Shulou(Shulou.com)06/02 Report--

After the amount of data comes up, when the single database and single table can not bear it, we need to introduce sub-database and sub-table. However, after dividing the database and tables, the primary key id cannot rely on MySQL and needs to obtain id from the outside.

At present, DBLE, which is used in our production environment as the middleware of sub-library and sub-table, has its own global id generator similar to snowflake scheme, which can also be implemented based on database.

However, in a larger company, global id is a widely used service that is usually provided independently as a public service.

Here we take the Leaf by Meituan Dianping as an example to see its effect.

Related documentation:

Https://tech.meituan.com/2019/03/07/open-source-project-leaf.html

Https://tech.meituan.com/2017/04/21/mt-leaf.html

Detailed explanation of https://blog.csdn.net/bskfnvjtlyzmv867/article/details/90175306 source code

The two official documents are introduced in detail, so we won't talk about it.

The following demonstration is based on the database for id distribution (the SLA of the database can be guaranteed by using MHA, pxc, or mgr)

0 environment

OS version: CentOS7

Node1 IP: 192.168.20.10

Node2 IP: 192.168.20.17

MySQL address: 192.168.20.10

1 create database tables, etc.

CREATE DATABASE leaf; use leaf; CREATE TABLE `leaf_ alloc` (`biz_ tag` varchar (12828) NOT NULL DEFAULT'',-- your biz unique name `max_ id` bigint (20) NOT NULL DEFAULT'1), `Step`int (11) NOT NULL, `substitution`varchar (256) DEFAULT NULL, `update_ time`timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP, PRIMARY KEY (`biz_ tag`) ENGINE=InnoDB -insert several identifiers of services that need to distribute id (I add two examples of card coupon id and post id here on the default basis) insert into leaf_alloc (biz_tag, max_id, step, description) values ('leaf-segment-test', 1, 2000,' Test leaf Segment Mode Get Id'); insert into leaf_alloc (biz_tag, max_id, step, description) values ('coupon', 1, 2000,' Get coupon Id') Insert into leaf_alloc (biz_tag, max_id, step, description) values ('tid', 1, 1000,' Get tiezi id'); select * from leaf_alloc +-- +-+ | biz_tag | max_id | step | description | update_time | | +-+-+ | coupon | 1 | 2000 | Get coupon Id | 2019-10-12 17:47:23 | | leaf-segment-test | 1 | 2000 | Test leaf Segment Mode Get Id | 2019-10-12 17:47:00 | | tid | 1 | 1000 | Get tiezi id | 2019-10-12 17:48:21 | +- -- +-+ 3 rows in set (0.001 sec)-- create a separate database account create user leaf@'%' identified by 'leaf1234' Grant select,update,delete,insert on leaf.* to leaf@'%'

2 compile leaf-server

Note: compilation and startup operations need to be performed separately on node1 and node2, and compilation + startup depends on maven and oracle-jdk

Cd / usr/local/git clone https://github.com/Meituan-Dianping/Leaf.gitcd leafmvn clean install-DskipTestscd leaf-server

Vim leaf-server/src/main/resources/leaf.properties modifies the connection of the database in the configuration file

Leaf.name=com.sankuai.leaf.opensource.testleaf.segment.enable=trueleaf.jdbc.url=jdbc:mysql://192.168.20.10:3306/leaf?useUnicode=true&characterEncoding=utf8&characterSetResults=utf8leaf.jdbc.username=leafleaf.jdbc.password=leaf1234leaf.snowflake.enable=false#leaf.snowflake.zk.address=#leaf.snowflake.port=

3 start leaf-server:

Cd / usr/local/leaf/leaf-servermvn spring-boot:run

4 Test

Test the number generator of the 3 biz_tag inserted above me:

Curl http://192.168.20.10:8080/api/segment/get/leaf-segment-test

Curl http://192.168.20.10:8080/api/segment/get/coupon

Curl

Http://192.168.20.10:8080/api/segment/get/tid

5 other

The number segment mode comes with a monitoring interface:

Http://192.168.20.10:8080/cache

Test the generation of id:

For i in {1... 2000}; do curl http://192.168.20.17:8080/api/segment/get/coupon; donefor i in {1... 2000}; do curl http://192.168.20.10:8080/api/segment/get/coupon; done

Then, when we obtain the id in the for loop, we disable the MySQL artificially. As shown below, we can see that at the beginning, the leaf can continue to send numbers because of the number range + double buffer, but when all the pre-allocated id is used up, an error will be reported:

As long as it doesn't take too long for our MySQL database to fail over, and if we set different biz_tag to different sizes of step, communication will not affect the leaf service.

The following is the embodiment of dual buffer in the web interface:

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.