @var array<string,ILoadBalancerForOwner> Map of (main section => tracked LoadBalancer) */ |
| 61 | private $mainLBs = []; |
| 62 | /** @var array<string,ILoadBalancerForOwner> Map of (external cluster => tracked LoadBalancer) */ |
| 63 | private $externalLBs = []; |
| 64 | |
| 65 | /** @var string[] Map of (server name => IP address) */ |
| 66 | private $hostsByServerName; |
| 67 | /** @var string[] Map of (database name => main section) */ |
| 68 | private $sectionsByDB; |
| 69 | /** @var int[][] Map of (main section => server name => load ratio) */ |
| 70 | private $sectionLoads; |
| 71 | /** @var int[][] Map of (external cluster => server name => load ratio) */ |
| 72 | private $externalLoadsByCluster; |
| 73 | /** @var array Server config map ("host", "serverName", and "load" ignored) */ |
| 74 | private $serverTemplate; |
| 75 | /** @var array Server config map overriding "serverTemplate" for all external servers */ |
| 76 | private $externalTemplateOverrides; |
| 77 | /** @var array[] Map of (main section => server config map overrides) */ |
| 78 | private $templateOverridesBySection; |
| 79 | /** @var array[] Map of (external cluster => server config map overrides) */ |
| 80 | private $templateOverridesByCluster; |
| 81 | /** @var array Server config override map for all main/external primary DB servers */ |
| 82 | private $masterTemplateOverrides; |
| 83 | /** @var array[] Map of (server name => server config map overrides) for all servers */ |
| 84 | private $templateOverridesByServer; |
| 85 | /** @var string[]|bool[] A map of (main section => read-only message) */ |
| 86 | private $readOnlyBySection; |
| 87 | /** @var array Configuration for the LoadMonitor to use within LoadBalancer instances */ |
| 88 | private $loadMonitorConfig; |
| 89 | /** @var DatabaseDomain[] Map of (domain ID => domain instance) */ |
| 90 | private $nonLocalDomainCache = []; |
| 91 | |
| 92 | /** |
| 93 | * Template override precedence (highest => lowest): |
| 94 | * - templateOverridesByServer |
| 95 | * - masterTemplateOverrides |
| 96 | * - templateOverridesBySection/templateOverridesByCluster |
| 97 | * - externalTemplateOverrides |
| 98 | * - serverTemplate |
| 99 | * Overrides only work on top level keys (so nested values will not be merged). |
| 100 | * |
| 101 | * Server config maps should be of the format Database::factory() requires. |
| 102 | * Additionally, a 'max lag' key should also be set on server maps, indicating how stale the |
| 103 | * data can be before the load balancer tries to avoid using it. The map can have 'is static' |
| 104 | * set to disable blocking replication sync checks (intended for archive servers with |
| 105 | * unchanging data). |
| 106 | * |
| 107 | * @see LBFactory::__construct() |
| 108 | * @param array $conf Additional parameters include: |
| 109 | * - hostsByName: map of (server name => IP address). [optional] |
| 110 | * - sectionsByDB: map of (database => main section). The database name "DEFAULT" is |
| 111 | * interpreted as a catch-all for all databases not otherwise mentioned. If no section |
| 112 | * name is specified for "DEFAULT", then the catch-all section is assumed to be named |
| 113 | * "DEFAULT". [optional] |
| 114 | * - sectionLoads: map of (main section => server name => load ratio); the first host |
| 115 | * listed in each section is the primary DB server for that section. [optional] |
| 116 | * - externalLoads: map of (cluster => server name => load ratio) map. [optional] |
| 117 | * - serverTemplate: server config map for Database::factory(). |
| 118 | * Note that "host", "serverName" and "load" entries will be overridden by "hostsByName". [optional] |
| 119 | * - externalTemplateOverrides: server config map overrides for external stores; |
| 120 | * respects the override precedence described above. [optional] |
| 121 | * - templateOverridesBySection: map of (main section => server config map overrides); |
| 122 | * respects the override precedence described above. [optional] |
| 123 | * - templateOverridesByCluster: map of (external cluster => server config map overrides); |
| 124 | * respects the override precedence described above. [optional] |
| 125 | * - masterTemplateOverrides: server config map overrides for masters; |
| 126 | * respects the override precedence described above. [optional] |
| 127 | * - templateOverridesByServer: map of (server name => server config map overrides); |
| 128 | * respects the override precedence described above and applies to both core |
| 129 | * and external storage. [optional] |
| 130 | * - loadMonitor: LoadMonitor::__construct() parameters with "class" field. [optional] |
| 131 | * - readOnlyBySection: map of (main section => message text or false). |
| 132 | * String values make sections read only, whereas anything else does not |
| 133 | * restrict read/write mode. [optional] |
| 134 | * - configCallback: A callback that returns a conf array that can be passed to |
| 135 | * the reconfigure() method. This will be used to autoReconfigure() to load |
| 136 | * any updated configuration. |
| 137 | */ |
| 138 | public function __construct( array $conf ) { |
| 139 | parent::__construct( $conf ); |
| 140 | |
| 141 | $this->hostsByServerName = $conf['hostsByName'] ?? []; |
| 142 | $this->sectionsByDB = ( $conf['sectionsByDB'] ?? [] ) + [ |
| 143 | self::CLUSTER_MAIN_DEFAULT => self::CLUSTER_MAIN_DEFAULT |
| 144 | ]; |
| 145 | $this->sectionLoads = $conf['sectionLoads'] ?? []; |
| 146 | $this->externalLoadsByCluster = $conf['externalLoads'] ?? []; |
| 147 | $this->serverTemplate = $conf['serverTemplate'] ?? []; |
| 148 | $this->externalTemplateOverrides = $conf['externalTemplateOverrides'] ?? []; |
| 149 | $this->templateOverridesBySection = $conf['templateOverridesBySection'] ?? []; |
| 150 | $this->templateOverridesByCluster = $conf['templateOverridesByCluster'] ?? []; |
| 151 | $this->masterTemplateOverrides = $conf['masterTemplateOverrides'] ?? []; |
| 152 | $this->templateOverridesByServer = $conf['templateOverridesByServer'] ?? []; |
| 153 | $this->readOnlyBySection = $conf['readOnlyBySection'] ?? []; |
| 154 | |
| 155 | if ( isset( $conf['loadMonitor'] ) ) { |
| 156 | $this->loadMonitorConfig = $conf['loadMonitor']; |
| 157 | } elseif ( isset( $conf['loadMonitorClass'] ) ) { // b/c |
| 158 | $this->loadMonitorConfig = [ 'class' => $conf['loadMonitorClass'] ]; |
| 159 | } else { |
| 160 | $this->loadMonitorConfig = [ 'class' => LoadMonitor::class ]; |
| 161 | } |
| 162 | |
| 163 | foreach ( $this->externalLoadsByCluster as $cluster => $_ ) { |
| 164 | if ( isset( $this->sectionLoads[$cluster] ) ) { |
| 165 | throw new LogicException( |
| 166 | "External cluster '$cluster' has the same name as a main section/cluster" |
| 167 | ); |
| 168 | } |
| 169 | } |
| 170 | } |
| 171 | |
| 172 | /** @inheritDoc */ |
| 173 | public function newMainLB( $domain = false ): ILoadBalancerForOwner { |
| 174 | $domainInstance = $this->resolveDomainInstance( $domain ); |
| 175 | $database = $domainInstance->getDatabase(); |
| 176 | $section = $this->getSectionFromDatabase( $database ); |
| 177 | |
| 178 | if ( !isset( $this->sectionLoads[$section] ) ) { |
| 179 | throw new UnexpectedValueException( "Section '$section' has no hosts defined." ); |
| 180 | } |
| 181 | |
| 182 | return $this->newLoadBalancer( |
| 183 | $section, |
| 184 | array_merge( |
| 185 | $this->serverTemplate, |
| 186 | $this->templateOverridesBySection[$section] ?? [] |
| 187 | ), |
| 188 | $this->sectionLoads[$section], |
| 189 | // Use the LB-specific read-only reason if everything isn't already read-only |
| 190 | is_string( $this->readOnlyReason ) |
| 191 | ? $this->readOnlyReason |
| 192 | : ( $this->readOnlyBySection[$section] ?? false ) |
| 193 | ); |
| 194 | } |
| 195 | |
| 196 | /** |
| 197 | * @param DatabaseDomain|string|false $domain |
| 198 | * @return DatabaseDomain |
| 199 | */ |
| 200 | private function resolveDomainInstance( $domain ) { |
| 201 | if ( $domain instanceof DatabaseDomain ) { |
| 202 | return $domain; // already a domain instance |
| 203 | } elseif ( $domain === false || $domain === $this->localDomain->getId() ) { |
| 204 | return $this->localDomain; |
| 205 | } elseif ( isset( $this->domainAliases[$domain] ) ) { |
| 206 | // This array acts as both the original map and as instance cache. |
| 207 | // Instances pass-through DatabaseDomain::newFromId as-is. |
| 208 | $this->domainAliases[$domain] = |
| 209 | DatabaseDomain::newFromId( $this->domainAliases[$domain] ); |
| 210 | |
| 211 | return $this->domainAliases[$domain]; |
| 212 | } |
| 213 | |
| 214 | $cachedDomain = $this->nonLocalDomainCache[$domain] ?? null; |
| 215 | if ( $cachedDomain === null ) { |
| 216 | $cachedDomain = DatabaseDomain::newFromId( $domain ); |
| 217 | $this->nonLocalDomainCache = [ $domain => $cachedDomain ]; |
| 218 | } |
| 219 | |
| 220 | return $cachedDomain; |
| 221 | } |
| 222 | |
| 223 | /** @inheritDoc */ |
| 224 | public function getMainLB( $domain = false ): ILoadBalancer { |
| 225 | $domainInstance = $this->resolveDomainInstance( $domain ); |
| 226 | $section = $this->getSectionFromDatabase( $domainInstance->getDatabase() ); |
| 227 | |
| 228 | if ( !isset( $this->mainLBs[$section] ) ) { |
| 229 | $this->mainLBs[$section] = $this->newMainLB( $domain ); |
| 230 | } |
| 231 | |
| 232 | return $this->mainLBs[$section]; |
| 233 | } |
| 234 | |
| 235 | /** @inheritDoc */ |
| 236 | public function newExternalLB( $cluster ): ILoadBalancerForOwner { |
| 237 | if ( !isset( $this->externalLoadsByCluster[$cluster] ) ) { |
| 238 | throw new InvalidArgumentException( "Unknown cluster '$cluster'" ); |
| 239 | } |
| 240 | return $this->newLoadBalancer( |
| 241 | $cluster, |
| 242 | array_merge( |
| 243 | $this->serverTemplate, |
| 244 | $this->externalTemplateOverrides, |
| 245 | $this->templateOverridesByCluster[$cluster] ?? [] |
| 246 | ), |
| 247 | $this->externalLoadsByCluster[$cluster], |
| 248 | $this->readOnlyReason |
| 249 | ); |
| 250 | } |
| 251 | |
| 252 | /** @inheritDoc */ |
| 253 | public function getExternalLB( $cluster ): ILoadBalancer { |
| 254 | if ( !isset( $this->externalLBs[$cluster] ) ) { |
| 255 | $this->externalLBs[$cluster] = $this->newExternalLB( |
| 256 | $cluster |
| 257 | ); |
| 258 | } |
| 259 | |
| 260 | return $this->externalLBs[$cluster]; |
| 261 | } |
| 262 | |
| 263 | public function getAllMainLBs(): array { |
| 264 | $lbs = []; |
| 265 | foreach ( $this->sectionsByDB as $db => $section ) { |
| 266 | if ( !isset( $lbs[$section] ) ) { |
| 267 | $lbs[$section] = $this->getMainLB( $db ); |
| 268 | } |
| 269 | } |
| 270 | |
| 271 | return $lbs; |
| 272 | } |
| 273 | |
| 274 | public function getAllExternalLBs(): array { |
| 275 | $lbs = []; |
| 276 | foreach ( $this->externalLoadsByCluster as $cluster => $unused ) { |
| 277 | $lbs[$cluster] = $this->getExternalLB( $cluster ); |
| 278 | } |
| 279 | |
| 280 | return $lbs; |
| 281 | } |
| 282 | |
| 283 | /** @inheritDoc */ |
| 284 | protected function getLBsForOwner() { |
| 285 | foreach ( $this->mainLBs as $lb ) { |
| 286 | yield $lb; |
| 287 | } |
| 288 | foreach ( $this->externalLBs as $lb ) { |
| 289 | yield $lb; |
| 290 | } |
| 291 | } |
| 292 | |
| 293 | /** |
| 294 | * Make a new load balancer object based on template and load array |
| 295 | * |
| 296 | * @param string $clusterName |
| 297 | * @param array $serverTemplate |
| 298 | * @param array $loads |
| 299 | * @param string|false $readOnlyReason |
| 300 | * @return LoadBalancer |
| 301 | */ |
| 302 | private function newLoadBalancer( |
| 303 | string $clusterName, |
| 304 | array $serverTemplate, |
| 305 | array $loads, |
| 306 | $readOnlyReason |
| 307 | ) { |
| 308 | $lb = new LoadBalancer( array_merge( |
| 309 | $this->baseLoadBalancerParams(), |
| 310 | [ |
| 311 | 'servers' => $this->makeServerConfigArrays( $serverTemplate, $loads ), |
| 312 | 'loadMonitor' => $this->loadMonitorConfig, |
| 313 | 'readOnlyReason' => $readOnlyReason, |
| 314 | 'clusterName' => $clusterName |
| 315 | ] |
| 316 | ) ); |
| 317 | $this->initLoadBalancer( $lb ); |
| 318 | |
| 319 | return $lb; |
| 320 | } |
| 321 | |
| 322 | /** |
| 323 | * Make a server array as expected by LoadBalancer::__construct() |
| 324 | * |
| 325 | * @param array $serverTemplate Server config map |
| 326 | * @param int[] $loads Map of (server name => load) |
| 327 | * @return array[] List of server config maps |
| 328 | */ |
| 329 | private function makeServerConfigArrays( array $serverTemplate, array $loads ) { |
| 330 | // Get the ordered map of (server name => load); the primary DB server is first |
| 331 | $servers = []; |
| 332 | foreach ( $loads as $serverName => $load ) { |
| 333 | $servers[] = array_merge( |
| 334 | $serverTemplate, |
| 335 | $servers ? [] : $this->masterTemplateOverrides, |
| 336 | $this->templateOverridesByServer[$serverName] ?? [], |
| 337 | [ |
| 338 | 'host' => $this->hostsByServerName[$serverName] ?? $serverName, |
| 339 | 'serverName' => $serverName, |
| 340 | 'load' => $load, |
| 341 | ] |
| 342 | ); |
| 343 | } |
| 344 | |
| 345 | return $servers; |
| 346 | } |
| 347 | |
| 348 | /** |
| 349 | * @param string|null $database |
| 350 | * @return string Main section name |
| 351 | */ |
| 352 | private function getSectionFromDatabase( $database ) { |
| 353 | if ( $database !== null && isset( $this->sectionsByDB[$database] ) ) { |
| 354 | return $this->sectionsByDB[$database]; |
| 355 | } |
| 356 | return $this->sectionsByDB[self::CLUSTER_MAIN_DEFAULT] |
| 357 | ?? self::CLUSTER_MAIN_DEFAULT; |
| 358 | } |
| 359 | |
| 360 | public function reconfigure( array $conf ): void { |
| 361 | if ( !$conf ) { |
| 362 | return; |
| 363 | } |
| 364 | |
| 365 | foreach ( $this->mainLBs as $lb ) { |
| 366 | // Approximate what LBFactoryMulti::__construct does (T346365) |
| 367 | $config = [ |
| 368 | 'servers' => $this->makeServerConfigArrays( |
| 369 | $conf['serverTemplate'] ?? [], |
| 370 | $conf['sectionLoads'][$lb->getClusterName()] |
| 371 | ) |
| 372 | ]; |
| 373 | $lb->reconfigure( $config ); |
| 374 | |
| 375 | } |
| 376 | foreach ( $this->externalLBs as $lb ) { |
| 377 | $config = [ |
| 378 | 'servers' => $this->makeServerConfigArrays( |
| 379 | $conf['serverTemplate'] ?? [], |
| 380 | $conf['externalLoads'][$lb->getClusterName()] |
| 381 | ) |
| 382 | ]; |
| 383 | $lb->reconfigure( $config ); |
| 384 | } |
| 385 | } |
| 386 | } |