The Slow Query Log

本文转载自：http://dev.mysql.com/doc/refman/5.5/en/slow-query-log.html

The Slow Query Log

The slow query log consists of SQL statements that took more than long_query_time seconds to execute and required at least min_examined_row_limit rows to be examined. The minimum and default values oflong_query_time are 0 and 10, respectively. The value can be specified to a resolution of microseconds. For logging to a file, times are written including the microseconds part. For logging to tables, only integer times are written; the microseconds part is ignored.

By default, administrative statements are not logged, nor are queries that do not use indexes for lookups. This behavior can be changed using --log-slow-admin-statements and log_queries_not_using_indexes, as described later.

The time to acquire the initial locks is not counted as execution time. mysqld writes a statement to the slow query log after it has been executed and after all locks have been released, so log order might differ from execution order.

By default, the slow query log is disabled. To specify the initial slow query log state explicitly, use --slow_query_log[={0|1}]. With no argument or an argument of 1, --slow_query_log enables the log. With an argument of 0, this option disables the log. To specify a log file name, use --slow_query_log_file=file_name. To specify the log destination, use --log-output (as described inSection 5.2.1, “Selecting General Query and Slow Query Log Output Destinations”). The older option to enable the slow query log file, --log-slow-queries, is deprecated.

If you specify no name for the slow query log file, the default name is host_name-slow.log. The server creates the file in the data directory unless an absolute path name is given to specify a different directory.

To disable or enable the slow query log or change the log file name at runtime, use the global slow_query_logand slow_query_log_file system variables. Set slow_query_log to 0 (or OFF) to disable the log or to 1 (orON) to enable it. Set slow_query_log_file to specify the name of the log file. If a log file already is open, it is closed and the new file is opened.

When the slow query log is enabled, the server writes output to any destinations specified by the --log-outputoption or log_output system variable. If you enable the log, the server opens the log file and writes startup messages to it. However, further logging of queries to the file does not occur unless the FILE log destination is selected. If the destination is NONE, the server writes no queries even if the slow query log is enabled. Setting the log file name has no effect on logging if the log destination value does not contain FILE.

The server writes less information to the slow query log (and binary log) if you use the --log-short-formatoption.

To include slow administrative statements in the statements written to the slow query log, use the --log-slow-admin-statements server option. Administrative statements include ALTER TABLE, ANALYZE TABLE, CHECK TABLE, CREATE INDEX, DROP INDEX, OPTIMIZE TABLE, and REPAIR TABLE.

To include queries that do not use indexes for row lookups in the statements written to the slow query log, enable the log_queries_not_using_indexes system variable. When such queries are logged, the slow query log may grow quickly.

The server uses the controlling parameters in the following order to determine whether to write a query to the slow query log:

1. The query must either not be an administrative statement, or --log-slow-admin-statements must have been specified.

2. The query must have taken at least long_query_time seconds, or log_queries_not_using_indexes must be enabled and the query used no indexes for row lookups.

3. The query must have examined at least min_examined_row_limit rows.

The server does not write queries handled by the query cache to the slow query log, nor queries that would not benefit from the presence of an index because the table has zero rows or one row.

By default, a replication slave does not write replicated queries to the slow query log. To change this, use the --log-slow-slave-statements server option.

The slow query log should be protected because logged statements might contain passwords. SeeSection 6.1.2.3, “Passwords and Logging”.

The slow query log can be used to find queries that take a long time to execute and are therefore candidates for optimization. However, examining a long slow query log can become a difficult task. To make this easier, you can process a slow query log file using the mysqldumpslow command to summarize the queries that appear in the log. See Section 4.6.8, “mysqldumpslow — Summarize Slow Query Log Files”.

Previous / Next / Up / Table of Contents

User Comments

Posted by Andrea Gangini on February 8 2005 4:29pm[Delete] [Edit]

Please note that the mysql slow query log will not show the SQL of your slow queries if your application uses prepared statement.
You may want to turn on General Query Log instead.

Posted by Karen Owen on March 30 2006 8:34pm[Delete] [Edit]

Here are instructions on getting mysqldumpslow to work in your environment if you have my.cnf in a non-standard (eg, non /etc) locale. Please note this is for Unix only.

1. Make sure it can find perl in /usr/local/bin/perl
2. Make sure you've got the slow log running first
3. Copy the slow log to datadir and name it specifically: servername-slow.log (eg, servername-slow.log). This assumes your actual slowlog is located somewhere else. If it is located by default in datadir then you still must make a copy of it, named hostname-slow.log in datadir directory.
4. Make a copy of mysqldumpslow in the $MYSQL_HOME/bin directory, and name it: mysqldumpslow_new and be sure it is chmod 750
5. Edit the mysqldumpslow_new as described in the next item.
6. Assuming you don't place your server's my.cnf in /etc, you must tell my_print_defaults where datadir and basedir are because they can't be set directly in mysqldumpslow nor can they be passed as a parameter on the command line to mysqldumpslow. 
This is done by altering the command line option to my_print_defaults, the results of which are fed into the mysqldumpslow perl script to establish datadir, basedir, and a few other variables. The following line in
mysqldumpslow_new should be changed.
Change from:
my $defaults = `my_print_defaults mysqld`;

to:
my $defaults = `my_print_defaults -c /location/where/config_file_lives/my.cnf mysqld`; 

Make sure the my.cnf pointed to in the -c option has datadir and basedir set under [mysqld] section.
7. To run the program now, simply type:

./mysqldumpslow_new -s c -t 3

8. You have now displayed output from the slow query log telling you the top 3 slow queries on the system.

Posted by Christopher Egner on July 19 2006 8:16pm[Delete] [Edit]

Andrea Gangini posted that slow queries made via prepared statements do not post to the slow query log. This is not exactly true.

In the case of MySQL 5.0.22 when accessed through JDBC (Java 1.5.0), slow queries constructed via PreparedStatement always appear in the slow query log. Slow queries constructed normally (via Statement), sometimes appear. I suspect that PreparedStatements bypass the query cache and so are always reported in the log whereas regular Statements are checked against the query cache. If the results are taken from the query cache and not by accessing the tables, the query does not appear in slow query log. Testing regular Statements with the SQL_CACHE and SQL_NO_CACHE hints seems to confirm this. (Remember MySQL does not expend much effort trying to recognise queries it's already seen - if the query strings aren't exactly the same, don't expect to get help from the query cache.)

Chris

Posted by lvaro G. Vicario on December 20 2011 2:43pm[Delete] [Edit]

Syntax for /etc/my.cnf in Red Hat 9:

[mysqld]
long_query_time=1
log-slow-queries=/var/log/mysql/log-slow-queries.log

You must create the file manually and change owners this way:

mkdir /var/log/mysql
touch /var/log/mysql/log-slow-queries.log
chown mysql.mysql -R /var/log/mysql

Posted by Dan Franklin on December 8 2008 8:11pm[Delete] [Edit]

If you are loading a large database dump created using mysqldump in the normal way, be aware that if an insert statement takes longer than long_query_time, the ENTIRE huge insert statement (with all its individual rows) will be logged to the slow query file. This can significantly extend the time it takes to load the database.

Posted by Roel Van de Paar on December 8 2008 10:07pm[Delete] [Edit]

An example:

[mysqld]
log-slow-queries = slow.log
long_query_time = 20
log-queries-not-using-indexes

- The first line under [mysqld] turns on slow query log and logs all slow queries to slow.log in the MySQL data directory.
- The second line indicates that any queries that took more than 20 seconds to execute need to be logged.
- The last line tells MySQL to log *any* queries that do not use indexes regardless of the setting in the second line.

If you like you can set the long_query_time very high so that only the queries that do not use indexes are logged (and no 'general' slow queries).

注：近期参加MySQL运维学习，老师推荐该文章作为学习和技术提高的扩展阅读，先记录到自己的博客中，随后慢慢消化、学习、提高。本文章与“日志系统”主题相关。