标签：group   nan   let   0 rows   locate   cal   全表扫描   page   sts   

之前在《MySQL sys Schema 简单介绍-1》中简单的介绍了，sys Schema库中的表。那么这些表都可以查询些什么信息呢？接下来本文将做下介绍。

1. 表的情况

1.1 统计下哪些表访问量占前十位

 select table_schema,table_name,sum(io_read_requests+io_write_requests) io from schema_table_statistics group by table_schema,table_name order by io desc limit 10;
 
+--------------+----------------------+----------+
| table_schema | table_name           | io       |
+--------------+----------------------+----------+
| tb0          | tb0                  | 16148445 |
| tb1          | tb1                  |   394803 |
| tb2          | tb2                  |    30842 |
| tb3          | tb3                  |    30544 |
| tb4          | tb4                  |    30505 |
| tb5          | tb5                  |    30041 |
| tb6          | tb6                  |    30011 |
| tb7          | tb7                  |    29836 |
| tb8          | tb8                  |    29730 |
| tb9          | tb9                  |    29603 |
+--------------+----------------------+----------+

1.2 哪个表占用了最多的buffer pool

mysql>  select * from innodb_buffer_stats_by_table order by allocated desc limit 1\G
*************************** 1. row ***************************
object_schema: db
  object_name: db1 
    allocated: 96.00 KiB
         data: 59.53 KiB
        pages: 6
 pages_hashed: 0
    pages_old: 3
  rows_cached: 797
1 row in set (1.11 sec)

1.3 查看表的全表扫描情况，看看哪些表需要做下优化

mysql> select * from schema_tables_with_full_table_scans limit 5 \G
*************************** 1. row ***************************
    object_schema: db0
      object_name: tb0
rows_full_scanned: 21072153342
          latency: 1.28 d
*************************** 2. row ***************************
    object_schema: db1
      object_name: tb1
rows_full_scanned: 19108957224
          latency: 1.55 d
*************************** 3. row ***************************
    object_schema: db2
      object_name: tb2
rows_full_scanned: 9778136634
          latency: 17.10 h
*************************** 4. row ***************************
    object_schema: db3
    object_name  : tb3
rows_full_scanned: 8340304503
          latency: 4.79 h
*************************** 5. row ***************************
    object_schema: db4
      object_name: tb4
rows_full_scanned: 6355100618
          latency: 11.86 h
5 rows in set (1.70 sec)

2. 索引的情况

2.1 有哪些无用的索引(即冗余索引)

mysql>  select * from schema_redundant_indexes \G
*************************** 1. row ***************************
              table_schema: tb1
                table_name: tb1
      redundant_index_name: xxx
   redundant_index_columns: xxx_prefix
redundant_index_non_unique: 1
       dominant_index_name: xxx_prefix
    dominant_index_columns: xxx_prefix,xxxx
 dominant_index_non_unique: 0
            subpart_exists: 0
            sql_drop_index: ALTER TABLE `db1`.`tb1` DROP INDEX `xxx`

2.2 哪些索引没有用到

mysql> select * from schema_unused_indexes limit 2 \G
*************************** 1. row ***************************
object_schema: db1
  object_name: tb1
   index_name: xxxxx
*************************** 2. row ***************************
object_schema: db2
  object_name: tb2
   index_name: xxxxxx
2 rows in set (0.04 sec)

2.3 查看索引的select \update\delete\insert情况

mysql> select * from schema_index_statistics limit 1 \G
*************************** 1. row ***************************
  table_schema: tb1
    table_name: tb1
    index_name: xxxxxxx
 rows_selected: 376085833
select_latency: 1.58 h
 rows_inserted: 0
insert_latency: 0 ps
  rows_updated: 39676093
update_latency: 28.00 m
  rows_deleted: 0
delete_latency: 0 ps
1 row in set (0.04 sec)

3. 语句相关

3.1 数据库中哪些SQL被频繁执行

mysql> select db,exec_count,query from statement_analysis order by exec_count desc limit 10;
+-------------+------------+-------------------------------------------------------------------+
| db          | exec_count | query                                                             |
+-------------+------------+-------------------------------------------------------------------+
| db1         |      20174 | SELECT `date_format` ( `locati ... AL ? SQL_TSI_SECOND ) LIMIT ?  |
| db2         |      19461 | SELECT `date_format` ( `locati ... AL ? SQL_TSI_SECOND ) LIMIT ?  |
+-------------+------------+-------------------------------------------------------------------+

3.2 哪些SQL执行了全表扫描或执行了排序操作

mysql> select * from statements_with_sorting limit 1 \G
*************************** 1. row ***************************
            query: SELECT `xxxx` FROM `xxx ... ER BY COUNT ( `xxxx` ) DESC
               db: db1
       exec_count: 85
    total_latency: 1.87 m
sort_merge_passes: 0
  avg_sort_merges: 0
sorts_using_scans: 85
 sort_using_range: 0
      rows_sorted: 2295
  avg_rows_sorted: 27
       first_seen: 2018-07-27 11:59:17
        last_seen: 2018-09-01 18:59:19
           digest: 1419efc6ce6a95e654f128cded18e49a

mysql> select * from statements_with_full_table_scans limit 1 \G
*************************** 1. row ***************************
                   query: SELECT `xxx` ( `xxx` , ?  ...  . `xxx` ORDER BY `xxx`
                      db: db1
              exec_count: 1
           total_latency: 998.92 us
     no_index_used_count: 1
no_good_index_used_count: 0
       no_index_used_pct: 100
               rows_sent: 48
           rows_examined: 192
           rows_sent_avg: 48
       rows_examined_avg: 192
              first_seen: 2018-10-15 16:58:03
               last_seen: 2018-10-15 16:58:03
                  digest: 1c7129051c2dbad8dfc34ad6009eda7f
1 row in set (0.19 sec)

3.3 哪些SQL语句使用了临时表，又有哪些用到了磁盘临时表

mysql> select db, query, tmp_tables, tmp_disk_tables  from statement_analysis where tmp_tables>0 or tmp_disk_tables >0 order by (tmp_tables+tmp_disk_tables) desc limit 3;
+-------+-------------------------------------------------------------------+------------+-----------------+
| db    | query                                                             | tmp_tables | tmp_disk_tables |
+-------+-------------------------------------------------------------------+------------+-----------------+
| xxxxx | SHOW TABLES LIKE ?                                                |      54917 |               0 |
| xxxxx | SELECT `xxxx` ( `xxxx` , ?  ...  . `xxx` ORDER BY `xxxxxxxxxxxx`  |      14957 |               0 |
| xxxxx | SELECT `xxxx` ( `xxxx` , ?  ...  . `xxx` ORDER BY `xxxxxxxxxxxx`  |      14957 |               0 |
+-------+-------------------------------------------------------------------+------------+-----------------+
3 rows in set (0.03 sec)

4. 用户相关

4.1 排查每个host的资源消耗情况

mysql> select * from host_summary \G
*************************** 1. row ***************************
                  host: xxxx
            statements: 2048
     statement_latency: 4.81 s
 statement_avg_latency: 2.35 ms
           table_scans: 917
              file_ios: 7882
       file_io_latency: 2.51 s
   current_connections: 0
     total_connections: 6
          unique_users: 1
        current_memory: 0 bytes
total_memory_allocated: 0 bytes

4.2 每个host的主要是在什么文件类型上耗费IO资源

mysql> select  * from host_summary_by_file_io_type limit 10;
+----------------+--------------------------------------+-------+---------------+-------------+
| host           | event_name                           | total | total_latency | max_latency |
+----------------+--------------------------------------+-------+---------------+-------------+
| xxxxxxxxxxxxx  | wait/io/file/sql/FRM                 |  2868 | 1.62 s        | 71.06 ms    |
| xxxxxxxxxxxxx  | wait/io/file/innodb/innodb_data_file |   387 | 870.14 ms     | 55.22 ms    |
| xxxxxxxxxxxxx  | wait/io/file/myisam/dfile            |  4627 | 29.50 ms      | 8.70 ms     |
| xxxxxxxxxxxxx  | wait/io/file/sql/FRM                 |    24 | 115.50 ms     | 48.00 ms    |
| xxxxxxxxxxxxx  | wait/io/file/myisam/dfile            |   145 | 105.90 ms     | 56.94 ms    |
| xxxxxxxxxxxxx  | wait/io/file/sql/FRM                 |    45 | 162.71 ms     | 28.65 ms    |
| xxxxxxxxxxxxx  | wait/io/file/myisam/dfile            |   581 | 61.59 ms      | 30.83 ms    |
| xxxxxxxxxxxxx  | wait/io/file/myisam/kfile            |     8 | 2.28 ms       | 1.37 ms     |
| xxxxxxxxxxxxx  | wait/io/file/sql/FRM                 |  1107 | 1.08 s        | 66.09 ms    |
| xxxxxxxxxxxxx  | wait/io/file/myisam/dfile            |  3321 | 39.72 ms      | 34.06 ms    |
+----------------+--------------------------------------+-------+---------------+-------------+
10 rows in set (0.02 sec)

4.3 查看每个host的语句执行情况

mysql> select  * from host_summary_by_statement_latency  limit 1 ;
+-----------+------------+---------------+-------------+--------------+-----------+---------------+---------------+------------+
| host      | total      | total_latency | max_latency | lock_latency | rows_sent | rows_examined | rows_affected | full_scans |
+-----------+------------+---------------+-------------+--------------+-----------+---------------+---------------+------------+
| localhost | 3958366834 | 5.98 d        | 20.24 s     | 5.10 h       | 692352951 |    1155909355 |      92068180 |   26536936 |
+-----------+------------+---------------+-------------+--------------+-----------+---------------+---------------+------------+
1 row in set (0.02 sec)

4.4 查看某个host 执行的语句在各个阶段的耗时

mysql> select * from host_summary_by_statement_type  where host = ‘xxxxxxxxxxxxxx‘  ;
+----------------+----------------+-------+---------------+-------------+--------------+-----------+---------------+---------------+------------+
| host           | statement      | total | total_latency | max_latency | lock_latency | rows_sent | rows_examined | rows_affected | full_scans |
+----------------+----------------+-------+---------------+-------------+--------------+-----------+---------------+---------------+------------+
| xxxxxxxxxxxxxx | select         |  6684 | 1.30 m        | 39.87 ms    | 1.75 s       |   2313061 |      25859716 |             0 |       6660 |
| xxxxxxxxxxxxxx | show_tables    |  6687 | 4.97 s        | 3.66 ms     | 723.36 ms    |     13962 |         13962 |             0 |       6687 |
| xxxxxxxxxxxxxx | begin          |   346 | 1.36 s        | 37.39 ms    | 0 ps         |         0 |             0 |             0 |          0 |
| xxxxxxxxxxxxxx | insert         |   692 | 167.14 ms     | 436.55 us   | 88.72 ms     |         0 |             0 |           692 |          0 |
| xxxxxxxxxxxxxx | delete         |    34 | 117.19 ms     | 67.60 ms    | 67.61 ms     |         0 |          5874 |           690 |          0 |
| xxxxxxxxxxxxxx | Field List     |   271 | 57.89 ms      | 527.03 us   | 0 ps         |         0 |             0 |             0 |          0 |
| xxxxxxxxxxxxxx | set_option     |   240 | 18.77 ms      | 133.34 us   | 0 ps         |         0 |             0 |             0 |          0 |
| xxxxxxxxxxxxxx | show_warnings  |   116 | 9.57 ms       | 149.68 us   | 0 ps         |         0 |             0 |             0 |          0 |
| xxxxxxxxxxxxxx | commit         |     2 | 7.69 ms       | 4.58 ms     | 0 ps         |         0 |             0 |             0 |          0 |
| xxxxxxxxxxxxxx | show_databases |     1 | 1.41 ms       | 1.41 ms     | 585.00 us    |         7 |             7 |             0 |          1 |
| xxxxxxxxxxxxxx | Quit           |    49 | 429.66 us     | 20.43 us    | 0 ps         |         0 |             0 |             0 |          0 |
| xxxxxxxxxxxxxx | Init DB        |     1 | 49.35 us      | 49.35 us    | 0 ps         |         0 |             0 |             0 |          0 |
+----------------+----------------+-------+---------------+-------------+--------------+-----------+---------------+---------------+------------+
12 rows in set (0.01 sec)

5. innodb 相关

5.1 查看下各个数据库的内存占用情况

mysql> select * from innodb_buffer_stats_by_schema ;
+---------------+------------+------------+-------+--------------+-----------+-------------+
| object_schema | allocated  | data       | pages | pages_hashed | pages_old | rows_cached |
+---------------+------------+------------+-------+--------------+-----------+-------------+
| XXXXXXXXXXX   | 344.08 MiB | 293.56 MiB | 22021 |        13900 |      6250 |      973010 |
| XXXXXXXXXXX   | 129.67 MiB | 77.38 MiB  |  8299 |         2639 |      4874 |     1022327 |
| XXXXXXXXXXX   | 48.55 MiB  | 39.89 MiB  |  3107 |            0 |      1375 |      552809 |
| XXXXXXXXXXX   | 24.81 MiB  | 22.56 MiB  |  1588 |           28 |        61 |       25596 |
| XXXXXXXXXXX   | 944.00 KiB | 537.57 KiB |    59 |           44 |        50 |        5227 |
| XXXXXXXXXXX   | 128.00 KiB | 24.60 KiB  |     8 |            3 |         6 |         157 |
| XXXXXXXXXXX   | 64.00 KiB  | 4.34 KiB   |     4 |            0 |         4 |          25 |
| XXXXXXXXXXX   | 16.00 KiB  | 1.50 KiB   |     1 |            0 |         0 |          17 |
| XXXXXXXXXXX   | 16.00 KiB  | 338 bytes  |     1 |            0 |         1 |           6 |
+---------------+------------+------------+-------+--------------+-----------+-------------+
9 rows in set (0.74 sec)

5.2 查看各个表内存占用情况

mysql> select * from innodb_buffer_stats_by_table limit 2 \G;
*************************** 1. row ***************************
object_schema: db1
  object_name: tb1
    allocated: 213.92 MiB
         data: 175.95 MiB
        pages: 13691
 pages_hashed: 12123
    pages_old: 2306
  rows_cached: 1154997
*************************** 2. row ***************************
object_schema: db2
  object_name: tb2
    allocated: 82.75 MiB
         data: 76.26 MiB
        pages: 5296
 pages_hashed: 0
    pages_old: 2183
  rows_cached: 1065643
2 rows in set (0.96 sec)

6.结束语

上述是我整理的一些常见的查询。当然还有很多的相关运用，大家可以查询下mysql的文档。

MySQL sys Schema 简单介绍-2

标签：group   nan   let   0 rows   locate   cal   全表扫描   page   sts   

原文地址：https://www.cnblogs.com/bush2582/p/10202159.html