■ The variable list is displayed. ‹ Click the Exit button to return to the administration panel. › Click the Display status button to display detailed server status. ■ The status information is displayed in a new window. ˇ Click Exit to return to the administration panel. MANAGE THE MYSQL SERVER 9 You may have noticed that there is often more than one way to achieve the same result. For example, if you want to display a list of MySQL variables and their current values, you can use several commands: show variables from the MySQL monitor, mysqladmin variables from the command prompt, the Variables tab in WinMySQLadmin, or the Show variables button in MySQLGUI. Each of these methods of MySQL administration has its advantages and disadvantages. The MySQL monitor and the mysqladmin utility are available on all operating systems and provide a consistent interface, while the MySQLGUI and WinMySQLadmin utilities provide a graphical interface with easy access to common options. You can use whichever commands are available to you, and whichever you find the most comfortable to work with. In addition to the administrative features discussed here, you can use MySQLGUI to send queries to the MySQL server and display the results. Unlike the command-line MySQL monitor, MySQLGUI includes options to save query results to a file after you view them, and saves a list of the most recent queries for easy access. While MySQLGUI looks simple, keep in mind that it is every bit as powerful as the command-line utilities. You can use it to delete an entire database or shut down the MySQL server. Be sure to select commands carefully. 187 516922 Ch09.F 9/26/02 12:01 PM Page 187 W hile MySQL will work remarkably well using the default settings, you can change a variety of factors to improve performance. Some of the most important factors that affect MySQL server performance include the speed of the server hardware, the design of tables, and the performance of particular queries. OPTIMIZE MYSQL MySQL Disk Access Disk access is usually the largest bottleneck affecting a MySQL server. Because tables are stored on disk, virtually every MySQL operation involves disk access. MySQL will benefit from a fast disk drive. Additionally, you may want to consider using a separate drive for MySQL data so that other server functions do not slow down MySQL. Processor Speed While not as important as disk speed, the processor, or CPU, of the MySQL server comes into play when working with data that has been read from disk: A faster processor will handle MySQL queries faster. It is also beneficial to analyze the other applications on the server. If a Web server or other software is using the CPU intensively, MySQL would benefit from a faster processor or a dedicated MySQL server machine. Memory A busy MySQL server requires a large amount of memory. This is used to store data temporarily while it is sent to clients, and for temporary tables. Adding memory will often improve performance, especially if the disk drives and CPU are already reasonably fast. OPTIMIZE THE MYSQL SERVER Using Fixed-Length Rows When a table is frequently changed, MySQL performs better using fixed-length rows. To use fixed-length rows on a table, avoid using variable-length column types such as VARCHAR, TEXT, and BLOB. The disadvantage of fixed-length rows is that all rows take the same amount of space. Variable-length rows will make more efficient use of disk space if there is a large amount of variation in the sizes of data items. Reduce Data Size The less data MySQL has to work with, the faster it will be. Use the smallest column sizes possible for your data and eliminate unnecessary columns to improve the MySQL server's performance working with a table. OPTIMIZE TABLE DESIGN 188 One aspect of MySQL that is relatively easy to optimize is the speed of the machine running the MySQL server. Upgrading disks, memory, or processor, or switching to a faster machine can dramatically improve performance. The design of a table can also affect performance. By considering performance when you design and create a table under MySQL, you can ensure that queries on the table can be quickly handled by the server. 516922 Ch10.F 9/26/02 11:38 AM Page 188 OPTIMIZE AND TROUBLESHOOT MYSQL 10 189 Using EXPLAIN You can use the EXPLAIN command with a SELECT query to display information about how MySQL will handle the query. When you use EXPLAIN, the MySQL server does not actually process the SELECT query. Instead, it displays a chart of information about the query. The information displayed by EXPLAIN includes the tables the query will use, the keys it can use to quickly find records, the number of rows the query will return, and any extra explanation the server can offer about the query. The Extra column will indicate whether a file sort is necessary to process an ORDER BY clause. It will also indicate whether a temporary table will be needed to handle a GROUP BY clause. If either of these is present, it indicates a major bottleneck for the query. You may be able to eliminate the problem by adding an index or changing the query. Example: EXPLAIN SELECT * FROM quotes; Improve Queries After you have isolated a slow query, you can try changing its syntax. Eliminate any unnecessary ORDER BY clauses, as they can slow down the query. Add WHERE clauses to target specific records wherever possible, and use the LIMIT clause to limit the number of records returned. See Chapter 6 for details about the syntax of these SELECT query clauses. Add Indexes If you frequently search for values in a particular column, you may be able to improve performance by adding an index on that column. You can use the ALTER TABLE command to add an index at any time. This is explained further in "Improve Performance with Indexes," later in this chapter. Lock Tables MySQL normally handles table locking automatically. You may be able to improve the performance of a complex query or series of queries by locking the table first. You can use the LOCK TABLES and UNLOCK TABLES commands to control table locking. This is described in "Manage Table Locking," later in this chapter. OPTIMIZE SLOW QUERIES Multiple Tables and Relationships Performance problems multiply when you are working with multiple tables. Because JOIN queries that retrieve data from multiple tables tend to be slow, do not divide data into more tables than necessary. When you do use multiple tables that have a relationship, be sure the columns that form the relationship are the same type, length, and preferably have the same name. This will ensure that the MySQL server can perform a JOIN query efficiently. Sort Table Data You can use the ORDER BY keywords with an ALTER TABLE command to sort the data in the table. This can improve performance when the contents of the table are often read in the same order, and are not changed frequently. Example: ALTER TABLE address ORDER BY name; OPTIMIZE TABLE DESIGN (CONTINUED) Often, a particular query that is used frequently can slow down the MySQL server. By optimizing a query's syntax, you can often improve its performance. The MySQL server keeps track of the number of slow queries, or queries that took more than a certain length of time, since the server started. You can display this value using the SHOW STATUS command in the MySQL monitor or the mysqladmin status command at the command prompt. The MySQL server can optionally maintain a slow query log, which keeps a record of each slow query. This is particularly useful for determining which queries are slowing down the server. This log file is described in detail in "View MySQL Log Files," later in this chapter. 516922 Ch10.F 9/26/02 11:38 AM Page 189 ⁄ From the command prompt, type cd /usr/local/ mysql/data/testdb and press Enter. ■ This switches to the directory where the testdb database is stored. Note: On Windows systems, the directory is usually c:\mysql\data\testdb. ¤ Type myisamchk quotes and press Enter. ■ This checks the table for errors and displays a report. Note: To check a different table, replace quotes with the name of the table. M ySQL includes a utility called myisamchk that allows you to check tables for errors and repair any errors that occur. You can use this utility as a regular check to watch for errors, or when you suspect a problem with a table. This utility is for the default MyISAM table type. An older utility, isamchk, works for ISAM tables. To check a table for errors, start in the directory where the database files are stored, typically /usr/local/mysql/data/ database_name. Type myisamchk followed by one or more table names. The check will be performed immediately, and may take several minutes on a large table. No clients can access the table while myisamchk is running, and myisamchk only works if no clients have a lock on the table. The following example checks the quotes table: cd /usr/local/mysql/data/testdb myisamchk quotes The myisamchk utility displays a series of status messages during the table check. If any of these indicate an error in the table, you can attempt to repair the table. Use the myisamchk -r option to attempt to repair a corrupt table. The following example repairs the quotes table: myisamchk -r quotes If an error message is displayed, and myisamchk is unable to repair the table using this option, you can try the -o option. This performs a slower recovery process that may work when the standard process fails. Unlike other MySQL utilities, the myisamchk utility does not require a MySQL username and password. It works directly with database files. For this reason, in the default installation, your UNIX username must have root access to run myisamchk. On Windows systems, this utility is available as c:\mysql\bin\myisamchk.exe. CHECK A TABLE FOR ERRORS MySQL 190 CHECK A TABLE FOR ERRORS 516922 Ch10.F 9/26/02 11:38 AM Page 190 ‹ Type myisamchk -m quotes and press Enter. ■ This performs a more detailed (medium) check on the table. › Type myisamchk -r quotes and press Enter. ■ This attempts to recover the table data. OPTIMIZE AND TROUBLESHOOT MYSQL 10 The myisamchk utility includes a number of additional options to control the check and repair process. Type myisamchk help for a complete list of options. The following table describes some of the most useful options: OPTION DESCRIPTION -c Check (default if no options are specified) -e Extended check — slow but more thorough -m Medium check — faster than extended -F Fast check — only checks improperly closed tables -C Checks only tables changed since the last check -i Displays information about the table while checking -f Automatically repairs the table if any errors are detected -T Does not mark table as checked -r Recover — attempts to repair table and recover data -o Safe recover — uses slower recovery method -q Quick recover — checks index files only -v Verbose — displays detailed information -V Displays the myisamchk version number -w Wait — waits until no clients are locking table before checking 191 516922 Ch10.F 9/26/02 11:38 AM Page 191 Note: This example uses the quotes table in the testdb database, but it would work with any table. ⁄ From the MySQL monitor, type USE testdb; and press Enter. ■ The database is now selected. ¤ Type OPTIMIZE TABLE quotes; and press Enter. ■ The table is scanned and optimized, and a report is displayed. W hen you delete rows from a MySQL table, they are not actually deleted. Instead, MySQL marks the rows as deleted and re-uses the space later when rows are inserted. If you have deleted a large number of rows from a table, you should optimize the table to reclaim the space. Optimizing is also necessary when a table with variable-length rows has been changed many times. To optimize a table, use the OPTIMIZE TABLE command within the MySQL monitor. To use this command, specify the table name. The following example optimizes the quotes table: OPTIMIZE TABLE quotes; Along with reclaiming space from deleted rows, the OPTIMIZE TABLE command also repairs minor errors in table rows, sorts the index files, and updates the table's statistics. You can use this command as often as you desire without damaging a table. However, the table is locked and cannot be used by clients during the optimization process. From time to time, you may run into a situation where a MySQL table becomes corrupted. This usually happens when a power outage or hardware failure causes the server to go down unexpectedly while a table is being updated. In most cases, you can easily repair the table. You can use the myisamchk utility discussed earlier or the REPAIR TABLE command to repair a damaged table. To use REPAIR TABLE, specify the table name. You can also specify the optional keyword QUICK for a quick repair or EXTENDED for an extended repair. If the regular repair does not work, the extended option may. The following command repairs the quotes table: REPAIR TABLE quotes; OPTIMIZE AND REPAIR TABLES MySQL 192 OPTIMIZE AND REPAIR TABLES 516922 Ch10.F 9/26/02 11:38 AM Page 192 ‹ Type REPAIR TABLE quotes; and press Enter. ■ This attempts to repair any errors in the table. › Type SELECT * FROM quotes; and press Enter. ■ This displays the contents of the table. Verify that the table rows are undamaged. OPTIMIZE AND TROUBLESHOOT MYSQL 10 Along with OPTIMIZE TABLE and REPAIR TABLE, MySQL includes a CHECK TABLE command. This command is equivalent to the myisamchk utility, but runs from the MySQL monitor or another client. To use CHECK TABLE, specify one or more table names. The following example checks the quotes table for errors: Example: CHECK TABLE quotes; You can also use several optional keywords with CHECK TABLE after the table name. These are described in the table below. KEYWORD DESCRIPTION QUICK Quick check — does not scan all table rows FAST Fast check — only checks improperly closed tables CHANGED Checks only tables that have changed since the last check MEDIUM Medium check — checks each table row (default) EXTENDED Extended check — comprehensive but slow on large tables 193 516922 Ch10.F 9/26/02 11:38 AM Page 193 T he MySQL server is a complex system, and includes a number of parameters you can use to tune the server's performance. You can modify these values by editing configuration files or with command-line options. Each configuration file is divided into sections for different MySQL components. For example, the line [mysqld] begins the section for the server, and [client] begins the section for clients. Within each section, each line can include an option from the program's command-line options or the set-variable command to set a system variable. MYSQL CONFIGURATION OPTIONS MySQL 194 The Global Configuration File The global configuration file is read first. On UNIX systems, this file is /etc/my.cnf. On Windows systems, the two global files are supported: my.ini in the Windows system directory, and my.cnf in the root directory, typically C:\. The User Configuration File On UNIX systems, each user can have their own configuration file, .my.cnf, in their home directory. You can use this file to set values for each user, typically for MySQL client utilities. The Server Configuration File The server configuration file affects a particular copy of the MySQL server software, and is only needed when multiple servers are installed on the same machine. This file also has the filename my.cnf. On UNIX systems, it is stored in the data directory under the MySQL installation, typically /usr/local/mysql/data. On Windows systems, the file is usually under C:\mysql\data\. The values you specify in the server configuration file override the values in the global file. CONFIGURATION FILES The [client] section in the configuration file includes options that affect all of the MySQL client programs, including the MySQL monitor, mysql, the mysqladmin utility, myisamchk, and other client tools. This section is particularly useful in a .my.cnf file in a user's home directory. For example, the following file excerpt sets the password option for MySQL clients. If you include this in the .my.cnf file in your home directory, you do not need to specify a password when using MySQL client programs. [client] password=mypassword The table below lists several options that may also be useful in the [client] section for a particular user. VALUE DESCRIPTION host=name Specifies a MySQL server to connect to user=name Username for the MySQL server, if different from UNIX username password=value Password for the MySQL server database=value Default database to select CLIENT OPTIONS MySQL supports a variety of configuration files. Each can contain the same commands and settings, but the order in which they are read determines which files can override others. 516922 Ch10.F 9/26/02 11:38 AM Page 194 OPTIMIZE AND TROUBLESHOOT MYSQL 10 VARIABLE DESCRIPTION back_log Maximum number of client requests waiting for threads concurrent_inserts Specifies ON to allow INSERT operations while clients are reading data connect_timeout Number of seconds the server waits for a connection before timing out delayed_insert_limit Number of rows of an INSERT DELAYED query to process at a time delayed_insert_timeout Time the server waits for additional INSERT DELAYED items delayed_queue_size Number of INSERT DELAYED rows to store flush_time Closes all tables after the specified number of seconds interactive_timeout Number of seconds of idle time before disconnecting interactive clients join_buffer_size Buffer size for full joins (queries from multiple tables) key_buffer_size Buffer size for index values long_query_time Amount of seconds before a query is considered slow max_allowed_packet Maximum size of a single packet of data max_connections Maximum number of simultaneous client connections allowed max_connect_errors Number of allowed errors before blocking connections from a host max_delayed_threads Maximum number of threads used for INSERT DELAYED queries max_join_size Maximum number of rows for JOIN queries max_sort_length The number of bytes of each BLOB or TEXT value to use when sorting max_user_connections Maximum number of connections for each username net_buffer_length Default size for the communication buffer net_read_timeout Number of seconds to wait before aborting when reading data net_write_timeout Number of seconds to wait before aborting when writing data thread_cache_size Number of threads kept standing by for use by clients wait_timeout Number of seconds of idle time before disconnecting a client SYSTEM VARIABLES Set Variables You can set MySQL system variables by including them in the [mysqld] section of a configuration file, using the set- variable command. The following example sets the max_allowed_packet variable: [mysqld] set-variable max_allowed_packet=1M You can also set variables using the set-variable option when mysqld is started. Display Current Values You can display the current values of all of the system variables using the SHOW VARIABLES command from the MySQL monitor or the mysqladmin variables command at the UNIX or Windows command prompt. This is useful if you are unsure whether the configuration files are being read correctly, and is the first thing you should do if a variable change has not produced the effect you expected. You can also use the LIKE operator with SHOW VARIABLES to show a section of the list. The following example shows the values of all variables that include the characters "max" in their names: SHOW VARIABLES LIKE '%max%'; 195 MySQL includes a number of system variables that control the behavior of the server. While these variables have sensible default values, setting them to optimal values for your system can dramatically improve the server's performance. The table below describes key system variables for optimizing performance. 516922 Ch10.F 9/26/02 11:38 AM Page 195 ⁄ From the MySQL monitor, type SHOW STATUS; and press Enter. ■ The current values of the status variables are displayed. ¤ Type SHOW STATUS LIKE '%Opened%'; and press Enter. ■ The variables that match the string you specified are displayed. T o optimize the performance of a MySQL server, the first step is to determine how it is currently performing. MySQL keeps a number of running status variables that you can examine to get a snapshot of the server's current performance. This will let you know how much traffic the server is handling, as well as early indications of performance problems. To display the status variables, you can use the SHOW STATUS command in the MySQL monitor or the mysqladmin extended-status command at the command prompt. The output of either command is a table of variables and their values. You can use the LIKE operator with SHOW STATUS to show only certain values. Most of the variables count the number of times something has happened since the server started. For example, the Opened_tables value is the number of table that have been opened, and the Questions value is the number of queries the server has received. The Uptime value gives you the number of seconds the server has been running, so you can use this and the various variables to get an idea of how the server performs over time. The Threads_connected value indicates how many client connections are currently open to the server, and the Max_used_connections value is the largest number of clients that are simultaneously connected. You can use these values to determine whether the server is busy and whether the max_concurrent_users system variable needs to be increased. Slow_queries is another value you should watch carefully. This is the number of queries that have taken more than the expected amount of time. The time for a slow query is set using the long_query_time system variable. If many slow queries are being counted, this means the server is running slower than expected, or some particular queries in use are slowing down the server. DISPLAY SERVER PERFORMANCE INFORMATION MySQL 196 DISPLAY SERVER PERFORMANCE INFORMATION 516922 Ch10.F 9/26/02 11:38 AM Page 196 [...]... OPTIMIZE AND TROUBLESHOOT MYSQL 10 Newer versions of MySQL support a binary update log This is similar to the update log, but stored in a binary rather than text format for greater efficiency To use the binary log, specify the log-bin option on the command line to mysqld MySQL includes a utility, mysqlbinlog, to convert the binary log to a text format You can maintain the MySQL log files at any time... locks it for your session The lock ends automatically when you close your connection to the server or use another LOCK TABLES command You can also use the UNLOCK TABLES command to end all current locks: UNLOCK TABLES; The MySQL server gives priority to WRITE locks If you add the LOW_PRIORITY keyword before WRITE in a LOCK TABLES command, the server allows READ locks for other threads while it waits for. .. incorrectly can reduce performance or even cripple the server After you make changes to the configuration file, save it and exit the editor MySQL does not read the new configuration automatically You will need to shut down the server using the mysqladmin shutdown command and restart it to read the modified configuration information Chapter 9 includes information on stopping and restarting the MySQL server After... Ch10.F 9/26/02 11:38 AM Page 202 MySQL IMPROVE PERFORMANCE WITH INDEXES A MySQL table can have one or more indexes associated with it While at least one unique index is usually used as a primary key to uniquely identify each row of the table, you can add additional indexes, unique or not, to improve performance The MySQL status values shown by the SHOW STATUS command or the mysqladmin extended-status command... were read by using the SHOW VARIABLES command from the MySQL monitor, or the mysqladmin variables command from the command prompt notepad c: \mysql\ my.cnf EDIT THE CONFIGURATION FILE IN WINDOWS Note: Select Run from the Start menu to display the Run dialog box ⁄ From the Run dialog ¤ Within MySQLadmin, › Click the Save box, type C: \mysql\ bin\ winmysqladmin and press Enter click the my.ini Setup tab Modification... and the err extension under UNIX, and is stored at c: \mysql\ data \mysql. err on Windows systems MySQL can optionally create a general query log, which logs each query sent to the server This log file is created if you specify the -l or log option when you start the mysqld server Unless you specify a different name in the command line, the filename for this log is the hostname of the server, and the extension... option on the command line VIEW MYSQL LOG FILES ⁄ From the command prompt, type cd /usr/local/ mysql/ data and press Enter s This switches to the location where logs are stored 200 Note: The location may vary depending on your installation Under Windows, the location is usually c: \mysql\ data ¤ To view the error log, type cat hostname.err, replacing hostname with the name of your server, and press Enter... clients have connected to the server Open_ tables Number of tables currently in use Open_ files Number of files currently in use Opened_tables Number of tables opened since the server started Questions Total number of queries received from clients Slow_queries Number of queries that have taken longer than a maximum amount of time Threads_cached Number of threads standing by for clients Threads_created Number... prompted for the s This selects a database ˇ Type field1 INT); and › Type CREATE TABLE press Enter newtable ( and press Enter next line s This creates a new table Á Type SHOW STATUS LIKE '%Opened%'; and press Enter s Notice that the number of opened tables has increased 197 516922 Ch10.F 9/26/02 11:38 AM Page 198 MySQL EDIT THE CONFIGURATION FILE IN UNIX I f you decide to change some of the MySQL options... improve performance and press Enter Á Type DROP INDEX authorind; and press Enter s You are prompted for the s This removes the added next line index 203 516922 Ch10.F 9/26/02 11:38 AM Page 204 MySQL MANAGE TABLE LOCKING M ySQL uses a system of table locking to ensure that the various threads working with a table do not conflict When a thread has a lock on a table, MySQL prevents other threads from performing . file is divided into sections for different MySQL components. For example, the line [mysqld] begins the section for the server, and [client] begins the section for clients. Within each section,. queries. OPTIMIZE MYSQL MySQL Disk Access Disk access is usually the largest bottleneck affecting a MySQL server. Because tables are stored on disk, virtually every MySQL operation involves disk access. MySQL. less data MySQL has to work with, the faster it will be. Use the smallest column sizes possible for your data and eliminate unnecessary columns to improve the MySQL server's performance working