3 * Database load balancing interface
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.
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.
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
23 namespace Wikimedia\Rdbms
;
26 use InvalidArgumentException
;
29 * Database cluster connection, tracking, load balancing, and transaction manager interface
31 * A "cluster" is considered to be one master database and zero or more replica databases.
32 * Typically, the replica DBs replicate from the master asynchronously. The first node in the
33 * "servers" configuration array is always considered the "master". However, this class can still
34 * be used when all or some of the "replica" DBs are multi-master peers of the master or even
35 * when all the DBs are non-replicating clones of each other holding read-only data. Thus, the
36 * role of "master" is in some cases merely nominal.
38 * By default, each DB server uses DBO_DEFAULT for its 'flags' setting, unless explicitly set
39 * otherwise in configuration. DBO_DEFAULT behavior depends on whether 'cliMode' is set:
40 * - In CLI mode, the flag has no effect with regards to LoadBalancer.
41 * - In non-CLI mode, the flag causes implicit transactions to be used; the first query on
42 * a database starts a transaction on that database. The transactions are meant to remain
43 * pending until either commitMasterChanges() or rollbackMasterChanges() is called. The
44 * application must have some point where it calls commitMasterChanges() near the end of
46 * Every iteration of beginMasterChanges()/commitMasterChanges() is called a "transaction round".
47 * Rounds are useful on the master DB connections because they make single-DB (and by and large
48 * multi-DB) updates in web requests all-or-nothing. Also, transactions on replica DBs are useful
49 * when REPEATABLE-READ or SERIALIZABLE isolation is used because all foriegn keys and constraints
50 * hold across separate queries in the DB transaction since the data appears within a consistent
51 * point-in-time snapshot.
53 * The typical caller will use LoadBalancer::getConnection( DB_* ) to yield a live database
54 * connection handle. The choice of which DB server to use is based on pre-defined loads for
55 * weighted random selection, adjustments thereof by LoadMonitor, and the amount of replication
56 * lag on each DB server. Lag checks might cause problems in certain setups, so they should be
57 * tuned in the server configuration maps as follows:
58 * - Master + N Replica(s): set 'max lag' to an appropriate threshold for avoiding any database
59 * lagged by this much or more. If all DBs are this lagged, then the load balancer considers
60 * the cluster to be read-only.
61 * - Galera Cluster: Seconds_Behind_Master will be 0, so there probably is nothing to tune.
62 * Note that lag is still possible depending on how wsrep-sync-wait is set server-side.
63 * - Read-only archive clones: set 'is static' in the server configuration maps. This will
64 * treat all such DBs as having 0 lag.
65 * - SQL load balancing proxy: any proxy should handle lag checks on its own, so the 'max lag'
66 * parameter should probably be set to INF in the server configuration maps. This will make
67 * the load balancer ignore whatever it detects as the lag of the logical replica is (which
68 * would probably just randomly bounce around).
70 * If using a SQL proxy service, it would probably be best to have two proxy hosts for the
71 * load balancer to talk to. One would be the 'host' of the master server entry and another for
72 * the (logical) replica server entry. The proxy could map the load balancer's "replica" DB to
73 * any number of physical replica DBs.
78 interface ILoadBalancer
{
79 /** @var int Request a replica DB connection */
80 const DB_REPLICA
= -1;
81 /** @var int Request a master DB connection */
84 /** @var string Domain specifier when no specific database needs to be selected */
85 const DOMAIN_ANY
= '';
87 /** @var int DB handle should have DBO_TRX disabled and the caller will leave it as such */
88 const CONN_TRX_AUTOCOMMIT
= 1;
90 /** @var string Manager of ILoadBalancer instances is running post-commit callbacks */
91 const STAGE_POSTCOMMIT_CALLBACKS
= 'stage-postcommit-callbacks';
92 /** @var string Manager of ILoadBalancer instances is running post-rollback callbacks */
93 const STAGE_POSTROLLBACK_CALLBACKS
= 'stage-postrollback-callbacks';
96 * Construct a manager of IDatabase connection objects
98 * @param array $params Parameter map with keys:
99 * - servers : Required. Array of server info structures.
100 * - localDomain: A DatabaseDomain or domain ID string.
101 * - loadMonitor : Name of a class used to fetch server lag and load.
102 * - readOnlyReason : Reason the master DB is read-only if so [optional]
103 * - waitTimeout : Maximum time to wait for replicas for consistency [optional]
104 * - maxLag: Try to avoid DB replicas with lag above this many seconds [optional]
105 * - srvCache : BagOStuff object for server cache [optional]
106 * - wanCache : WANObjectCache object [optional]
107 * - chronologyCallback: Callback to run before the first connection attempt [optional]
108 * - hostname : The name of the current server [optional]
109 * - cliMode: Whether the execution context is a CLI script. [optional]
110 * - profiler : Class name or instance with profileIn()/profileOut() methods. [optional]
111 * - trxProfiler: TransactionProfiler instance. [optional]
112 * - replLogger: PSR-3 logger instance. [optional]
113 * - connLogger: PSR-3 logger instance. [optional]
114 * - queryLogger: PSR-3 logger instance. [optional]
115 * - perfLogger: PSR-3 logger instance. [optional]
116 * - errorLogger : Callback that takes an Exception and logs it. [optional]
117 * - deprecationLogger: Callback to log a deprecation warning. [optional]
118 * - roundStage: STAGE_POSTCOMMIT_* class constant; for internal use [optional]
119 * - ownerId: integer ID of an LBFactory instance that manages this instance [optional]
120 * @throws InvalidArgumentException
122 public function __construct( array $params );
125 * Get the local (and default) database domain ID of connection handles
127 * @see DatabaseDomain
128 * @return string Database domain ID; this specifies DB name, schema, and table prefix
131 public function getLocalDomainID();
134 * @param DatabaseDomain|string|bool $domain Database domain
135 * @return string Value of $domain if it is foreign or the local domain otherwise
138 public function resolveDomainID( $domain );
141 * Close all connection and redefine the local domain for testing or schema creation
143 * @param DatabaseDomain|string $domain
146 public function redefineLocalDomain( $domain );
149 * Get the index of the reader connection, which may be a replica DB
151 * This takes into account load ratios and lag times. It should
152 * always return a consistent index during a given invocation.
154 * Side effect: opens connections to databases
155 * @param string|bool $group Query group, or false for the generic group
156 * @param string|bool $domain Domain ID, or false for the current domain
158 * @return bool|int|string
160 public function getReaderIndex( $group = false, $domain = false );
163 * Set the master position to reach before the next generic group DB handle query
165 * If a generic replica DB connection is already open then this immediately waits
166 * for that DB to catch up to the specified replication position. Otherwise, it will
167 * do so once such a connection is opened.
169 * If a timeout happens when waiting, then getLaggedReplicaMode()/laggedReplicaUsed()
172 * @param DBMasterPos|bool $pos Master position or false
174 public function waitFor( $pos );
177 * Set the master wait position and wait for a generic replica DB to catch up to it
179 * This can be used a faster proxy for waitForAll()
181 * @param DBMasterPos|bool $pos Master position or false
182 * @param int|null $timeout Max seconds to wait; default is mWaitTimeout
183 * @return bool Success (able to connect and no timeouts reached)
185 public function waitForOne( $pos, $timeout = null );
188 * Set the master wait position and wait for ALL replica DBs to catch up to it
190 * @param DBMasterPos|bool $pos Master position or false
191 * @param int|null $timeout Max seconds to wait; default is mWaitTimeout
192 * @return bool Success (able to connect and no timeouts reached)
194 public function waitForAll( $pos, $timeout = null );
197 * Get any open connection to a given server index, local or foreign
199 * Use CONN_TRX_AUTOCOMMIT to only look for connections opened with that flag.
200 * Avoid the use of begin() or startAtomic() on any such connections.
202 * @param int $i Server index or DB_MASTER/DB_REPLICA
203 * @param int $flags Bitfield of CONN_* class constants
204 * @return Database|bool False if no such connection is open
206 public function getAnyOpenConnection( $i, $flags = 0 );
209 * Get a connection handle by server index
211 * The CONN_TRX_AUTOCOMMIT flag is ignored for databases with ATTR_DB_LEVEL_LOCKING
212 * (e.g. sqlite) in order to avoid deadlocks. ILoadBalancer::getServerAttributes()
213 * can be used to check such flags beforehand.
215 * If the caller uses $domain or sets CONN_TRX_AUTOCOMMIT in $flags, then it must
216 * also call ILoadBalancer::reuseConnection() on the handle when finished using it.
217 * In all other cases, this is not necessary, though not harmful either.
218 * Avoid the use of begin() or startAtomic() on any such connections.
220 * @param int $i Server index (overrides $groups) or DB_MASTER/DB_REPLICA
221 * @param array|string|bool $groups Query group(s), or false for the generic reader
222 * @param string|bool $domain Domain ID, or false for the current domain
223 * @param int $flags Bitfield of CONN_* class constants
225 * @note This method throws DBAccessError if ILoadBalancer::disable() was called
230 public function getConnection( $i, $groups = [], $domain = false, $flags = 0 );
233 * Mark a foreign connection as being available for reuse under a different DB domain
235 * This mechanism is reference-counted, and must be called the same number of times
236 * as getConnection() to work.
238 * @param IDatabase $conn
239 * @throws InvalidArgumentException
241 public function reuseConnection( IDatabase
$conn );
244 * Get a database connection handle reference
246 * The handle's methods simply wrap those of a Database handle
248 * The CONN_TRX_AUTOCOMMIT flag is ignored for databases with ATTR_DB_LEVEL_LOCKING
249 * (e.g. sqlite) in order to avoid deadlocks. ILoadBalancer::getServerAttributes()
250 * can be used to check such flags beforehand. Avoid the use of begin() or startAtomic()
251 * on any CONN_TRX_AUTOCOMMIT connections.
253 * @see ILoadBalancer::getConnection() for parameter information
255 * @param int $i Server index or DB_MASTER/DB_REPLICA
256 * @param array|string|bool $groups Query group(s), or false for the generic reader
257 * @param string|bool $domain Domain ID, or false for the current domain
258 * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTOCOMMIT)
261 public function getConnectionRef( $i, $groups = [], $domain = false, $flags = 0 );
264 * Get a database connection handle reference without connecting yet
266 * The handle's methods simply wrap those of a Database handle
268 * The CONN_TRX_AUTOCOMMIT flag is ignored for databases with ATTR_DB_LEVEL_LOCKING
269 * (e.g. sqlite) in order to avoid deadlocks. ILoadBalancer::getServerAttributes()
270 * can be used to check such flags beforehand. Avoid the use of begin() or startAtomic()
271 * on any CONN_TRX_AUTOCOMMIT connections.
273 * @see ILoadBalancer::getConnection() for parameter information
275 * @param int $i Server index or DB_MASTER/DB_REPLICA
276 * @param array|string|bool $groups Query group(s), or false for the generic reader
277 * @param string|bool $domain Domain ID, or false for the current domain
278 * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTOCOMMIT)
281 public function getLazyConnectionRef( $i, $groups = [], $domain = false, $flags = 0 );
284 * Get a maintenance database connection handle reference for migrations and schema changes
286 * The handle's methods simply wrap those of a Database handle
288 * The CONN_TRX_AUTOCOMMIT flag is ignored for databases with ATTR_DB_LEVEL_LOCKING
289 * (e.g. sqlite) in order to avoid deadlocks. ILoadBalancer::getServerAttributes()
290 * can be used to check such flags beforehand. Avoid the use of begin() or startAtomic()
291 * on any CONN_TRX_AUTOCOMMIT connections.
293 * @see ILoadBalancer::getConnection() for parameter information
295 * @param int $i Server index or DB_MASTER/DB_REPLICA
296 * @param array|string|bool $groups Query group(s), or false for the generic reader
297 * @param string|bool $domain Domain ID, or false for the current domain
298 * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTOCOMMIT)
299 * @return MaintainableDBConnRef
301 public function getMaintenanceConnectionRef( $i, $groups = [], $domain = false, $flags = 0 );
304 * Open a connection to the server given by the specified index
306 * The index must be an actual index into the array. If a connection to the server is
307 * already open and not considered an "in use" foreign connection, this simply returns it.
309 * Avoid using CONN_TRX_AUTOCOMMIT for databases with ATTR_DB_LEVEL_LOCKING (e.g. sqlite)
310 * in order to avoid deadlocks. ILoadBalancer::getServerAttributes() can be used to check
311 * such flags beforehand.
313 * If the caller uses $domain or sets CONN_TRX_AUTOCOMMIT in $flags, then it must
314 * also call ILoadBalancer::reuseConnection() on the handle when finished using it.
315 * In all other cases, this is not necessary, though not harmful either.
316 * Avoid the use of begin() or startAtomic() on any such connections.
318 * @note This method throws DBAccessError if ILoadBalancer::disable() was called
320 * @param int $i Server index (does not support DB_MASTER/DB_REPLICA)
321 * @param string|bool $domain Domain ID, or false for the current domain
322 * @param int $flags Bitfield of CONN_* class constants (e.g. CONN_TRX_AUTOCOMMIT)
323 * @return Database|bool Returns false on errors
324 * @throws DBAccessError
326 public function openConnection( $i, $domain = false, $flags = 0 );
331 public function getWriterIndex();
334 * Returns true if the specified index is a valid server index
339 public function haveIndex( $i );
342 * Returns true if the specified index is valid and has non-zero load
347 public function isNonZeroLoad( $i );
350 * Get the number of defined servers (not the number of open connections)
354 public function getServerCount();
357 * Get the host name or IP address of the server with the specified index
360 * @return string Readable name if available or IP/host otherwise
362 public function getServerName( $i );
365 * Return the server info structure for a given index, or false if the index is invalid.
370 public function getServerInfo( $i );
373 * Get DB type of the server with the specified index
376 * @return string One of (mysql,postgres,sqlite,...) or "unknown" for bad indexes
379 public function getServerType( $i );
382 * @param int $i Server index
383 * @return array (Database::ATTRIBUTE_* constant => value) for all such constants
386 public function getServerAttributes( $i );
389 * Get the current master position for chronology control purposes
390 * @return DBMasterPos|bool Returns false if not applicable
392 public function getMasterPos();
395 * Disable this load balancer. All connections are closed, and any attempt to
396 * open a new connection will result in a DBAccessError.
398 public function disable();
401 * Close all open connections
403 public function closeAll();
408 * Using this function makes sure the LoadBalancer knows the connection is closed.
409 * If you use $conn->close() directly, the load balancer won't update its state.
411 * @param IDatabase $conn
413 public function closeConnection( IDatabase
$conn );
416 * Commit transactions on all open connections
417 * @param string $fname Caller name
418 * @param int|null $owner ID of the calling instance (e.g. the LBFactory ID)
419 * @throws DBExpectedError
421 public function commitAll( $fname = __METHOD__
, $owner = null );
424 * Run pre-commit callbacks and defer execution of post-commit callbacks
426 * Use this only for mutli-database commits
428 * @param string $fname Caller name
429 * @param int|null $owner ID of the calling instance (e.g. the LBFactory ID)
430 * @return int Number of pre-commit callbacks run (since 1.32)
432 public function finalizeMasterChanges( $fname = __METHOD__
, $owner = null );
435 * Perform all pre-commit checks for things like replication safety
437 * Use this only for mutli-database commits
439 * @param array $options Includes:
440 * - maxWriteDuration : max write query duration time in seconds
441 * @param string $fname Caller name
442 * @param int|null $owner ID of the calling instance (e.g. the LBFactory ID)
443 * @throws DBTransactionError
445 public function approveMasterChanges( array $options, $fname, $owner = null );
448 * Flush any master transaction snapshots and set DBO_TRX (if DBO_DEFAULT is set)
450 * The DBO_TRX setting will be reverted to the default in each of these methods:
451 * - commitMasterChanges()
452 * - rollbackMasterChanges()
454 * This allows for custom transaction rounds from any outer transaction scope.
456 * @param string $fname Caller name
457 * @param int|null $owner ID of the calling instance (e.g. the LBFactory ID)
458 * @throws DBExpectedError
460 public function beginMasterChanges( $fname = __METHOD__
, $owner = null );
463 * Issue COMMIT on all open master connections to flush changes and view snapshots
464 * @param string $fname Caller name
465 * @param int|null $owner ID of the calling instance (e.g. the LBFactory ID)
466 * @throws DBExpectedError
468 public function commitMasterChanges( $fname = __METHOD__
, $owner = null );
471 * Consume and run all pending post-COMMIT/ROLLBACK callbacks and commit dangling transactions
473 * @param string $fname Caller name
474 * @param int|null $owner ID of the calling instance (e.g. the LBFactory ID)
475 * @return Exception|null The first exception or null if there were none
477 public function runMasterTransactionIdleCallbacks( $fname = __METHOD__
, $owner = null );
480 * Run all recurring post-COMMIT/ROLLBACK listener callbacks
482 * @param string $fname Caller name
483 * @param int|null $owner ID of the calling instance (e.g. the LBFactory ID)
484 * @return Exception|null The first exception or null if there were none
486 public function runMasterTransactionListenerCallbacks( $fname = __METHOD__
, $owner = null );
489 * Issue ROLLBACK only on master, only if queries were done on connection
490 * @param string $fname Caller name
491 * @param int|null $owner ID of the calling instance (e.g. the LBFactory ID)
492 * @throws DBExpectedError
494 public function rollbackMasterChanges( $fname = __METHOD__
, $owner = null );
497 * Commit all replica DB transactions so as to flush any REPEATABLE-READ or SSI snapshots
499 * @param string $fname Caller name
501 public function flushReplicaSnapshots( $fname = __METHOD__
);
504 * Commit all master DB transactions so as to flush any REPEATABLE-READ or SSI snapshots
506 * An error will be thrown if a connection has pending writes or callbacks
508 * @param string $fname Caller name
510 public function flushMasterSnapshots( $fname = __METHOD__
);
513 * @return bool Whether a master connection is already open
515 public function hasMasterConnection();
518 * Whether there are pending changes or callbacks in a transaction by this thread
521 public function hasMasterChanges();
524 * Get the timestamp of the latest write query done by this thread
525 * @return float|bool UNIX timestamp or false
527 public function lastMasterChangeTimestamp();
530 * Check if this load balancer object had any recent or still
531 * pending writes issued against it by this PHP thread
533 * @param float|null $age How many seconds ago is "recent" [defaults to mWaitTimeout]
536 public function hasOrMadeRecentMasterChanges( $age = null );
539 * Get the list of callers that have pending master changes
541 * @return string[] List of method names
543 public function pendingMasterChangeCallers();
546 * @note This method will trigger a DB connection if not yet done
547 * @param string|bool $domain Domain ID, or false for the current domain
548 * @return bool Whether the database for generic connections this request is highly "lagged"
550 public function getLaggedReplicaMode( $domain = false );
553 * Checks whether the database for generic connections this request was both:
554 * - a) Already choosen due to a prior connection attempt
555 * - b) Considered highly "lagged"
557 * @note This method will never cause a new DB connection
560 public function laggedReplicaUsed();
563 * @note This method may trigger a DB connection if not yet done
564 * @param string|bool $domain Domain ID, or false for the current domain
565 * @param IDatabase|null $conn DB master connection; used to avoid loops [optional]
566 * @return string|bool Reason the master is read-only or false if it is not
568 public function getReadOnlyReason( $domain = false, IDatabase
$conn = null );
571 * Disables/enables lag checks
572 * @param null|bool $mode
575 public function allowLagged( $mode = null );
580 public function pingAll();
583 * Call a function with each open connection object
584 * @param callable $callback
585 * @param array $params
587 public function forEachOpenConnection( $callback, array $params = [] );
590 * Call a function with each open connection object to a master
591 * @param callable $callback
592 * @param array $params
594 public function forEachOpenMasterConnection( $callback, array $params = [] );
597 * Call a function with each open replica DB connection object
598 * @param callable $callback
599 * @param array $params
601 public function forEachOpenReplicaConnection( $callback, array $params = [] );
604 * Get the hostname and lag time of the most-lagged replica DB
606 * This is useful for maintenance scripts that need to throttle their updates.
607 * May attempt to open connections to replica DBs on the default DB. If there is
608 * no lag, the maximum lag will be reported as -1.
610 * @param bool|string $domain Domain ID, or false for the default database
611 * @return array ( host, max lag, index of max lagged host )
613 public function getMaxLag( $domain = false );
616 * Get an estimate of replication lag (in seconds) for each server
618 * Results are cached for a short time in memcached/process cache
620 * Values may be "false" if replication is too broken to estimate
622 * @param string|bool $domain
623 * @return int[] Map of (server index => float|int|bool)
625 public function getLagTimes( $domain = false );
628 * Get the lag in seconds for a given connection, or zero if this load
629 * balancer does not have replication enabled.
631 * This should be used in preference to Database::getLag() in cases where
632 * replication may not be in use, since there is no way to determine if
633 * replication is in use at the connection level without running
634 * potentially restricted queries such as SHOW SLAVE STATUS. Using this
635 * function instead of Database::getLag() avoids a fatal error in this
636 * case on many installations.
638 * @param IDatabase $conn
639 * @return int|bool Returns false on error
641 public function safeGetLag( IDatabase
$conn );
644 * Wait for a replica DB to reach a specified master position
646 * This will connect to the master to get an accurate position if $pos is not given
648 * @param IDatabase $conn Replica DB
649 * @param DBMasterPos|bool $pos Master position; default: current position
650 * @param int $timeout Timeout in seconds [optional]
651 * @return bool Success
653 public function safeWaitForMasterPos( IDatabase
$conn, $pos = false, $timeout = 10 );
656 * Set a callback via IDatabase::setTransactionListener() on
657 * all current and future master connections of this load balancer
659 * @param string $name Callback name
660 * @param callable|null $callback
662 public function setTransactionListener( $name, callable
$callback = null );
665 * Set a new table prefix for the existing local domain ID for testing
667 * @param string $prefix
670 public function setLocalDomainPrefix( $prefix );
673 * Make certain table names use their own database, schema, and table prefix
674 * when passed into SQL queries pre-escaped and without a qualified database name
676 * For example, "user" can be converted to "myschema.mydbname.user" for convenience.
677 * Appearances like `user`, somedb.user, somedb.someschema.user will used literally.
679 * Calling this twice will completely clear any old table aliases. Also, note that
680 * callers are responsible for making sure the schemas and databases actually exist.
682 * @param array[] $aliases Map of (table => (dbname, schema, prefix) map)
684 public function setTableAliases( array $aliases );
687 * Convert certain index names to alternative names before querying the DB
689 * Note that this applies to indexes regardless of the table they belong to.
691 * This can be employed when an index was renamed X => Y in code, but the new Y-named
692 * indexes were not yet built on all DBs. After all the Y-named ones are added by the DBA,
693 * the aliases can be removed, and then the old X-named indexes dropped.
695 * @param string[] $aliases
698 public function setIndexAliases( array $aliases );