3.7 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 3.4 HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 3 315 Implementation classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316 5.9 HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317 services based on stateless ser‐ vices. At the same time, it can sense and use the underlying database’s HA solution to achieve its overall high availability. Data Migra‐ tion Data migration is the key

3.7 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3.4 HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 3 342 Implementation classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343 5.9 HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344 storage and computing platform. Guar‐ antee the HA of your distributed database cluster with ShardingSphere’s Operator on Ku‐ bernetes, and the native HA of your existing data sources. Data Mi‐ gra‐

分布式治理的相关研发工作 02 Apache ShardingSphere 高可用介绍 Apache ShardingSphere 高可用源码解析 Apache ShardingSphere HA & MySQL MGR 实战演练 目录 01 03 Apache ShardingSphere 高可用介绍 ShardingSphere-Proxy 高可用 Apache ShardingSphere Apache ShardingSphere 高可用介绍 Apache ShardingSphere 高可用源码解析 目录 03 02 01 Apache ShardingSphere HA & MySQL MGR 实战演练 Apache ShardingSphere 高可用源码解析 数据库发现模块结构 Apache ShardingSphere 高可用源码解析 ShardingSphere 高可用介绍 Apache ShardingSphere 高可用源码解析 Apache ShardingSphere HA & MySQL MGR 实战演练 目录 01 03 02 Apache ShardingSphere HA & MySQL MGR 实战演练 准备工作 : • ShardingSphere-Proxy 5.1.2 or master branch

56 Unsupported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 4.6 HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 4 245 6.8.2 ReplicaLoadBalanceAlgorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 6.9 HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246 splitting model, primary nodes need to be used for both reading and writing in the transaction. 4.6 HA 4.6.1 Background High availability is the most basic requirement of modern systems. As the cornerstone

iii Unsupported . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 4.6 HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 4 233 6.8.2 ReplicaLoadBalanceAlgorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 6.9 HA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 splitting model, primary nodes need to be used for both reading and writing in the transaction. 4.6 HA 4.6.1 Background High availability is the most basic requirement of modern systems. As the cornerstone

la tc h N O Distributed locks elected by the master nodeUsed for dis‐ tributed locks of curator s ha rd in g: ra w‐ la te x: ne ce ss ar y N O The flag for re‐sharding. If the total number of shards sharding is over or the master node crashes f ai lo ve r: ra w‐ la te x: it em s :r aw ‐l at ex :ˋ s ha rd ˋ it em N O Once a job crashes, it will record to this node.When there is an idle job server mn ty pe R e q u i r e d Describe id V AR CH AR (4 0) Y e s Primary key jo b_n ame VA RC HA R( 10 0) Y e s Job name o rig ina l_t ask _id V AR CH AR (1 00 0) Y e s Original task name

ShardingSphere-Proxy 186 Apache ShardingSphere document, v5.1.0 规则操作 • 创建数据库发现规则 CREATE DB_DISCOVERY RULE ha_group_0 ( RESOURCES(ds_0, ds_1), TYPE(NAME=mgr,PROPERTIES('group-name'='92504d5b-6dec')), HEART HEARTBEAT(PROPERTIES('keep-alive-cron'='0/5 * * * * ?')) ); • 修改数据库发现规则 ALTER DB_DISCOVERY RULE ha_group_0 ( RESOURCES(ds_0, ds_1, ds_2), TYPE(NAME=mgr,PROPERTIES('group-name'='92504d5b-6dec')), HEAR ')) ); • 删除数据库发现规则 DROP DB_DISCOVERY RULE ha_group_0; • 删除数据库发现类型 DROP DB_DISCOVERY TYPE ha_group_0_mgr; • 删除数据库发现心跳 DROP DB_DISCOVERY HEARTBEAT ha_group_0_heartbeat; • 删除数据源 DROP RESOURCE ds_0

CREATE DB_DISCOVERY RULE ha_group_0 ( RESOURCES(resource_0,resource_1), TYPE(NAME=mgr,PROPERTIES(groupName='92504d5b-6dec',keepAliveCron='')) ); ALTER DB_DISCOVERY RULE ha_group_0 ( RESOURCES(resource_0 ); 5.2. ShardingSphere-Proxy 178 Apache ShardingSphere document, v5.0.0 DROP DB_DISCOVERY RULE ha_group_0; Shadow Definition CREATE SHADOW RULE shadowRuleDefinition [, shadowRuleDefinition] ... database for execution s ou rc eD at aS ou rc eN am es List Source data source names s ha do wD at aS ou rc eN am es List Shadow data source names Governance Configuration Item

CREATE DB_DISCOVERY RULE ha_group_0 ( RESOURCES(resource_0,resource_1), TYPE(NAME=mgr,PROPERTIES(groupName='92504d5b-6dec',keepAliveCron='')) ); ALTER DB_DISCOVERY RULE ha_group_0 ( RESOURCES(resource_0 TYPE(NAME=mgr,PROPERTIES(groupName='92504d5b-6dec' ,keepAliveCron='')) ); DROP DB_DISCOVERY RULE ha_group_0; RQL 语法 本章节将对 RQL 的语法进行详细说明，并列举所有查询列的含义。 数据源资源 定义 SHOW RESOURCES [FROM schemaName] 说明

CREATE DB_DISCOVERY RULE ha_group_0 ( RESOURCES(resource_0,resource_1), TYPE(NAME=mgr,PROPERTIES(groupName='92504d5b-6dec',keepAliveCron='')) ); ALTER DB_DISCOVERY RULE ha_group_0 ( RESOURCES(resource_0 TYPE(NAME=mgr,PROPERTIES(groupName='92504d5b-6dec' ,keepAliveCron='')) ); DROP DB_DISCOVERY RULE ha_group_0; 影子库压测 定义 CREATE SHADOW RULE shadowRuleDefinition [, shadowRuleDefinition] ... ALTER SHADOW