本文介绍了规则和拓扑类查询语句。

SHOW RULE [FROM tablename]

使用说明:
  • show rule:查看数据库下每一个逻辑表的拆分情况;
  • show rule from tablename:查看数据库下指定逻辑表的拆分情况。
SHOW RULE;
+----+------------+-----------+------------------+---------------------+--------------------+------------------+---------------------+--------------------+
| ID | TABLE_NAME | BROADCAST | DB_PARTITION_KEY | DB_PARTITION_POLICY | DB_PARTITION_COUNT | TB_PARTITION_KEY | TB_PARTITION_POLICY | TB_PARTITION_COUNT |
+----+------------+-----------+------------------+---------------------+--------------------+------------------+---------------------+--------------------+
|  0 | k_1        |         0 | k                | hash                | 40                 | k                | hash                | 2                  |
|  1 | k_2        |         0 | k                | hash                | 40                 | k                | hash                | 2                  |
|  2 | sbtest1    |         0 | id               | hash                | 40                 | id               | hash                | 2                  |
|  3 | t1         |         0 | id               | hash                | 40                 | id               | hash                | 4                  |
+----+------------+-----------+------------------+---------------------+--------------------+------------------+---------------------+--------------------+
4 rows in set (0.05 sec)            
重要列详解:
  • BROADCAST:是否为广播表(0:否,1:是);
  • DB_PARTITION_KEY:分库的拆分键,没有分库的话,值为空;
  • DB_PARTITION_POLICY:分库的拆分策略,取值包括哈希或YYYYMM、YYYYDD、YYYYWEEK等日期策略;
  • DB_PARTITION_COUNT:分库数;
  • TB_PARTITION_KEY:分表的拆分键,没有分表的话,值为空;
  • TB_PARTITION_POLICY:分表的拆分策略,取值包括哈希或MM、DD、MMDD、WEEK等日期策略;
  • TB_PARTITION_COUNT:分表数。

SHOW FULL RULE [FROM tablename]

查看数据库下逻辑表的拆分规则,比SHOW RULE指令展示的信息更加详细。

SHOW FULL RULE;
+----+------------+-----------+------------+-----------------------+----------------------+--------------------------------------------------+-------------------+----------------------------------------+----------------+--------------------+
| ID | TABLE_NAME | BROADCAST | JOIN_GROUP | ALLOW_FULL_TABLE_SCAN | DB_NAME_PATTERN      | DB_RULES_STR                                     | TB_NAME_PATTERN   | TB_RULES_STR                           | PARTITION_KEYS | DEFAULT_DB_INDEX   |
+----+------------+-----------+------------+-----------------------+----------------------+--------------------------------------------------+-------------------+----------------------------------------+----------------+--------------------+
|  0 | k_1        |         0 | NULL       |                     1 | TEST1_{000000}_GROUP | ((#k,1,80#).longValue().abs() % 80).intdiv(2)    | k_1_cewR_{00}     | ((#k,1,80#).longValue().abs() % 80)    | k              | TEST1_SINGLE_GROUP |
|  1 | k_2        |         0 | NULL       |                     1 | TEST1_{000000}_GROUP | ((#k,1,80#).longValue().abs() % 80).intdiv(2)    | k_2_1xsQ_{00}     | ((#k,1,80#).longValue().abs() % 80)    | k              | TEST1_SINGLE_GROUP |
|  2 | sbtest1    |         0 | NULL       |                     1 | TEST1_{000000}_GROUP | ((#id,1,80#).longValue().abs() % 80).intdiv(2)   | sbtest1_wO5k_{00} | ((#id,1,80#).longValue().abs() % 80)   | id             | TEST1_SINGLE_GROUP |
|  3 | t1         |         0 | NULL       |                     1 | TEST1_{000000}_GROUP | ((#id,1,160#).longValue().abs() % 160).intdiv(4) | t1_EMrC_{000}     | ((#id,1,160#).longValue().abs() % 160) | id             | TEST1_SINGLE_GROUP |
+----+------------+-----------+------------+-----------------------+----------------------+--------------------------------------------------+-------------------+----------------------------------------+----------------+--------------------+
重要列详解:
  • BROADCAST:是否为广播表(0:否,1:是);
  • JOIN_GROUP:保留字段,暂时无意义。
  • ALLOW_FULL_TABLE_SCAN:分库分表在没有指定分表键值的情况下是否允许查询数据,如果配置为true,此时需要扫描每一个物理表来查找出符合条件的数据,简称为全表扫描;
  • DB_NAME_PATTERN:DB_NAME_PATTERN中{}之间的0为占位符,执行SQL时会被DB_RULES_STR计算出的值替代,并保持位数。比如,DB_NAME_PATTERN的值为SEQ_{0000}_RDS,DB_RULES_STR的值为[1,2,3,4],则会产生4个DB_NAME,分别为SEQ_0001_RDS、SEQ_0002_RDS、SEQ_0003_RDS、SEQ_0004_RDS;
  • DB_RULES_STR:具体的分库规则;
  • TB_NAME_PATTERN:TB_NAME_PATTERN中{}之间的0为占位符,执行SQL时会被TB_RULES_STR计算出的值替代,并保持位数。比如,TB_NAME_PATTERN的值为table_{00},TB_RULES_STR的值为[1,2,3,4,5,6,7,8],则会产生8张表,分别为table_01、table_02、table_03、table_04、table_05、table_06、table_07、table_08;
  • TB_RULES_STR:分表规则;
  • PARTITION_KEYS:分库和分表键集合,对于既分库又分表的情形,分库键在前,分表键在后;
  • DEFAULT_DB_INDEX:单库单表存放的分库。

SHOW TOPOLOGY FROM tablename

查看指定逻辑表的拓扑分布,展示该逻辑表保存在哪些分库中,每个分库下包含哪些分表。

SHOW TOPOLOGY FROM EMP; 
+------+--------------------------------------------------+------------+  
| ID   | GROUP_NAME                                       | TABLE_NAME |  
+------+--------------------------------------------------+------------+  
|    0 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0000_RDS | emp_0      |   
|    1 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0000_RDS | emp_1      |   
|    2 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0001_RDS | emp_0      |    
|    3 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0001_RDS | emp_1      |    
|    4 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0002_RDS | emp_0      |    
|    5 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0002_RDS | emp_1      |    
|    6 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0003_RDS | emp_0      |    
|    7 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0003_RDS | emp_1      |    
|    8 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0004_RDS | emp_0      |    
|    9 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0004_RDS | emp_1      |    
|   10 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0005_RDS | emp_0      |    
|   11 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0005_RDS | emp_1      |    
|   12 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0006_RDS | emp_0      |    
|   13 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0006_RDS | emp_1      |   
|   14 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0007_RDS | emp_0      |   
|   15 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0007_RDS | emp_1      |    
+------+--------------------------------------------------+------------+

SHOW PARTITIONS FROM tablename

查看分库分表键集合,分库键和分表键之间用逗号分割。如果最终结果有两个值,说明是既分库又分表的情形,第一个是分库键,第二个是分表键。如果结果只有一个值,说明是分库不分表的情形,该值是分库键。

SHOW PARTITIONS FROM EMP;
+-----------+
| KEYS      |
+-----------+
| emp_no,id |
+-----------+
1 row in set (0.00 sec)
            

SHOW BROADCASTS

查看广播表列表。

SHOW BROADCASTS;
+------+------------+
| ID   | TABLE_NAME |
+------+------------+
|    0 | brd2       |
|    1 | brd_tbl    |
+------+------------+
2 rows in set (0.01 sec)
            

SHOW DATASOURCES

查看底层存储信息,包含数据库名、数据库分组名、连接信息、用户名、底层存储类型、读写权重、连接池信息等。

SHOW DATASOURCES;
+------+----------------------------+------------------------------------------------+--------------------------------------------------+----------------------------------------------------------------------------------+-----------+-------+------+------+------+--------------+----------+--------------+---------------+----------------------------------------------+-------------+--------------+   
| ID   | SCHEMA                     | NAME                                           | GROUP                                            | URL                                                                              | USER      | TYPE  | INIT | MIN  | MAX  | IDLE_TIMEOUT | MAX_WAIT | ACTIVE_COUNT | POOLING_COUNT | ATOM                                         | READ_WEIGHT | WRITE_WEIGHT |
+------+----------------------------+------------------------------------------------+--------------------------------------------------+----------------------------------------------------------------------------------+-----------+-------+------+------+------+--------------+----------+--------------+---------------+----------------------------------------------+-------------+--------------+  
|    0 | seq_test_1487767780814rgkk | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0000_iiab_1 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0000_RDS | jdbc:mysql://rds1ur80kcv8g3t6p3ol.mysql.rds.aliyuncs.com:3306/seq_test_wnjg_0000 | jnkinsea0 | xdb   | 0    | 24   | 72   | 15           | 5000     | 0            | 1             | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0000_iiab | 10          | 10           |  
|    1 | seq_test_1487767780814rgkk | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0001_iiab_2 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0001_RDS | jdbc:mysql://rds1ur80kcv8g3t6p3ol.mysql.rds.aliyuncs.com:3306/seq_test_wnjg_0001 | jnkinsea0 | xdb   | 0    | 24   | 72   | 15           | 5000     | 0            | 1             | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0001_iiab | 10          | 10           |   
|    2 | seq_test_1487767780814rgkk | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0002_iiab_3 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0002_RDS | jdbc:mysql://rds1ur80kcv8g3t6p3ol.mysql.rds.aliyuncs.com:3306/seq_test_wnjg_0002 | jnkinsea0 | xdb   | 0    | 24   | 72   | 15           | 5000     | 0            | 1             | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0002_iiab | 10          | 10           | 
|    3 | seq_test_1487767780814rgkk | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0003_iiab_4 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0003_RDS | jdbc:mysql://rds1ur80kcv8g3t6p3ol.mysql.rds.aliyuncs.com:3306/seq_test_wnjg_0003 | jnkinsea0 | xdb   | 0    | 24   | 72   | 15           | 5000     | 0            | 1             | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0003_iiab | 10          | 10           |   
|    4 | seq_test_1487767780814rgkk | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0004_iiab_5 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0004_RDS | jdbc:mysql://rds1ur80kcv8g3t6p3ol.mysql.rds.aliyuncs.com:3306/seq_test_wnjg_0004 | jnkinsea0 | xdb   | 0    | 24   | 72   | 15           | 5000     | 0            | 1             | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0004_iiab | 10          | 10           |   
|    5 | seq_test_1487767780814rgkk | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0005_iiab_6 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0005_RDS | jdbc:mysql://rds1ur80kcv8g3t6p3ol.mysql.rds.aliyuncs.com:3306/seq_test_wnjg_0005 | jnkinsea0 | xdb   | 0    | 24   | 72   | 15           | 5000     | 0            | 1             | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0005_iiab | 10          | 10           |  
|    6 | seq_test_1487767780814rgkk | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0006_iiab_7 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0006_RDS | jdbc:mysql://rds1ur80kcv8g3t6p3ol.mysql.rds.aliyuncs.com:3306/seq_test_wnjg_0006 | jnkinsea0 | xdb   | 0    | 24   | 72   | 15           | 5000     | 0            | 1             | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0006_iiab | 10          | 10           |  
|    7 | seq_test_1487767780814rgkk | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0007_iiab_8 | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0007_RDS | jdbc:mysql://rds1ur80kcv8g3t6p3ol.mysql.rds.aliyuncs.com:3306/seq_test_wnjg_0007 | jnkinsea0 | xdb   | 0    | 24   | 72   | 15           | 5000     | 0            | 1             | rds1ur80kcv8g3t6p3ol_seq_test_wnjg_0007_iiab | 10          | 10           |  
+------+----------------------------+------------------------------------------------+--------------------------------------------------+----------------------------------------------------------------------------------+-----------+-------+------+------+------+--------------+----------+--------------+---------------+----------------------------------------------+-------------+--------------+

重要列详解:

  • SCHEMA:数据库名;
  • GROUP:数据库分组名。分组的目标是管理多组数据完全相同的数据库,比如通过 RDS(MySQL)进行数据复制后的主备数据库。主要用来解决读写分离、主备切换的问题;
  • URL:底层RDS(MySQL)的连接信息;
  • TYPE:底层存储类型。
  • READ_WEIGHT:读权重。在主实例的读压力比较大的时候,可以通过读写分离功能将读流量进行分流,减轻RDS主实例的压力。PolarDB-X会自动识别读写流量,引导写流量进入RDS主实例,读流量则按配置的权重流向所有RDS实例;
  • WRITE_WEIGHT:写权重。

SHOW NODE

查看物理库的读写次数(历史累计数据)、读写权重(历史累计数据)。

SHOW NODE;
+------+--------------------------------------------------+-------------------+------------------+---------------------+--------------------+
| ID   | NAME                                             | MASTER_READ_COUNT | SLAVE_READ_COUNT | MASTER_READ_PERCENT | SLAVE_READ_PERCENT |
+------+--------------------------------------------------+-------------------+------------------+---------------------+--------------------+
| 0    | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0000_RDS |                12 |                0 | 100%                | 0%                 |
| 1    | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0001_RDS |                 0 |                0 | 0%                  | 0%                 |
| 2    | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0002_RDS |                 0 |                0 | 0%                  | 0%                 |
| 3    | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0003_RDS |                 0 |                0 | 0%                  | 0%                 |
| 4    | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0004_RDS |                 0 |                0 | 0%                  | 0%                 |
| 5    | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0005_RDS |                 0 |                0 | 0%                  | 0%                 |
| 6    | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0006_RDS |                 0 |                0 | 0%                  | 0%                 |
| 7    | SEQ_TEST_1487767780814RGKKSEQ_TEST_WNJG_0007_RDS |                 0 |                0 | 0%                  | 0%                 |
+------+--------------------------------------------------+-------------------+------------------+---------------------+--------------------+
8 rows in set (0.01 sec)
            

重要列详解:

  • MASTER_COUNT:RDS主实例处理的读写查询次数(历史累计数据);
  • SLAVE_COUNT:RDS备实例处理的只读查询次数(历史累计数据);
  • MASTER_PERCENT:RDS主实例处理的读写查询占比(注意该列显示的是累计的实际数据占比,并不是用户配置的百分比);
  • SLAVE_PERCENT:RDS备实例处理的读写查询占比(注意该列显示的是累计的实际数据占比,并不是用户配置的百分比)。
说明
  • 事务中的只读查询会被发送到RDS主实例;
  • 由于MASTER_PERCENTSLAVE_PERCENT这两列代表的是历史累计数据,更改读写权重的配比后,这几个数值并不能立即反应最新的读写权重配比,需累计一段比较长的时间才行。