MySQL中使用SHOW PROFILE命令分析性能的用法整理 |
show profile是由Jeremy Cole捐献给MySQL社区版本的 。默认的是关闭的,但是会话级别可以开启这个功能 。开启它可以让MySQL收集在执行语句的时候所使用的资源 。为了统计报表,把profiling设为1 mysql> SET profiling = 1; mysql> SELECT COUNT(DISTINCT actor.first_name) AS cnt_name, COUNT(*) AS cnt -> FROM sakila.film_actor -> INNER JOIN sakila.actor USING(actor_id) -> GROUP BY sakila.film_actor.film_id -> ORDER BY cnt_name DESC; ... 997 rows in set (0.03 sec) mysql> SHOW PROFILES\G *************************** 1. row *************************** Query_ID: 1 Duration: 0.02596900 Query: SELECT COUNT(DISTINCT actor.first_name) AS cnt_name,... mysql> SHOW PROFILE; +------------------------+-----------+ | Status | Duration | +------------------------+-----------+ | (initialization) | 0.000005 | | Opening tables | 0.000033 | | System lock | 0.000037 | | Table lock | 0.000024 | | init | 0.000079 | | optimizing | 0.000024 | | statistics | 0.000079 | | preparing | 0.00003 | | Creating tmp table | 0.000124 | | executing | 0.000008 | | Copying to tmp table | 0.010048 | | Creating sort index | 0.004769 | | Copying to group table | 0.0084880 | | Sorting result | 0.001136 | | Sending data | 0.000925 | | end | 0.00001 | | removing tmp table | 0.00004 | | end | 0.000005 | | removing tmp table | 0.00001 | | end | 0.000011 | | query end | 0.00001 | | freeing items | 0.000025 | | removing tmp table | 0.00001 | | freeing items | 0.000016 | | closing tables | 0.000017 | | logging slow query | 0.000006 | +------------------------+-----------+ mysql> SHOW PROFILE CPU FOR QUERY 1; SHOW PROFILES显示最近发给服务器的多条语句,条数根据会话变量profiling_history_size定义,默认是15,最大值为100 。设为0等价于关闭分析功能 。 SHOW PROFILE FOR QUERY n,这里的n就是对应SHOW PROFILES输出中的Query_ID 。
mysql> show profiles; +----------+-------------+---------------------------------------+ | Query_ID | Duration | Query | +----------+-------------+---------------------------------------+ | 1 | 0.00037700 | alter table table1 drop column c3 int | | 2 | 70.37123800 | alter table table1 drop column c3 | | 3 | 0.00124500 | show tables | | 4 | 0.00569800 | select * from table1 where id=2 | | 5 | 0.00068500 | select count(1) from tables1 | | 6 | 0.00197900 | select count(1) from table1 | | 7 | 0.00105900 | alter table tables1 drop c1 | | 8 | 0.00800200 | alter table table1 drop c1 | +----------+-------------+---------------------------------------+ 8 rows in set (0.00 sec) mysql> SHOW PROFILE FOR QUERY 2; #查看alter table table1 drop column c3的分析 +------------------------------+-----------+ | Status | Duration | +------------------------------+-----------+ | starting | 0.000183 | | checking permissions | 0.000057 | | checking permissions | 0.000059 | | init | 0.000060 | | Opening tables | 0.000071 | | System lock | 0.000062 | | setup | 0.000080 | | creating table | 0.005052 | | After create | 0.000220 | | copy to tmp table | 0.000244 | | rename result table | 70.364027 | | end | 0.000575 | | Waiting for query cache lock | 0.000062 | | end | 0.000075 | | query end | 0.000057 | | closing tables | 0.000061 | | freeing items | 0.000080 | | logging slow query | 0.000056 | | logging slow query | 0.000098 | | cleaning up | 0.000059 | +------------------------------+-----------+ 20 rows in set (0.00 sec) 如果没有指定FOR QUERY,那么输出最近一条语句的信息 。 LIMIT部分的用法与SELECT中LIMIT子句一致,不赘述 。 type是可选的,取值范围可以如下:
例: mysql> show profile cpu for query 2; +------------------------------+-----------+----------+------------+ | Status | Duration | CPU_user | CPU_system | +------------------------------+-----------+----------+------------+ | starting | 0.000183 | 0.000000 | 0.000000 | | checking permissions | 0.000057 | 0.000000 | 0.000000 | | checking permissions | 0.000059 | 0.000000 | 0.000000 | | init | 0.000060 | 0.000000 | 0.000000 | | Opening tables | 0.000071 | 0.000000 | 0.000000 | | System lock | 0.000062 | 0.000000 | 0.000000 | | setup | 0.000080 | 0.000000 | 0.001000 | | creating table | 0.005052 | 0.003000 | 0.001000 | | After create | 0.000220 | 0.000000 | 0.000000 | | copy to tmp table | 0.000244 | 0.000000 | 0.000000 | | rename result table | 70.364027 | 7.470864 | 41.612674 | | end | 0.000575 | 0.000000 | 0.001000 | | Waiting for query cache lock | 0.000062 | 0.000000 | 0.000000 | | end | 0.000075 | 0.000000 | 0.000000 | | query end | 0.000057 | 0.000000 | 0.000000 | | closing tables | 0.000061 | 0.000000 | 0.000000 | | freeing items | 0.000080 | 0.000000 | 0.000000 | | logging slow query | 0.000056 | 0.000000 | 0.000000 | | logging slow query | 0.000098 | 0.000000 | 0.000000 | | cleaning up | 0.000059 | 0.000000 | 0.000000 | +------------------------------+-----------+----------+------------+ 20 rows in set (0.00 sec) ps: 作用范围 开启分析功能后,所有本会话中的语句都被分析(甚至包括执行错误的语句),除了SHOW PROFILE和SHOW PROFILES两句本身 。 应用示例:使用SHOW PROFILE分析查询缓存命中的性能优势 。 mysql> set profiling=1; Query OK, 0 rows affected (0.00 sec) mysql> select count(1) as cnt from tran_excution; +-------+ | cnt | +-------+ | 14225 | +-------+ 1 row in set (0.00 sec) 由于已经启用了查询缓存,相同查询(非分区表)可以直接在查询缓存中命中 。 mysql> select count(1) as cnt from tran_excution; 我们仔细看下两个同样的语句的分析 。 mysql> show profile source for query 1; +--------------------------------+----------+-----------------------+---------------+-------------+ | Status | Duration | Source_function | Source_file | Source_line | +--------------------------------+----------+-----------------------+---------------+-------------+ | starting | 0.000048 | NULL | NULL | NULL | | Waiting for query cache lock | 0.000013 | try_lock | sql_cache.cc | 454 | | checking query cache for query | 0.000040 | send_result_to_client | sql_cache.cc | 1561 | | checking permissions | 0.000023 | check_access | sql_parse.cc | 4751 | | Opening tables | 0.000040 | open_tables | sql_base.cc | 4831 | | System lock | 0.000020 | mysql_lock_tables | lock.cc | 299 | | Waiting for query cache lock | 0.000037 | try_lock | sql_cache.cc | 454 | | init | 0.000020 | mysql_select | sql_select.cc | 2579 | | optimizing | 0.000015 | optimize | sql_select.cc | 865 | | statistics | 0.000017 | optimize | sql_select.cc | 1056 | | preparing | 0.000016 | optimize | sql_select.cc | 1078 | | executing | 0.000015 | exec | sql_select.cc | 1836 | | Sending data | 0.003875 | exec | sql_select.cc | 2380 | | end | 0.000018 | mysql_select | sql_select.cc | 2615 | | query end | 0.000015 | mysql_execute_command | sql_parse.cc | 4440 | | closing tables | 0.000016 | mysql_execute_command | sql_parse.cc | 4492 | | freeing items | 0.000016 | mysql_parse | sql_parse.cc | 5640 | | Waiting for query cache lock | 0.000012 | try_lock | sql_cache.cc | 454 | | freeing items | 0.000032 | NULL | NULL | NULL | | Waiting for query cache lock | 0.000017 | try_lock | sql_cache.cc | 454 | | freeing items | 0.000016 | NULL | NULL | NULL | | storing result in query cache | 0.000017 | end_of_result | sql_cache.cc | 1020 | | logging slow query | 0.000018 | log_slow_statement | sql_parse.cc | 1461 | | logging slow query | 0.000050 | log_slow_statement | sql_parse.cc | 1470 | | cleaning up | 0.000018 | dispatch_command | sql_parse.cc | 1417 | +--------------------------------+----------+-----------------------+---------------+-------------+ 25 rows in set (0.00 sec) mysql> show profile source for query 2; +--------------------------------+----------+-----------------------+--------------+-------------+ | Status | Duration | Source_function | Source_file | Source_line | +--------------------------------+----------+-----------------------+--------------+-------------+ | starting | 0.000051 | NULL | NULL | NULL | | Waiting for query cache lock | 0.000014 | try_lock | sql_cache.cc | 454 | | checking query cache for query | 0.000016 | send_result_to_client | sql_cache.cc | 1561 | | checking privileges on cached | 0.000013 | send_result_to_client | sql_cache.cc | 1652 | | checking permissions | 0.000015 | check_access | sql_parse.cc | 4751 | | sending cached result to clien | 0.000036 | send_result_to_client | sql_cache.cc | 1749 | | logging slow query | 0.000017 | log_slow_statement | sql_parse.cc | 1461 | | cleaning up | 0.000018 | dispatch_command | sql_parse.cc | 1417 | +--------------------------------+----------+-----------------------+--------------+-------------+ 8 rows in set (0.00 sec) 可以清晰地看到缓存中命中时,大大节省了后台的开销 。当然缓存的使用也需要根据各种场景(表的数据规模,更新频率等)考察使用,并不是启用缓存就一定能够提高查询效率 。这里仅仅作为SHOW PROFILE的一个应用示例 。 |