13.7.6.3 FLUSH Statement

FLUSH [NO_WRITE_TO_BINLOG | LOCAL] {
    flush_option [, flush_option] ...
  | tables_option
}

flush_option: {
    BINARY LOGS
  | DES_KEY_FILE
  | ENGINE LOGS
  | ERROR LOGS
  | GENERAL LOGS
  | HOSTS
  | LOGS
  | PRIVILEGES
  | QUERY CACHE
  | RELAY LOGS
  | SLOW LOGS
  | STATUS
  | USER_RESOURCES
}

tables_option: {
    TABLES
  | TABLES tbl_name [, tbl_name] ...
  | TABLES WITH READ LOCK
  | TABLES tbl_name [, tbl_name] ... WITH READ LOCK
  | TABLES tbl_name [, tbl_name] ... FOR EXPORT
}

The FLUSH statement has several variant forms that clear or reload various internal caches, flush tables, or acquire locks. To execute FLUSH, you must have the RELOAD privilege. Specific flush options might require additional privileges, as described later.

Note

It is not possible to issue FLUSH statements within stored functions or triggers. However, you may use FLUSH in stored procedures, so long as these are not called from stored functions or triggers. See Section 20.8, “Restrictions on Stored Programs”.

By default, the server writes FLUSH statements to the binary log so that they replicate to replication slaves. To suppress logging, specify the optional NO_WRITE_TO_BINLOG keyword or its alias LOCAL.

Note

FLUSH LOGS, FLUSH BINARY LOGS, FLUSH TABLES WITH READ LOCK (with or without a table list), and FLUSH TABLES tbl_name ... FOR EXPORT are not written to the binary log in any case because they would cause problems if replicated to a slave.

The FLUSH statement causes an implicit commit. See Section 13.3.3, “Statements That Cause an Implicit Commit”.

The mysqladmin utility provides a command-line interface to some flush operations, using commands such as flush-hosts, flush-logs, flush-privileges, flush-status, and flush-tables. See Section 4.5.2, “mysqladmin — Client for Administering a MySQL Server”.

Sending a SIGHUP signal to the server causes several flush operations to occur that are similar to various forms of the FLUSH statement. Signals can be sent by root or the account that owns the server process. They enable the applicable flush operations to be performed without having to connect to the server (which for these operations requires an account that has the RELOAD privilege). See Section 4.10, “Unix Signal Handling in MySQL”.

The RESET statement is similar to FLUSH. See Section 13.7.6.6, “RESET Statement”, for information about using the RESET statement with replication.

The following list describes the permitted FLUSH statement flush_option values. For descriptions of FLUSH TABLES variants, see FLUSH TABLES Syntax.

• FLUSH BINARY LOGS

  Closes and reopens any binary log file to which the server is writing. If binary logging is enabled, the sequence number of the binary log file is incremented by one relative to the previous file.

• FLUSH DES_KEY_FILE

  Reloads the DES keys from the file that was specified with the --des-key-file option at server startup time.

• FLUSH ENGINE LOGS

  Closes and reopens any flushable logs for installed storage engines. This causes InnoDB to flush its logs to disk.

• FLUSH ERROR LOGS

  Closes and reopens any error log file to which the server is writing.

• FLUSH GENERAL LOGS

  Closes and reopens any general query log file to which the server is writing.

• FLUSH HOSTS

  Empties the host cache and the Performance Schema host_cache table that exposes the cache contents, and unblocks any blocked hosts. See Section 8.12.5.2, “DNS Lookup Optimization and the Host Cache”.

  Flush the host cache if some of your hosts change IP address or if the error message Host 'host_name' is blocked occurs for connections from legitimate hosts. (See Section B.4.2.6, “Host 'host_name' is blocked”.) When more than max_connect_errors errors occur successively for a given host while connecting to the MySQL server, MySQL assumes that something is wrong and blocks the host from further connection requests. Flushing the host cache enables further connection attempts from the host. The default value of max_connect_errors is 100. To avoid this error message, start the server with max_connect_errors set to a large value.

• FLUSH LOGS

  Closes and reopens any log file to which the server is writing. If binary logging is enabled, the sequence number of the binary log file is incremented by one relative to the previous file. If relay logging is enabled, the sequence number of the relay log file is incremented by one relative to the previous file.

  FLUSH LOGS has no effect on tables used for the general query log or for the slow query log (see Section 5.4.1, “Selecting General Query Log and Slow Query Log Output Destinations”). It also has no effect on tables used for the binary or relay log (as controlled by the master_info_repository and relay_log_info_repository system variables).

• FLUSH PRIVILEGES

  Reloads the privileges from the grant tables in the mysql system database.

  If the --skip-grant-tables option was specified at server startup to disable the MySQL privilege system, FLUSH PRIVILEGES provides a way to enable the privilege system at runtime.

  Frees memory cached by the server as a result of GRANT, CREATE USER, CREATE SERVER, and INSTALL PLUGIN statements. This memory is not released by the corresponding REVOKE, DROP USER, DROP SERVER, and UNINSTALL PLUGIN statements, so for a server that executes many instances of the statements that cause caching, there will be an increase in cached memory use unless it is freed with FLUSH PRIVILEGES.

• FLUSH QUERY CACHE

  Defragment the query cache to better utilize its memory. FLUSH QUERY CACHE does not remove any queries from the cache, unlike FLUSH TABLES or RESET QUERY CACHE.

• FLUSH RELAY LOGS

  Closes and reopens any relay log file to which the server is writing. If relay logging is enabled, the sequence number of the relay log file is incremented by one relative to the previous file.

• FLUSH SLOW LOGS

  Closes and reopens any slow query log file to which the server is writing.

• FLUSH STATUS

  This option adds the current thread's session status variable values to the global values and resets the session values to zero. Some global variables may be reset to zero as well. It also resets the counters for key caches (default and named) to zero and sets Max_used_connections to the current number of open connections. This information may be of use when debugging a query. See Section 1.7, “How to Report Bugs or Problems”.

• FLUSH USER_RESOURCES

  Resets all per-hour user resources to zero. This enables clients that have reached their hourly connection, query, or update limits to resume activity immediately. FLUSH USER_RESOURCES does not apply to the limit on maximum simultaneous connections that is controlled by the max_user_connections system variable. See Section 6.2.13, “Setting Account Resource Limits”.

In MySQL 5.6.11 only, gtid_next must be set to AUTOMATIC before issuing this statement. (Bug #16062608, Bug #16715809, Bug #69045)

FLUSH TABLES Syntax

FLUSH TABLES flushes tables, and, depending on the variant used, acquires locks. Any TABLES variant used in a FLUSH statement must be the only option used. FLUSH TABLE is a synonym for FLUSH TABLES.

Note

The descriptions here that indicate tables are flushed by closing them apply differently for InnoDB, which flushes table contents to disk but leaves them open. This still permits table files to be copied while the tables are open, as long as other activity does not modify them.

• FLUSH TABLES

  Closes all open tables, forces all tables in use to be closed, and flushes the query cache and prepared statement cache. FLUSH TABLES also removes all query results from the query cache, like the RESET QUERY CACHE statement. For information about query caching and prepared statement caching, see Section 8.10.3, “The MySQL Query Cache”. and Section 8.10.4, “Caching of Prepared Statements and Stored Programs”.

  FLUSH TABLES is not permitted when there is an active LOCK TABLES ... READ. To flush and lock tables, use FLUSH TABLES tbl_name ... WITH READ LOCK instead.

• FLUSH TABLES tbl_name [, tbl_name] ...

  With a list of one or more comma-separated table names, this statement is like FLUSH TABLES with no names except that the server flushes only the named tables. If a named table does not exist, no error occurs.

• FLUSH TABLES WITH READ LOCK

  Closes all open tables and locks all tables for all databases with a global read lock. This is a very convenient way to get backups if you have a file system such as Veritas or ZFS that can take snapshots in time. Use UNLOCK TABLES to release the lock.

  FLUSH TABLES WITH READ LOCK acquires a global read lock rather than table locks, so it is not subject to the same behavior as LOCK TABLES and UNLOCK TABLES with respect to table locking and implicit commits:

  • UNLOCK TABLES implicitly commits any active transaction only if any tables currently have been locked with LOCK TABLES. The commit does not occur for UNLOCK TABLES following FLUSH TABLES WITH READ LOCK because the latter statement does not acquire table locks.

  • Beginning a transaction causes table locks acquired with LOCK TABLES to be released, as though you had executed UNLOCK TABLES. Beginning a transaction does not release a global read lock acquired with FLUSH TABLES WITH READ LOCK.

  FLUSH TABLES WITH READ LOCK is not compatible with XA transactions.

  FLUSH TABLES WITH READ LOCK does not prevent the server from inserting rows into the log tables (see Section 5.4.1, “Selecting General Query Log and Slow Query Log Output Destinations”).

• FLUSH TABLES tbl_name [, tbl_name] ... WITH READ LOCK

  This statement flushes and acquires read locks for the named tables. The statement first acquires exclusive metadata locks for the tables, so it waits for transactions that have those tables open to complete. Then the statement flushes the tables from the table cache, reopens the tables, acquires table locks (like LOCK TABLES ... READ), and downgrades the metadata locks from exclusive to shared. After the statement acquires locks and downgrades the metadata locks, other sessions can read but not modify the tables.

  Because this statement acquires table locks, you must have the LOCK TABLES privilege for each table, in addition to the RELOAD privilege that is required to use any FLUSH statement.

  This statement applies only to existing base (non-TEMPORARY) tables. If a name refers to a base table, that table is used. If it refers to a TEMPORARY table, it is ignored. If a name applies to a view, an ER_WRONG_OBJECT error occurs. Otherwise, an ER_NO_SUCH_TABLE error occurs.

  Use UNLOCK TABLES to release the locks, LOCK TABLES to release the locks and acquire other locks, or START TRANSACTION to release the locks and begin a new transaction.

  This FLUSH TABLES variant enables tables to be flushed and locked in a single operation. It provides a workaround for the restriction that FLUSH TABLES is not permitted when there is an active LOCK TABLES ... READ.

  This statement does not perform an implicit UNLOCK TABLES, so an error results if you use the statement while there is any active LOCK TABLES or use it a second time without first releasing the locks acquired.

  If a flushed table was opened with HANDLER, the handler is implicitly flushed and loses its position.

• FLUSH TABLES tbl_name [, tbl_name] ... FOR EXPORT

  This FLUSH TABLES variant applies to InnoDB tables. It ensures that changes to the named tables have been flushed to disk so that binary table copies can be made while the server is running.

  The statement works like this:

  1. It acquires shared metadata locks for the named tables. The statement blocks as long as other sessions have active transactions that have modified those tables or hold table locks for them. When the locks have been acquired, the statement blocks transactions that attempt to update the tables, while permitting read-only operations to continue.

  2. It checks whether all storage engines for the tables support FOR EXPORT. If any do not, an ER_ILLEGAL_HA error occurs and the statement fails.

  3. The statement notifies the storage engine for each table to make the table ready for export. The storage engine must ensure that any pending changes are written to disk.

  4. The statement puts the session in lock-tables mode so that the metadata locks acquired earlier are not released when the FOR EXPORT statement completes.

  The FLUSH TABLES ... FOR EXPORT statement requires that you have the SELECT privilege for each table. Because this statement acquires table locks, you must also have the LOCK TABLES privilege for each table, in addition to the RELOAD privilege that is required to use any FLUSH statement.

  This statement applies only to existing base (non-TEMPORARY) tables. If a name refers to a base table, that table is used. If it refers to a TEMPORARY table, it is ignored. If a name applies to a view, an ER_WRONG_OBJECT error occurs. Otherwise, an ER_NO_SUCH_TABLE error occurs.

  InnoDB supports FOR EXPORT for tables that have their own .ibd file file (that is, tables created with the innodb_file_per_table setting enabled). InnoDB ensures when notified by the FOR EXPORT statement that any changes have been flushed to disk. This permits a binary copy of table contents to be made while the FOR EXPORT statement is in effect because the .ibd file is transaction consistent and can be copied while the server is running. FOR EXPORT does not apply to InnoDB system tablespace files, or to InnoDB tables that have FULLTEXT indexes.

  FLUSH TABLES ...FOR EXPORT does not work with partitioned InnoDB tables prior to MySQL 5.6.17, but is supported for such tables in MySQL 5.6.17 and later. (Bug #16943907)

  When notified by FOR EXPORT, InnoDB writes to disk certain kinds of data that is normally held in memory or in separate disk buffers outside the tablespace files. For each table, InnoDB also produces a file named table_name.cfg in the same database directory as the table. The .cfg file contains metadata needed to reimport the tablespace files later, into the same or different server.

  When the FOR EXPORT statement completes, InnoDB will have flushed all dirty pages to the table data files. Any change buffer entries are merged prior to flushing. At this point, the tables are locked and quiescent: The tables are in a transactionally consistent state on disk and you can copy the .ibd tablespace files along with the corresponding .cfg files to get a consistent snapshot of those tables.

  For the procedure to reimport the copied table data into a MySQL instance, see Section 14.6.1.3, “Importing InnoDB Tables”.

  After you are done with the tables, use UNLOCK TABLES to release the locks, LOCK TABLES to release the locks and acquire other locks, or START TRANSACTION to release the locks and begin a new transaction.

  While any of these statements is in effect within the session, attempts to use FLUSH TABLES ... FOR EXPORT produce an error:

  FLUSH TABLES ... WITH READ LOCK
  FLUSH TABLES ... FOR EXPORT
  LOCK TABLES ... READ
  LOCK TABLES ... WRITE

  While FLUSH TABLES ... FOR EXPORT is in effect within the session, attempts to use any of these statements produce an error:

  FLUSH TABLES WITH READ LOCK
  FLUSH TABLES ... WITH READ LOCK
  FLUSH TABLES ... FOR EXPORT