分布式主动感知在智能运维中的实践|分享实录(物联网中全面感知特点能够感知哪些数据)
530
2022-05-28
Gaussdb 100在集群部署下,支持通过路由规则将SQL语句发送给对应的DN节点执行,执行过程如下图所示:
以下为访问示例(演示环境为三节点分布式集群)。
新建路由规则
DROP DISTRIBUTE RULE IF EXISTS enmo_muggle; CREATE DISTRIBUTE RULE enmo_muggle(key char(32) not null) AS LIST(key) ( groupid 1 values ('dn1'), groupid 2 values ('dn2'), groupid 3 values ('dn3') );
创建表
DROP TABLE IF EXISTS muggle; CREATE TABLE muggle(id int, key char(32) not null) DISTRIBUTE BY LIST(key) ( groupid 1 values ('dn1'), groupid 2 values ('dn2'), groupid 3 values ('dn3') );
3.未设置路由规则,操作数据时,会自动将数据按(groupid)落盘至指定DN节点。
SQL> route by null; Succeed. SQL> SQL> INSERT INTO muggle VALUES(1, 'dn1'); 1 rows affected. SQL> INSERT INTO muggle VALUES(2, 'dn2'); 1 rows affected. SQL> INSERT INTO muggle VALUES(3, 'dn3'); 1 rows affected. SQL> commit; Succeed. SQL> select * from muggle; ID KEY ------------ -------------------------------- 1 dn1 2 dn2 3 dn3 3 rows fetched. SQL> route by node 1; Succeed. SQL> select * from muggle; ID KEY ------------ -------------------------------- 1 dn1 1 rows fetched. SQL> route by node 2; Succeed. SQL> select * from muggle; ID KEY ------------ -------------------------------- 2 dn2 1 rows fetched. SQL> route by node 3; Succeed. SQL> select * from muggle; ID KEY ------------ -------------------------------- 3 dn3 1 rows fetched. SQL>
指定路由规则名操作数据,则按路由规则名落盘数据。
SQL> ROUTE BY RULE enmo_muggle(key) VALUES('dn1'); Succeed. SQL> INSERT INTO muggle VALUES(1, 'dn1'); 1 rows affected. SQL> ROUTE BY RULE enmo_muggle(key) VALUES('dn2'); Succeed. SQL> INSERT INTO muggle VALUES(2, 'dn2'); 1 rows affected. SQL> ROUTE BY RULE enmo_muggle(key) VALUES('dn3'); Succeed. SQL> INSERT INTO muggle VALUES(3, 'dn3'); 1 rows affected. SQL> commit; Succeed. SQL> route by node 1; Succeed. SQL> select * from muggle; ID KEY ------------ -------------------------------- 1 dn1 1 rows fetched. SQL> route by node 2; Succeed. SQL> select * from muggle; ID KEY ------------ -------------------------------- 2 dn2 1 rows fetched. SQL> route by node 3; Succeed. SQL> select * from muggle; ID KEY ------------ -------------------------------- 3 dn3 1 rows fetched. SQL> route by null; Succeed. SQL> select * from muggle; ID KEY ------------ -------------------------------- 1 dn1 2 dn2 3 dn3 3 rows fetched.
操作数据时,若指定按(groupid)节点ID,则数据落盘在指定的DN节点,操作数据不会按规则名进行落盘。
SQL> route by node 1; Succeed. SQL> insert into muggle values(1,'dn1'); 1 rows affected. SQL> insert into muggle values(2,'dn2'); 1 rows affected. SQL> insert into muggle values(3,'dn3'); 1 rows affected. SQL> commit; Succeed. SQL> select * from muggle; ID KEY ------------ -------------------------------- 1 dn1 2 dn2 3 dn3 3 rows fetched. SQL> route by node 2; Succeed. SQL> select * from muggle; ID KEY ------------ -------------------------------- 0 rows fetched. SQL> route by node 3; Succeed. SQL> select * from muggle; ID KEY ------------ -------------------------------- 0 rows fetched.
本文转自“墨天轮”社区GaussDB频道
GaussDB
版权声明:本文内容由网络用户投稿,版权归原作者所有,本站不拥有其著作权,亦不承担相应法律责任。如果您发现本站中有涉嫌抄袭或描述失实的内容,请联系我们jiasou666@gmail.com 处理,核实后本网站将在24小时内删除侵权内容。