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
22 * @author Aaron Schulz
24 namespace Wikimedia\Rdbms
;
28 use MaintainableDBConnRef
;
31 use DBTransactionError
;
34 use InvalidArgumentException
;
37 * Database cluster connection, tracking, load balancing, and transaction manager interface
39 * A "cluster" is considered to be one master database and zero or more replica databases.
40 * Typically, the replica DBs replicate from the master asynchronously. The first node in the
41 * "servers" configuration array is always considered the "master". However, this class can still
42 * be used when all or some of the "replica" DBs are multi-master peers of the master or even
43 * when all the DBs are non-replicating clones of each other holding read-only data. Thus, the
44 * role of "master" is in some cases merely nominal.
46 * By default, each DB server uses DBO_DEFAULT for its 'flags' setting, unless explicitly set
47 * otherwise in configuration. DBO_DEFAULT behavior depends on whether 'cliMode' is set:
48 * - In CLI mode, the flag has no effect with regards to LoadBalancer.
49 * - In non-CLI mode, the flag causes implicit transactions to be used; the first query on
50 * a database starts a transaction on that database. The transactions are meant to remain
51 * pending until either commitMasterChanges() or rollbackMasterChanges() is called. The
52 * application must have some point where it calls commitMasterChanges() near the end of
54 * Every iteration of beginMasterChanges()/commitMasterChanges() is called a "transaction round".
55 * Rounds are useful on the master DB connections because they make single-DB (and by and large
56 * multi-DB) updates in web requests all-or-nothing. Also, transactions on replica DBs are useful
57 * when REPEATABLE-READ or SERIALIZABLE isolation is used because all foriegn keys and constraints
58 * hold across separate queries in the DB transaction since the data appears within a consistent
59 * point-in-time snapshot.
61 * The typical caller will use LoadBalancer::getConnection( DB_* ) to yield a live database
62 * connection handle. The choice of which DB server to use is based on pre-defined loads for
63 * weighted random selection, adjustments thereof by LoadMonitor, and the amount of replication
64 * lag on each DB server. Lag checks might cause problems in certain setups, so they should be
65 * tuned in the server configuration maps as follows:
66 * - Master + N Replica(s): set 'max lag' to an appropriate threshold for avoiding any database
67 * lagged by this much or more. If all DBs are this lagged, then the load balancer considers
68 * the cluster to be read-only.
69 * - Galera Cluster: Seconds_Behind_Master will be 0, so there probably is nothing to tune.
70 * Note that lag is still possible depending on how wsrep-sync-wait is set server-side.
71 * - Read-only archive clones: set 'is static' in the server configuration maps. This will
72 * treat all such DBs as having 0 lag.
73 * - SQL load balancing proxy: any proxy should handle lag checks on its own, so the 'max lag'
74 * parameter should probably be set to INF in the server configuration maps. This will make
75 * the load balancer ignore whatever it detects as the lag of the logical replica is (which
76 * would probably just randomly bounce around).
78 * If using a SQL proxy service, it would probably be best to have two proxy hosts for the
79 * load balancer to talk to. One would be the 'host' of the master server entry and another for
80 * the (logical) replica server entry. The proxy could map the load balancer's "replica" DB to
81 * any number of physical replica DBs.
86 interface ILoadBalancer
{
87 /** @var integer Request a replica DB connection */
88 const DB_REPLICA
= -1;
89 /** @var integer Request a master DB connection */
92 /** @var string Domain specifier when no specific database needs to be selected */
93 const DOMAIN_ANY
= '';
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 * - srvCache : BagOStuff object for server cache [optional]
105 * - memCache : BagOStuff object for cluster memory cache [optional]
106 * - wanCache : WANObjectCache object [optional]
107 * - hostname : The name of the current server [optional]
108 * - cliMode: Whether the execution context is a CLI script. [optional]
109 * - profiler : Class name or instance with profileIn()/profileOut() methods. [optional]
110 * - trxProfiler: TransactionProfiler instance. [optional]
111 * - replLogger: PSR-3 logger instance. [optional]
112 * - connLogger: PSR-3 logger instance. [optional]
113 * - queryLogger: PSR-3 logger instance. [optional]
114 * - perfLogger: PSR-3 logger instance. [optional]
115 * - errorLogger : Callback that takes an Exception and logs it. [optional]
116 * @throws InvalidArgumentException
118 public function __construct( array $params );
121 * Get the index of the reader connection, which may be a replica DB
123 * This takes into account load ratios and lag times. It should
124 * always return a consistent index during a given invocation.
126 * Side effect: opens connections to databases
127 * @param string|bool $group Query group, or false for the generic reader
128 * @param string|bool $domain Domain ID, or false for the current domain
130 * @return bool|int|string
132 public function getReaderIndex( $group = false, $domain = false );
135 * Set the master wait position
137 * If a DB_REPLICA connection has been opened already, then wait immediately.
138 * Otherwise sets a variable telling it to wait if such a connection is opened.
140 * @param DBMasterPos $pos
142 public function waitFor( $pos );
145 * Set the master wait position and wait for a "generic" replica DB to catch up to it
147 * This can be used a faster proxy for waitForAll()
149 * @param DBMasterPos $pos
150 * @param int $timeout Max seconds to wait; default is mWaitTimeout
151 * @return bool Success (able to connect and no timeouts reached)
153 public function waitForOne( $pos, $timeout = null );
156 * Set the master wait position and wait for ALL replica DBs to catch up to it
158 * @param DBMasterPos $pos
159 * @param int $timeout Max seconds to wait; default is mWaitTimeout
160 * @return bool Success (able to connect and no timeouts reached)
162 public function waitForAll( $pos, $timeout = null );
165 * Get any open connection to a given server index, local or foreign
167 * @param int $i Server index or DB_MASTER/DB_REPLICA
168 * @return Database|bool False if no such connection is open
170 public function getAnyOpenConnection( $i );
173 * Get a connection by index
175 * @param int $i Server index or DB_MASTER/DB_REPLICA
176 * @param array|string|bool $groups Query group(s), or false for the generic reader
177 * @param string|bool $domain Domain ID, or false for the current domain
182 public function getConnection( $i, $groups = [], $domain = false );
185 * Mark a foreign connection as being available for reuse under a different DB domain
187 * This mechanism is reference-counted, and must be called the same number of times
188 * as getConnection() to work.
190 * @param IDatabase $conn
191 * @throws InvalidArgumentException
193 public function reuseConnection( $conn );
196 * Get a database connection handle reference
198 * The handle's methods simply wrap those of a Database handle
200 * @see ILoadBalancer::getConnection() for parameter information
202 * @param int $i Server index or DB_MASTER/DB_REPLICA
203 * @param array|string|bool $groups Query group(s), or false for the generic reader
204 * @param string|bool $domain Domain ID, or false for the current domain
207 public function getConnectionRef( $i, $groups = [], $domain = false );
210 * Get a database connection handle reference without connecting yet
212 * The handle's methods simply wrap those of a Database handle
214 * @see ILoadBalancer::getConnection() for parameter information
216 * @param int $i Server index or DB_MASTER/DB_REPLICA
217 * @param array|string|bool $groups Query group(s), or false for the generic reader
218 * @param string|bool $domain Domain ID, or false for the current domain
221 public function getLazyConnectionRef( $i, $groups = [], $domain = false );
224 * Get a maintenance database connection handle reference for migrations and schema changes
226 * The handle's methods simply wrap those of a Database handle
228 * @see ILoadBalancer::getConnection() for parameter information
230 * @param int $db Server index or DB_MASTER/DB_REPLICA
231 * @param array|string|bool $groups Query group(s), or false for the generic reader
232 * @param string|bool $domain Domain ID, or false for the current domain
233 * @return MaintainableDBConnRef
235 public function getMaintenanceConnectionRef( $db, $groups = [], $domain = false );
238 * Open a connection to the server given by the specified index
239 * Index must be an actual index into the array.
240 * If the server is already open, returns it.
242 * @note If disable() was called on this LoadBalancer, this method will throw a DBAccessError.
244 * @param int $i Server index or DB_MASTER/DB_REPLICA
245 * @param string|bool $domain Domain ID, or false for the current domain
246 * @return Database|bool Returns false on errors
247 * @throws DBAccessError
249 public function openConnection( $i, $domain = false );
254 public function getWriterIndex();
257 * Returns true if the specified index is a valid server index
262 public function haveIndex( $i );
265 * Returns true if the specified index is valid and has non-zero load
270 public function isNonZeroLoad( $i );
273 * Get the number of defined servers (not the number of open connections)
277 public function getServerCount();
280 * Get the host name or IP address of the server with the specified index
281 * Prefer a readable name if available.
285 public function getServerName( $i );
288 * Return the server info structure for a given index, or false if the index is invalid.
292 public function getServerInfo( $i );
295 * Sets the server info structure for the given index. Entry at index $i
296 * is created if it doesn't exist
298 * @param array $serverInfo
300 public function setServerInfo( $i, array $serverInfo );
303 * Get the current master position for chronology control purposes
304 * @return DBMasterPos|bool Returns false if not applicable
306 public function getMasterPos();
309 * Disable this load balancer. All connections are closed, and any attempt to
310 * open a new connection will result in a DBAccessError.
312 public function disable();
315 * Close all open connections
317 public function closeAll();
322 * Using this function makes sure the LoadBalancer knows the connection is closed.
323 * If you use $conn->close() directly, the load balancer won't update its state.
325 * @param IDatabase $conn
327 public function closeConnection( IDatabase
$conn );
330 * Commit transactions on all open connections
331 * @param string $fname Caller name
332 * @throws DBExpectedError
334 public function commitAll( $fname = __METHOD__
);
337 * Perform all pre-commit callbacks that remain part of the atomic transactions
338 * and disable any post-commit callbacks until runMasterPostTrxCallbacks()
340 * Use this only for mutli-database commits
342 public function finalizeMasterChanges();
345 * Perform all pre-commit checks for things like replication safety
347 * Use this only for mutli-database commits
349 * @param array $options Includes:
350 * - maxWriteDuration : max write query duration time in seconds
351 * @throws DBTransactionError
353 public function approveMasterChanges( array $options );
356 * Flush any master transaction snapshots and set DBO_TRX (if DBO_DEFAULT is set)
358 * The DBO_TRX setting will be reverted to the default in each of these methods:
359 * - commitMasterChanges()
360 * - rollbackMasterChanges()
362 * This allows for custom transaction rounds from any outer transaction scope.
364 * @param string $fname
365 * @throws DBExpectedError
367 public function beginMasterChanges( $fname = __METHOD__
);
370 * Issue COMMIT on all master connections where writes where done
371 * @param string $fname Caller name
372 * @throws DBExpectedError
374 public function commitMasterChanges( $fname = __METHOD__
);
377 * Issue all pending post-COMMIT/ROLLBACK callbacks
379 * Use this only for mutli-database commits
381 * @param int $type IDatabase::TRIGGER_* constant
382 * @return Exception|null The first exception or null if there were none
384 public function runMasterPostTrxCallbacks( $type );
387 * Issue ROLLBACK only on master, only if queries were done on connection
388 * @param string $fname Caller name
389 * @throws DBExpectedError
391 public function rollbackMasterChanges( $fname = __METHOD__
);
394 * Suppress all pending post-COMMIT/ROLLBACK callbacks
396 * Use this only for mutli-database commits
398 * @return Exception|null The first exception or null if there were none
400 public function suppressTransactionEndCallbacks();
403 * Commit all replica DB transactions so as to flush any REPEATABLE-READ or SSI snapshot
405 * @param string $fname Caller name
407 public function flushReplicaSnapshots( $fname = __METHOD__
);
410 * @return bool Whether a master connection is already open
412 public function hasMasterConnection();
415 * Determine if there are pending changes in a transaction by this thread
418 public function hasMasterChanges();
421 * Get the timestamp of the latest write query done by this thread
422 * @return float|bool UNIX timestamp or false
424 public function lastMasterChangeTimestamp();
427 * Check if this load balancer object had any recent or still
428 * pending writes issued against it by this PHP thread
430 * @param float $age How many seconds ago is "recent" [defaults to mWaitTimeout]
433 public function hasOrMadeRecentMasterChanges( $age = null );
436 * Get the list of callers that have pending master changes
438 * @return string[] List of method names
440 public function pendingMasterChangeCallers();
443 * @note This method will trigger a DB connection if not yet done
444 * @param string|bool $domain Domain ID, or false for the current domain
445 * @return bool Whether the generic connection for reads is highly "lagged"
447 public function getLaggedReplicaMode( $domain = false );
450 * @note This method will never cause a new DB connection
451 * @return bool Whether any generic connection used for reads was highly "lagged"
453 public function laggedReplicaUsed();
456 * @note This method may trigger a DB connection if not yet done
457 * @param string|bool $domain Domain ID, or false for the current domain
458 * @param IDatabase|null $conn DB master connection; used to avoid loops [optional]
459 * @return string|bool Reason the master is read-only or false if it is not
461 public function getReadOnlyReason( $domain = false, IDatabase
$conn = null );
464 * Disables/enables lag checks
465 * @param null|bool $mode
468 public function allowLagged( $mode = null );
473 public function pingAll();
476 * Call a function with each open connection object
477 * @param callable $callback
478 * @param array $params
480 public function forEachOpenConnection( $callback, array $params = [] );
483 * Call a function with each open connection object to a master
484 * @param callable $callback
485 * @param array $params
487 public function forEachOpenMasterConnection( $callback, array $params = [] );
490 * Call a function with each open replica DB connection object
491 * @param callable $callback
492 * @param array $params
494 public function forEachOpenReplicaConnection( $callback, array $params = [] );
497 * Get the hostname and lag time of the most-lagged replica DB
499 * This is useful for maintenance scripts that need to throttle their updates.
500 * May attempt to open connections to replica DBs on the default DB. If there is
501 * no lag, the maximum lag will be reported as -1.
503 * @param bool|string $domain Domain ID, or false for the default database
504 * @return array ( host, max lag, index of max lagged host )
506 public function getMaxLag( $domain = false );
509 * Get an estimate of replication lag (in seconds) for each server
511 * Results are cached for a short time in memcached/process cache
513 * Values may be "false" if replication is too broken to estimate
515 * @param string|bool $domain
516 * @return int[] Map of (server index => float|int|bool)
518 public function getLagTimes( $domain = false );
521 * Get the lag in seconds for a given connection, or zero if this load
522 * balancer does not have replication enabled.
524 * This should be used in preference to Database::getLag() in cases where
525 * replication may not be in use, since there is no way to determine if
526 * replication is in use at the connection level without running
527 * potentially restricted queries such as SHOW SLAVE STATUS. Using this
528 * function instead of Database::getLag() avoids a fatal error in this
529 * case on many installations.
531 * @param IDatabase $conn
532 * @return int|bool Returns false on error
534 public function safeGetLag( IDatabase
$conn );
537 * Wait for a replica DB to reach a specified master position
539 * This will connect to the master to get an accurate position if $pos is not given
541 * @param IDatabase $conn Replica DB
542 * @param DBMasterPos|bool $pos Master position; default: current position
543 * @param int $timeout Timeout in seconds [optional]
544 * @return bool Success
546 public function safeWaitForMasterPos( IDatabase
$conn, $pos = false, $timeout = 10 );
549 * Set a callback via IDatabase::setTransactionListener() on
550 * all current and future master connections of this load balancer
552 * @param string $name Callback name
553 * @param callable|null $callback
555 public function setTransactionListener( $name, callable
$callback = null );
558 * Set a new table prefix for the existing local domain ID for testing
560 * @param string $prefix
562 public function setDomainPrefix( $prefix );
565 * Make certain table names use their own database, schema, and table prefix
566 * when passed into SQL queries pre-escaped and without a qualified database name
568 * For example, "user" can be converted to "myschema.mydbname.user" for convenience.
569 * Appearances like `user`, somedb.user, somedb.someschema.user will used literally.
571 * Calling this twice will completely clear any old table aliases. Also, note that
572 * callers are responsible for making sure the schemas and databases actually exist.
574 * @param array[] $aliases Map of (table => (dbname, schema, prefix) map)
576 public function setTableAliases( array $aliases );