mysql 5.0.41 mysqld 引數大全

./mysqld  Ver 5.0.41-log for pc-linux-gnu on i686 (MySQL Community Server (GPL))
Copyright (C) 2000 MySQL AB, by Monty and others
This software comes with ABSOLUTELY NO WARRANTY. This is free software,
and you are welcome to modify and redistribute it under the GPL license

Starts the MySQL database server

Usage: ./mysqld [OPTIONS]

Default options are read from the following files in the given order:
/etc/my.cnf ~/.my.cnf /usr/local/mysql/etc/my.cnf
The following groups are read: mysql_cluster mysqld server mysqld-5.0
The following options may be given as the first argument:
--print-defaults    Print the program argument list and exit
--no-defaults        Don't read default options from any options file
--defaults-file=#    Only read default options from the given file #
--defaults-extra-file=# Read this file after the global files are read

  -?, --help          Display this help and exit.
  --abort-slave-event-count=#
                      Option used by mysql-test for debugging and testing of
                      replication.
  --allow-suspicious-udfs
                      Allows use of UDFs consisting of only one symbol xxx()
                      without corresponding xxx_init() or xxx_deinit(). That
                      also means that one can load any function from any
                      library, for example exit() from libc.so
  -a, --ansi          Use ANSI SQL syntax instead of MySQL syntax. This mode
                      will also set transaction isolation level 'serializable'.
  --auto-increment-increment[=#]
                      Auto-increment columns are incremented by this
  --auto-increment-offset[=#]
                      Offset added to Auto-increment columns. Used when
                      auto-increment-increment != 1
  --automatic-sp-privileges
                      Creating and dropping stored procedures alters ACLs.
                      Disable with --skip-automatic-sp-privileges.
  -b, --basedir=name  Path to installation directory. All paths are usually
                      resolved relative to this.
  --bdb               Enable Berkeley DB (if this version of MySQL supports
                      it). Disable with --skip-bdb (will save memory).
  --big-tables        Allow big result sets by saving all temporary sets on
                      file (Solves most 'table full' errors).
  --bind-address=name IP address to bind to.
  --binlog-do-db=name Tells the master it should log updates for the specified
                      database, and exclude all others not explicitly
                      mentioned.
  --binlog-ignore-db=name
                      Tells the master that updates to the given database
                      should not be logged tothe binary log.
  --bootstrap         Used by mysql installation scripts.
  --character-set-client-handshake
                      Don't ignore client side character set value sent during
                      handshake.
  --character-set-filesystem=name
                      Set the filesystem character set.
  -C, --character-set-server=name
                      Set the default character set.
  --character-sets-dir=name
                      Directory where character sets are.
  -r, --chroot=name   Chroot mysqld daemon during startup.
  --collation-server=name
                      Set the default collation.
  --completion-type=# Default completion type.
  --concurrent-insert[=#]
                      Use concurrent insert with MyISAM. Disable with
                      --concurrent-insert=0
  --console           Write error output on screen; Don't remove the console
                      window on windows.
  --core-file         Write core on errors.
  -h, --datadir=name  Path to the database root.
  -C, --default-character-set=name
                      Set the default character set (deprecated option, use
                      --character-set-server instead).
  --default-collation=name
                      Set the default collation (deprecated option, use
                      --collation-server instead).
  --default-storage-engine=name
                      Set the default storage engine (table type) for tables.
  --default-table-type=name
                      (deprecated) Use --default-storage-engine.
  --default-time-zone=name
                      Set the default time zone.
  --delay-key-write[=name]
                      Type of DELAY_KEY_WRITE.
  --delay-key-write-for-all-tables
                      Don't flush key buffers between writes for any MyISAM
                      table (Deprecated option, use --delay-key-write=all
                      instead).
  --des-key-file=name Load keys for des_encrypt() and des_encrypt from given
                      file.
  --disconnect-slave-event-count=#
                      Option used by mysql-test for debugging and testing of
                      replication.
  --enable-locking    Deprecated option, use --external-locking instead.
  --enable-pstack     Print a symbolic stack trace on failure.
  --engine-condition-pushdown
                      Push supported query conditions to the storage engine.
  -T, --exit-info[=#] Used for debugging;  Use at your own risk!
  --external-locking  Use system (external) locking (disabled by default). With
                      this option enabled you can run myisamchk to test (not
                      repair) tables while the MySQL server is running. Disable
                      with --skip-external-locking.
  --flush             Flush tables to disk between SQL commands.
  --gdb               Set up signals usable for debugging
  --large-pages       Enable support for large pages. Disable with
                      --skip-large-pages.
  --init-connect=name Command(s) that are executed for each new connection
  --init-file=name    Read SQL commands from this file at startup.
  --init-rpl-role=name
                      Set the replication role.
  --init-slave=name   Command(s) that are executed when a slave connects to
                      this master
  --innodb            Enable InnoDB (if this version of MySQL supports it).
                      Disable with --skip-innodb (will save memory).
  --innodb_checksums  Enable InnoDB checksums validation (enabled by default).
                      Disable with --skip-innodb-checksums.
  --innodb_data_file_path=name
                      Path to individual files and their sizes.
  --innodb_data_home_dir=name
                      The common part for InnoDB table spaces.
  --innodb_doublewrite
                      Enable InnoDB doublewrite buffer (enabled by default).
                      Disable with --skip-innodb-doublewrite.
  --innodb_fast_shutdown[=#]
                      Speeds up the shutdown process of the InnoDB storage
                      engine. Possible values are 0, 1 (faster) or 2 (fastest -
                      crash-like).
  --innodb_file_per_table
                      Stores each InnoDB table to an .ibd file in the database
                      dir.
  --innodb_flush_log_at_trx_commit[=#]
                      Set to 0 (write and flush once per second), 1 (write and
                      flush at each commit) or 2 (write at commit, flush once
                      per second).
  --innodb_flush_method=name
                      With which method to flush data.
  --innodb_locks_unsafe_for_binlog
                      Force InnoDB not to use next-key locking. Instead use
                      only row-level locking
  --innodb_log_arch_dir=name
                      Where full logs should be archived.
  --innodb_log_archive[=#]
                      Set to 1 if you want to have logs archived.
  --innodb_log_group_home_dir=name
                      Path to InnoDB log files.
  --innodb_max_dirty_pages_pct=#
                      Percentage of dirty pages allowed in bufferpool.
  --innodb_max_purge_lag=#
                      Desired maximum length of the purge queue (0 = no limit)
  --innodb_rollback_on_timeout
                      Roll back the complete transaction on lock wait timeout,
                      for 4.x compatibility (disabled by default)
  --innodb_status_file
                      Enable SHOW INNODB STATUS output in the
                      innodb_status. file
  --innodb_support_xa Enable InnoDB support for the XA two-phase commit
  --innodb_table_locks
                      Enable InnoDB locking in LOCK TABLES
  --isam              Obsolete. ISAM storage engine is no longer supported.
  -L, --language=name Client error messages in given language. May be given as
                      a full path.
  --local-infile      Enable/disable LOAD DATA LOCAL INFILE (takes values 1|0).
  -l, --log[=name]    Log connections and queries to file.
  --log-bin[=name]    Log update queries in binary format. Optional (but
                      strongly recommended to avoid replication problems if
                      server's hostname changes) argument should be the chosen
                      location for the binary log files.
  --log-bin-index=name
                      File that holds the names for last binary log files.
  --log-bin-trust-function-creators
                      If equal to 0 (the default), then when --log-bin is used,
                      creation of a stored function is allowed only to users
                      having the SUPER privilege and only if this function may
                      not break binary logging.
  --log-bin-trust-routine-creators
                      (deprecated) Use log-bin-trust-function-creators.
  --log-error[=name]  Error log file.
  --log-isam[=name]   Log all MyISAM changes to file.
  -0, --log-long-format
                      Log some extra information to update log. Please note
                      that this option is deprecated; see --log-short-format
                      option.
  --log-queries-not-using-indexes
                      Log queries that are executed without benefit of any
                      index to the slow log if it is open.
  --log-short-format  Don't log extra information to update and slow-query
                      logs.
  --log-slave-updates Tells the slave to log the updates from the slave thread
                      to the binary log. You will need to turn it on if you
                      plan to daisy-chain the slaves.
  --log-slow-admin-statements
                      Log slow OPTIMIZE, ANALYZE, ALTER and other
                      administrative statements to the slow log if it is open.
  --log-slow-queries[=name]
                      Log slow queries to this log file. Defaults logging to
                      hostname-slow.log file. Must be enabled to activate other
                      slow log options.
  --log-tc=name       Path to transaction coordinator log (used for
                      transactions that affect more than one storage engine,
                      when binary log is disabled)
  --log-tc-size=#     Size of transaction coordinator log.
  --log-update[=name] The update log is deprecated since version 5.0, is
                      replaced by the binary log and this option justs turns on
                      --log-bin instead.
  -W, --log-warnings[=#]
                      Log some not critical warnings to the log file.
  --low-priority-updates
                      INSERT/DELETE/UPDATE has lower priority than selects.
  --master-connect-retry=#
                      The number of seconds the slave thread will sleep before
                      retrying to connect to the master in case the master goes
                      down or the connection is lost.
  --master-host=name  Master hostname or IP address for replication. If not
                      set, the slave thread will not be started. Note that the
                      setting of master-host will be ignored if there exists a
                      valid master.info file.
  --master-info-file=name
                      The location and name of the file that remembers the
                      master and where the I/O replication thread is in the
                      master's binlogs.
  --master-password=name
                      The password the slave thread will authenticate with when
                      connecting to the master. If not set, an empty password
                      is assumed.The value in master.info will take precedence
                      if it can be read.
  --master-port=#     The port the master is listening on. If not set, the
                      compiled setting of MYSQL_PORT is assumed. If you have
                      not tinkered with configure options, this should be 3306.
                      The value in master.info will take precedence if it can
                      be read.
  --master-retry-count=#
                      The number of tries the slave will make to connect to the
                      master before giving up.
  --master-ssl        Enable the slave to connect to the master using SSL.
  --master-ssl-ca[=name]
                      Master SSL CA file. Only applies if you have enabled
                      master-ssl.
  --master-ssl-capath[=name]
                      Master SSL CA path. Only applies if you have enabled
                      master-ssl.
  --master-ssl-cert[=name]
                      Master SSL certificate file name. Only applies if you
                      have enabled master-ssl
  --master-ssl-cipher[=name]
                      Master SSL cipher. Only applies if you have enabled
                      master-ssl.
  --master-ssl-key[=name]
                      Master SSL keyfile name. Only applies if you have enabled
                      master-ssl.
  --master-user=name  The username the slave thread will use for authentication
                      when connecting to the master. The user must have FILE
                      privilege. If the master user is not set, user test is
                      assumed. The value in master.info will take precedence if
                      it can be read.
  --max-binlog-dump-events=#
                      Option used by mysql-test for debugging and testing of
                      replication.
  --memlock           Lock mysqld in memory.
  --merge             Enable Merge storage engine. Disable with --skip-merge.
  --myisam-recover[=name]
                      Syntax: myisam-recover[=option[,option...]], where option
                      can be DEFAULT, BACKUP, FORCE or QUICK.
  --ndbcluster        Enable NDB Cluster (if this version of MySQL supports
                      it). Disable with --skip-ndbcluster (will save memory).
  --ndb-connectstring=name
                      Connect string for ndbcluster.
  --ndb-mgmd-host=name
                      Set host and port for ndb_mgmd. Syntax: hostname[:port]
  --ndb-nodeid=#      Nodeid for this mysqlserver in the cluster.
  --ndb-autoincrement-prefetch-sz=#
                      Specify number of autoincrement values that are
                      prefetched.
  --ndb-force-send    Force send of buffers to ndb immediately without waiting
                      for other threads.
  --ndb_force_send    same as --ndb-force-send.
  --ndb-use-exact-count
                      Use exact records count during query planning and for
                      fast select count(*), disable for faster queries.
  --ndb_use_exact_count
                      same as --ndb-use-exact-count.
  --ndb-use-transactions
                      Use transactions for large inserts, if enabled then large
                      inserts will be split into several smaller transactions
  --ndb_use_transactions
                      same as --ndb-use-transactions.
  --ndb-shm           Use shared memory connections when available.
  --ndb-optimized-node-selection
                      Select nodes for transactions in a more optimal way.
  --ndb-cache-check-time=#
                      A dedicated thread is created to, at the given
                      millisecons interval, invalidate the query cache if
                      another MySQL server in the cluster has changed the data
                      in the database.
  -n, --new           Use very new possible 'unsafe' functions.
  --old-passwords     Use old password encryption method (needed for 4.0 and
                      older clients).
  --old-style-user-limits
                      Enable old-style. user limits (before 5.0.3 user resources
                      were counted per each user+host vs. per account)
  --pid-file=name     Pid file used by safe_mysqld.
  -P, --port=#        Port number to use for connection.
  --port-open-timeout=#
                      Maximum time in seconds to wait for the port to become
                      free. (Default: no wait)
  --profiling_history_size=#
                      Limit of query profiling memory
  --relay-log=name    The location and name to use for relay logs.
  --relay-log-index=name
                      The location and name to use for the file that keeps a
                      list of the last relay logs.
  --relay-log-info-file=name
                      The location and name of the file that remembers where
                      the SQL replication thread is in the relay logs.
  --replicate-do-db=name
                      Tells the slave thread to restrict replication to the
                      specified database. To specify more than one database,
                      use the directive multiple times, once for each database.
                      Note that this will only work if you do not use
                      cross-database queries such as UPDATE some_db.some_table
                      SET foo='bar' while having selected a different or no
                      database. If you need cross database updates to work,
                      make sure you have 3.23.28 or later, and use
                      replicate-wild-do-table=db_name.%.
  --replicate-do-table=name
                      Tells the slave thread to restrict replication to the
                      specified table. To specify more than one table, use the
                      directive multiple times, once for each table. This will
                      work for cross-database updates, in contrast to
                      replicate-do-db.
  --replicate-ignore-db=name
                      Tells the slave thread to not replicate to the specified
                      database. To specify more than one database to ignore,
                      use the directive multiple times, once for each database.
                      This option will not work if you use cross database
                      updates. If you need cross database updates to work, make
                      sure you have 3.23.28 or later, and use
                      replicate-wild-ignore-table=db_name.%.
  --replicate-ignore-table=name
                      Tells the slave thread to not replicate to the specified
                      table. To specify more than one table to ignore, use the
                      directive multiple times, once for each table. This will
                      work for cross-datbase updates, in contrast to
                      replicate-ignore-db.
  --replicate-rewrite-db=name
                      Updates to a database with a different name than the
                      original. Example:
                      replicate-rewrite-db=master_db_name->slave_db_name.
  --replicate-same-server-id
                      In replication, if set to 1, do not skip events having
                      our server id. Default value is 0 (to break infinite
                      loops in circular replication). Can't be set to 1 if
                      --log-slave-updates is used.
  --replicate-wild-do-table=name
                      Tells the slave thread to restrict replication to the
                      tables that match the specified wildcard pattern. To
                      specify more than one table, use the directive multiple
                      times, once for each table. This will work for
                      cross-database updates. Example:
                      replicate-wild-do-table=foo%.bar% will replicate only
                      updates to tables in all databases that start with foo
                      and whose table names start with bar.
  --replicate-wild-ignore-table=name
                      Tells the slave thread to not replicate to the tables
                      that match the given wildcard pattern. To specify more
                      than one table to ignore, use the directive multiple
                      times, once for each table. This will work for
                      cross-database updates. Example:
                      replicate-wild-ignore-table=foo%.bar% will not do updates
                      to tables in databases that start with foo and whose
                      table names start with bar.
  --report-host=name  Hostname or IP of the slave to be reported to to the
                      master during slave registration. Will appear in the
                      output of SHOW SLAVE HOSTS. Leave unset if you do not
                      want the slave to register itself with the master. Note
                      that it is not sufficient for the master to simply read
                      the IP of the slave off the socket once the slave
                      connects. Due to NAT and other routing issues, that IP
                      may not be valid for connecting to the slave from the
                      master or other hosts.
  --report-password=name
                      Undocumented.
  --report-port=#     Port for connecting to slave reported to the master
                      during slave registration. Set it only if the slave is
                      listening on a non-default port or if you have a special
                      tunnel from the master or other clients to the slave. If
                      not sure, leave this option unset.
  --report-user=name  Undocumented.
  --rpl-recovery-rank=#
                      Undocumented.
  --safe-mode         Skip some optimize stages (for testing).
  --safe-show-database
                      Deprecated option; use GRANT SHOW DATABASES instead...
  --safe-user-create  Don't allow new user creation by the user who has no
                      write privileges to the mysql.user table.
  --safemalloc-mem-limit=#
                      Simulate memory shortage when compiled with the
                      --with-debug=full option.
  --secure-auth       Disallow authentication for accounts that have old
                      (pre-4.1) passwords.
  --secure-file-priv=name
                      Limit LOAD DATA, SELECT ... OUTFILE, and LOAD_FILE() to
                      files within specified directory
  --server-id=#       Uniquely identifies the server instance in the community
                      of replication partners.
  -O, --set-variable=name
                      Change the value of a variable. Please note that this
                      option is deprecated;you can set variables directly with
                      --variable-name=value.
  --show-slave-auth-info
                      Show user and password in SHOW SLAVE HOSTS on this master
  --skip-grant-tables Start without grant tables. This gives all users FULL
                      ACCESS to all tables!
  --skip-host-cache   Don't cache host names.
  --skip-locking      Deprecated option, use --skip-external-locking instead.
  --skip-name-resolve Don't resolve hostnames. All hostnames are IP's or
                      'localhost'.
  --skip-networking   Don't allow connection with TCP/IP.
  --skip-new          Don't use new, possible wrong routines.
  --skip-show-database
                      Don't allow 'SHOW DATABASE' commands.
  --skip-slave-start  If set, slave is not autostarted.
  --skip-stack-trace  Don't print a stack trace on failure.
  --skip-symlink      Don't allow symlinking of tables. Deprecated option.  Use
                      --skip-symbolic-links instead.
  --skip-thread-priority
                      Don't give threads different priorities.
  --slave-load-tmpdir=name
                      The location where the slave should put its temporary
                      files when replicating a LOAD DATA INFILE command.
  --slave-skip-errors=name
                      Tells the slave thread to continue replication when a
                      query returns an error from the provided list.
  --socket=name       Socket file to use for connection.
  --sporadic-binlog-dump-fail
                      Option used by mysql-test for debugging and testing of
                      replication.
  --sql-bin-update-same=#
                      The update log is deprecated since version 5.0, is
                      replaced by the binary log and this option does nothing
                      anymore.
  --sql-mode=name     Syntax: sql-mode=option[,option[,option...]] where option
                      can be one of: REAL_AS_FLOAT, PIPES_AS_CONCAT,
                      ANSI_QUOTES, IGNORE_SPACE, ONLY_FULL_GROUP_BY,
                      NO_UNSIGNED_SUBTRACTION.
  --ssl               Enable SSL for connection (automatically enabled with
                      other flags). Disable with --skip-ssl.
  --ssl-ca=name       CA file in PEM format (check OpenSSL docs, implies
                      --ssl).
  --ssl-capath=name   CA directory (check OpenSSL docs, implies --ssl).
  --ssl-cert=name     X509 cert in PEM format (implies --ssl).
  --ssl-cipher=name   SSL cipher to use (implies --ssl).
  --ssl-key=name      X509 key in PEM format (implies --ssl).
  -s, --symbolic-links
                      Enable symbolic link support.
  --sysdate-is-now    Non-default option to alias SYSDATE() to NOW() to make it
                      safe-replicable. Since 5.0, SYSDATE() returns a `dynamic'
                      value different for different invocations, even within
                      the same statement.
  --tc-heuristic-recover=name
                      Decision to use in heuristic recover process. Possible
                      values are COMMIT or ROLLBACK.
  --temp-pool         Using this option will cause most temporary files created
                      to use a small set of names, rather than a unique name
                      for each new file.
  --timed_mutexes     Specify whether to time mutexes (only InnoDB mutexes are
                      currently supported)
  -t, --tmpdir=name   Path for temporary files. Several paths may be specified,
                      separated by a colon (:), in this case they are used in a
                      round-robin fashion.
  --transaction-isolation=name
                      Default transaction isolation level.
  -s, --use-symbolic-links
                      Enable symbolic link support. Deprecated option; use
                      --symbolic-links instead.
  -u, --user=name     Run mysqld daemon as user.
  -v, --verbose       Used with --help option for detailed help
  -V, --version       Output version information and exit.
  -W, --warnings[=#]  Deprecated; use --log-warnings instead.
  --back_log=#        The number of outstanding connection requests MySQL can
                      have. This comes into play when the main MySQL thread
                      gets very many connection requests in a very short time.
  --binlog_cache_size=#
                      The size of the cache to hold the SQL statements for the
                      binary log during a transaction. If you often use big,
                      multi-statement transactions you can increase this to get
                      more performance.
  --bulk_insert_buffer_size=#
                      Size of tree cache used in bulk insert optimisation. Note
                      that this is a limit per thread!
  --connect_timeout=# The number of seconds the mysqld server is waiting for a
                      connect packet before responding with 'Bad handshake'.
  --date_format=name  The DATE format (For future).
  --datetime_format=name
                      The DATETIME/TIMESTAMP format (for future).
  --default_week_format=#
                      The default week format used by WEEK() functions.
  --delayed_insert_limit=#
                      After inserting delayed_insert_limit rows, the INSERT
                      DELAYED handler will check if there are any SELECT
                      statements pending. If so, it allows these to execute
                      before continuing.
  --delayed_insert_timeout=#
                      How long a INSERT DELAYED thread should wait for INSERT
                      statements before terminating.
  --delayed_queue_size=#
                      What size queue (in rows) should be allocated for
                      handling INSERT DELAYED. If the queue becomes full, any
                      client that does INSERT DELAYED will wait until there is
                      room in the queue again.
  --div_precision_increment=#
                      Precision of the result of '/' operator will be increased
                      on that value.
  --expire_logs_days=#
                   &nb