dataSources: # Omit the data source configuration, please refer to the usage
rules:
- !SHARDING
tables: # Sharding table configuration
<logic-table-name> (+): # Logic table name
actualDataNodes (?): # Describe data source names and actual tables (refer to Inline syntax rules)
databaseStrategy (?): # Databases sharding strategy, use default databases sharding strategy if absent. sharding strategy below can choose only one.
standard: # For single sharding column scenario
shardingColumn: # Sharding column name
shardingAlgorithmName: # Sharding algorithm name
complex: # For multiple sharding columns scenario
shardingColumns: # Sharding column names, multiple columns separated with comma
shardingAlgorithmName: # Sharding algorithm name
hint: # Sharding by hint
shardingAlgorithmName: # Sharding algorithm name
none: # Do not sharding
tableStrategy: # Tables sharding strategy, same as database sharding strategy
keyGenerateStrategy: # Key generator strategy
column: # Column name of key generator
keyGeneratorName: # Key generator name
autoTables: # Auto Sharding table configuration
t_order_auto: # Logic table name
actualDataSources (?): # Data source names
shardingStrategy: # Sharding strategy
standard: # For single sharding column scenario
shardingColumn: # Sharding column name
shardingAlgorithmName: # Auto sharding algorithm name
bindingTables (+): # Binding tables
- <logic_table_name_1, logic_table_name_2, ...>
- <logic_table_name_1, logic_table_name_2, ...>
broadcastTables (+): # Broadcast tables
- <table-name>
- <table-name>
defaultDatabaseStrategy: # Default strategy for database sharding
defaultTableStrategy: # Default strategy for table sharding
defaultKeyGenerateStrategy: # Default Key generator strategy
# Sharding algorithm configuration
shardingAlgorithms:
<sharding-algorithm-name> (+): # Sharding algorithm name
type: # Sharding algorithm type
props: # Sharding algorithm properties
# ...
# Key generate algorithm configuration
keyGenerators:
<key-generate-algorithm-name> (+): # Key generate algorithm name
type: # Key generate algorithm type
props: # Key generate algorithm properties
# ...
props:
# ...
dataSources: # Omit the data source configuration, please refer to the usage
rules:
- !REPLICA_QUERY
dataSources:
<data-source-name> (+): # Logic data source name of replica query
primaryDataSourceName: # Primary data source name
replicaDataSourceNames:
- <replica-data_source-name> (+) # Replica data source name
loadBalancerName: # Load balance algorithm name
# Load balance algorithm configuration
loadBalancers:
<load-balancer-name> (+): # Load balance algorithm name
type: # Load balance algorithm type
props: # Load balance algorithm properties
# ...
props:
# ...
Please refer to Built-in Load Balance Algorithm List for more details about type of algorithm.
dataSource: # Omit the data source configuration, please refer to the usage
rules:
- !ENCRYPT
tables:
<table-name> (+): # Encrypt table name
columns:
<column-name> (+): # Encrypt logic column name
cipherColumn: # Cipher column name
assistedQueryColumn (?): # Assisted query column name
encryptorName: # Encrypt algorithm name
# Encrypt algorithm configuration
encryptors:
<encrypt-algorithm-name> (+): # Encrypt algorithm name
type: # Encrypt algorithm type
props: # Encrypt algorithm properties
# ...
Please refer to Built-in Encrypt Algorithm List for more details about type of algorithm.
dataSources: # Omit the data source configuration, please refer to the usage
rules:
- !SHADOW
column: # Shadow column name
sourceDataSourceNames: # Source Data Source names
# ...
shadowDataSourceNames: # Shadow Data Source names
# ...
props:
# ...
governance:
name: # Governance name
registryCenter: # Registry center
type: # Governance instance type. Example:Zookeeper, etcd
serverLists: # The list of servers that connect to governance instance, including IP and port number; use commas to separate
overwrite: # Whether to overwrite local configurations with config center configurations; if it can, each initialization should refer to local configurations
dataSources: # Data sources configuration, multiple `data_source_name` available
<data_source_name>: # <!!Data source pool implementation class> `!!` means class instantiation
driverClassName: # Class name of database driver
url: # Database URL
username: # Database username
password: # Database password
# ... Other properties for data source pool
shardingRule:
tables: # Sharding rule configuration, multiple `logic_table_name` available
<logic_table_name>: # Name of logic table
actualDataNodes: # Describe data source names and actual tables, delimiter as point, multiple data nodes separated with comma, support inline expression. Absent means sharding databases only. Example: ds${0..7}.tbl${0..7}
databaseStrategy: # Databases sharding strategy, use default databases sharding strategy if absent. sharding strategy below can choose only one
standard: # Standard sharding scenario for single sharding column
shardingColumn: # Name of sharding column
preciseAlgorithmClassName: # Precise algorithm class name used for `=` and `IN`. This class need to implements PreciseShardingAlgorithm, and require a no argument constructor
rangeAlgorithmClassName: # Range algorithm class name used for `BETWEEN`. This class need to implements RangeShardingAlgorithm, and require a no argument constructor
complex: # Complex sharding scenario for multiple sharding columns
shardingColumns: # Names of sharding columns. Multiple columns separated with comma
algorithmClassName: # Complex sharding algorithm class name. This class need to implements ComplexKeysShardingAlgorithm, and require a no argument constructor
inline: # Inline expression sharding scenario for single sharding column
shardingColumn: # Name of sharding column
algorithmInlineExpression: # Inline expression for sharding algorithm
hint: # Hint sharding strategy
algorithmClassName: # Hint sharding algorithm class name. This class need to implements HintShardingAlgorithm, and require a no argument constructor
none: # Do not sharding
tableStrategy: # Tables sharding strategy, Same as databases sharding strategy
keyGenerator:
column: # Column name of key generator
type: # Type of key generator, use default key generator if absent, and there are three types to choose, that is, SNOWFLAKE/UUID
props: # Properties, Notice: when use SNOWFLAKE, `max.tolerate.time.difference.milliseconds` for `SNOWFLAKE` need to be set. To use the generated value of this algorithm as sharding value, it is recommended to configure `max.vibration.offset`
bindingTables: # Binding table rule configurations
- <logic_table_name1, logic_table_name2, ...>
- <logic_table_name3, logic_table_name4, ...>
- <logic_table_name_x, logic_table_name_y, ...>
broadcastTables: # Broadcast table rule configurations
- table_name1
- table_name2
- table_name_x
defaultDataSourceName: # If table not configure at table rule, will route to defaultDataSourceName
defaultDatabaseStrategy: # Default strategy for sharding databases, same as databases sharding strategy
defaultTableStrategy: # Default strategy for sharding tables, same as tables sharding strategy
defaultKeyGenerator:
type: # Type of default key generator, use user-defined ones or built-in ones, e.g. SNOWFLAKE, UUID. Default key generator is `org.apache.shardingsphere.core.keygen.generator.impl.SnowflakeKeyGenerator`
column: # Column name of default key generator
props: # Properties of default key generator, e.g. `max.tolerate.time.difference.milliseconds` for `SNOWFLAKE`
masterSlaveRules: # Read-write splitting rule configuration, more details can reference Read-write splitting part
<data_source_name>: # Data sources configuration, need consist with data source map, multiple `data_source_name` available
masterDataSourceName: # more details can reference Read-write splitting part
slaveDataSourceNames: # more details can reference Read-write splitting part
loadBalanceAlgorithmType: # more details can reference Read-write splitting part
props: # Properties configuration of load balance algorithm
<property-name>: # property key value pair
props: # Properties
sql.show: # To show SQLS or not, default value: false
executor.size: # The number of working threads, default value: CPU count
check.table.metadata.enabled: # To check the metadata consistency of all the tables or not, default value : false
max.connections.size.per.query: # The maximum connection number allocated by each query of each physical database. default value: 1
dataSources: # Omit data source configurations; keep it consistent with data sharding
masterSlaveRule:
name: # Read-write split data source name
masterDataSourceName: # Master data source name
slaveDataSourceNames: # Slave data source name
- <data_source_name1>
- <data_source_name2>
- <data_source_name_x>
loadBalanceAlgorithmType: # Slave database load balance algorithm type; optional value, ROUND_ROBIN and RANDOM, can be omitted if `loadBalanceAlgorithmClassName` exists
props: # Properties configuration of load balance algorithm
<property-name>: # property key value pair
props: # Property configuration
sql.show: # Show SQL or not; default value: false
executor.size: # Executing thread number; default value: CPU core number
check.table.metadata.enabled: # Whether to check table metadata consistency when it initializes; default value: false
max.connections.size.per.query: # The maximum connection number allocated by each query of each physical database. default value: 1
Create a DataSource through the YamlMasterSlaveDataSourceFactory factory class:
DataSource dataSource = YamlMasterSlaveDataSourceFactory.createDataSource(yamlFile);
dataSource: # Ignore data sources configuration
encryptRule:
encryptors:
<encryptor-name>:
type: # Encryptor type
props: # Properties, e.g. `aes.key.value` for AES encryptor
aes.key.value:
tables:
<table-name>:
columns:
<logic-column-name>:
cipherColumn: # Ciphertext column name
assistedQueryColumn: # AssistedColumns for query，when use ShardingQueryAssistedEncryptor, it can help query encrypted data
encryptor: # Encrypt name
dataSources: # Omit data source configurations
shardingRule: # Omit sharding rule configurations
masterSlaveRule: # Omit read-write split rule configurations
encryptRule: # Omit encrypt rule configurations
orchestration:
name: # Orchestration instance name
overwrite: # Whether to overwrite local configurations with registry center configurations; if it can, each initialization should refer to local configurations
registry: # Registry center configuration
type: # Registry center type. Example:zookeeper
serverLists: # The list of servers that connect to registry center, including IP and port number; use commas to seperate addresses, such as: host1:2181,host2:2181
namespace: # Registry center namespace
digest: # The token that connects to the registry center; default means there is no need for authentication
operationTimeoutMilliseconds: # Default value: 500 milliseconds
maxRetries: # Maximum retry time after failing; default value: 3 times
retryIntervalMilliseconds: # Interval time to retry; default value: 500 milliseconds
timeToLiveSeconds: # Living time of temporary nodes; default value: 60 seconds
dataSources: # Data sources configuration, multiple `data_source_name` available
<data_source_name>: # <!!Data source pool implementation class> `!!` means class instantiation
driverClassName: # Class name of database driver
url: # Database URL
username: # Database username
password: # Database password
# ... Other properties for data source pool
shardingRule:
tables: # Sharding rule configuration, multiple `logic_table_name` available
<logic_table_name>: # Name of logic table
actualDataNodes: # Describe data source names and actual tables, delimiter as point, multiple data nodes separated with comma, support inline expression. Absent means sharding databases only. Example: ds${0..7}.tbl${0..7}
databaseStrategy: # Databases sharding strategy, use default databases sharding strategy if absent. sharding strategy below can choose only one
standard: # Standard sharding scenario for single sharding column
shardingColumn: # Name of sharding column
preciseAlgorithmClassName: # Precise algorithm class name used for `=` and `IN`. This class need to implements PreciseShardingAlgorithm, and require a no argument constructor
rangeAlgorithmClassName: # Range algorithm class name used for `BETWEEN`. This class need to implements RangeShardingAlgorithm, and require a no argument constructor
complex: # Complex sharding scenario for multiple sharding columns
shardingColumns: # Names of sharding columns. Multiple columns separated with comma
algorithmClassName: # Complex sharding algorithm class name. This class need to implements ComplexKeysShardingAlgorithm, and require a no argument constructor
inline: # Inline expression sharding scenario for single sharding column
shardingColumn: # Name of sharding column
algorithmInlineExpression: # Inline expression for sharding algorithm
hint: # Hint sharding strategy
algorithmClassName: # Hint sharding algorithm class name. This class need to implements HintShardingAlgorithm, and require a no argument constructor
none: # Do not sharding
tableStrategy: # Tables sharding strategy, Same as databases sharding strategy
keyGeneratorColumnName: # Column name of key generator, do not use Key generator if absent
keyGeneratorClassName: # Key generator, use default key generator if absent. This class need to implements KeyGenerator, and require a no argument constructor
logicIndex: # Name if logic index. If use `DROP INDEX XXX` SQL in Oracle/PostgreSQL, This property needs to be set for finding the actual tables
bindingTables: # Binding table rule configurations
- <logic_table_name1, logic_table_name2, ...>
- <logic_table_name3, logic_table_name4, ...>
- <logic_table_name_x, logic_table_name_y, ...>
bindingTables: # Broadcast table rule configurations
- table_name1
- table_name2
- table_name_x
defaultDataSourceName: # If table not configure at table rule, will route to defaultDataSourceName
defaultDatabaseStrategy: # Default strategy for sharding databases, same as databases sharding strategy
defaultTableStrategy: # Default strategy for sharding tables, same as tables sharding strategy
defaultKeyGeneratorClassName: # Default key generator class name, default value is `io.shardingsphere.core.keygen.DefaultKeyGenerator`. This class need to implements KeyGenerator, and require a no argument constructor
masterSlaveRules: # Read-write splitting rule configuration, more details can reference Read-write splitting part
<data_source_name>: # Data sources configuration, need consist with data source map, multiple `data_source_name` available
masterDataSourceName: # more details can reference Read-write splitting part
slaveDataSourceNames: # more details can reference Read-write splitting part
loadBalanceAlgorithmType: # more details can reference Read-write splitting part
loadBalanceAlgorithmClassName: # more details can reference Read-write splitting part
configMap: # User-defined arguments
key1: value1
key2: value2
keyx: valuex
props: # Properties
sql.show: # To show SQLS or not, default value: false
executor.size: # The number of working threads, default value: CPU count
check.table.metadata.enabled: #T o check the metadata consistency of all the tables or not, default value : false
configMap: # User-defined arguments
key1: value1
key2: value2
keyx: valuex
dataSources: # Ignore data sources configuration, same as sharding
masterSlaveRule:
name: # Name of master slave data source
masterDataSourceName: # Name of master data source
slaveDataSourceNames: # Names of Slave data sources
- <data_source_name1>
- <data_source_name2>
- <data_source_name_x>
loadBalanceAlgorithmClassName: # Load balance algorithm class name. This class need to implements MasterSlaveLoadBalanceAlgorithm, and require a no argument constructor
loadBalanceAlgorithmType: # Load balance algorithm type, values should be: `ROUND_ROBIN` or `RANDOM`. Ignore if `loadBalanceAlgorithmClassName` is present
props: # Properties
sql.show: # To show SQLS or not, default value: false
executor.size: # The number of working threads, default value: CPU count
check.table.metadata.enabled: # To check the metadata consistency of all the tables or not, default value : false
configMap: # User-defined arguments
key1: value1
key2: value2
keyx: valuex
Create a DataSource through the YamlMasterSlaveDataSourceFactory factory class:
DataSource dataSource = MasterSlaveDataSourceFactory.createDataSource(yamlFile);
dataSources: # Ignore data sources configuration
shardingRule: # Ignore sharding rule configuration
masterSlaveRule: # Ignore master slave rule configuration
orchestration:
name: # Name of orchestration instance
overwrite: # Use local configuration to overwrite registry center or not
registry: # Registry configuration
serverLists: # Registry servers list, multiple split as comma. Example: host1:2181,host2:2181
namespace: # Namespace of registry
digest: # Digest for registry. Default is not need digest.
operationTimeoutMilliseconds: # Operation timeout time in milliseconds, default value is 500 milliseconds
maxRetries: # Max number of times to retry, default value is 3
retryIntervalMilliseconds: # Time interval in milliseconds on each retry, default value is 500 milliseconds
timeToLiveSeconds: # Time to live in seconds of ephemeral keys, default value is 60 seconds
dataSources:
db0: !!org.apache.commons.dbcp.BasicDataSource
driverClassName: org.h2.Driver
url: jdbc:h2:mem:db0;DB_CLOSE_DELAY=-1;DATABASE_TO_UPPER=false;MODE=MYSQL
username: sa
password:
maxActive: 100
db1: !!org.apache.commons.dbcp.BasicDataSource
driverClassName: org.h2.Driver
url: jdbc:h2:mem:db1;DB_CLOSE_DELAY=-1;DATABASE_TO_UPPER=false;MODE=MYSQL
username: sa
password:
maxActive: 100
shardingRule:
tables:
config:
actualDataNodes: db${0..1}.t_config
t_order:
actualDataNodes: db${0..1}.t_order_${0..1}
databaseStrategy:
standard:
shardingColumn: user_id
preciseAlgorithmClassName: io.shardingjdbc.core.yaml.fixture.SingleAlgorithm
tableStrategy:
inline:
shardingColumn: order_id
algorithmInlineExpression: t_order_${order_id % 2}
keyGeneratorColumnName: order_id
keyGeneratorClass: io.shardingjdbc.core.yaml.fixture.IncrementKeyGenerator
t_order_item:
actualDataNodes: db${0..1}.t_order_item_${0..1}
# The strategy of binding the rest of the tables in the table is the same as the strategy of the first table
databaseStrategy:
standard:
shardingColumn: user_id
preciseAlgorithmClassName: io.shardingjdbc.core.yaml.fixture.SingleAlgorithm
tableStrategy:
inline:
shardingColumn: order_id
algorithmInlineExpression: t_order_item_${order_id % 2}
bindingTables:
- t_order,t_order_item
# Default database sharding strategy
defaultDatabaseStrategy:
none:
defaultTableStrategy:
complex:
shardingColumns: id, order_id
algorithmClassName: io.shardingjdbc.core.yaml.fixture.MultiAlgorithm
props:
sql.show: true
In order to relieve the pressure on the database, the write and read operations are separated into different data sources. The write library is called the master library, and the read library is called the slave library. One master library can be configured with multiple slave libraries.
dataSources:
db_master: !!org.apache.commons.dbcp.BasicDataSource
driverClassName: org.h2.Driver
url: jdbc:h2:mem:db_master;DB_CLOSE_DELAY=-1;DATABASE_TO_UPPER=false;MODE=MYSQL
username: sa
password:
maxActive: 100
db_slave_0: !!org.apache.commons.dbcp.BasicDataSource
driverClassName: org.h2.Driver
url: jdbc:h2:mem:db_slave_0;DB_CLOSE_DELAY=-1;DATABASE_TO_UPPER=false;MODE=MYSQL
username: sa
password:
maxActive: 100
db_slave_1: !!org.apache.commons.dbcp.BasicDataSource
driverClassName: org.h2.Driver
url: jdbc:h2:mem:db_slave_1;DB_CLOSE_DELAY=-1;DATABASE_TO_UPPER=false;MODE=MYSQL
username: sa
password:
maxActive: 100
masterSlaveRule:
name: db_ms
masterDataSourceName: db_master
slaveDataSourceNames: [db_slave_0, db_slave_1]
configMap:
key1: value1
Create a DataSource through the MasterSlaveDataSourceFactory factory class:
DataSource dataSource = MasterSlaveDataSourceFactory.createDataSource(yamlFile);
Zookeeper sharding table and database Orchestration Configuration Item Explanation
dataSources: Data sources configuration
shardingRule: Sharding rule configuration
orchestration: Zookeeper Orchestration Configuration
name: Orchestration name
overwrite: Whether to overwrite local configurations with config center configurations; if it can, each initialization should refer to local configurations
zookeeper: Registry center Configuration
namespace: Registry center namespace
serverLists: The list of servers that connect to governance instance, including IP and port number, use commas to separate, such as: host1:2181,host2:2181
baseSleepTimeMilliseconds: The initial millisecond value of the interval to wait for retry
maxSleepTimeMilliseconds: The maximum millisecond value of the interval to wait for retry
maxRetries: The maximum retry count
sessionTimeoutMilliseconds: The session timeout milliseconds
connectionTimeoutMilliseconds: The connecton timeout milliseconds
digest: Permission token to connect to Zookeeper. default no authorization is required
Etcd sharding table and database Orchestration Configuration Item Explanation
dataSources: Data sources configuration
shardingRule: Sharding rule configuration
orchestration: Etcd Orchestration Configuration
name: Orchestration name
overwrite: Whether to overwrite local configurations with config center configurations; if it can, each initialization should refer to local configurations
etcd: Registry center Configuration
serverLists: The list of servers that connect to governance instance, including IP and port number, use commas to separate, such as: http://host1:2379,http://host2:2379
timeToLiveSeconds: Time to live seconds for ephemeral nodes
timeoutMilliseconds: The request timeout milliseconds
maxRetries: The maximum retry count
retryIntervalMilliseconds: The retry interval milliseconds
Sharding table and database Data source construction method
DataSource dataSource = OrchestrationShardingDataSourceFactory.createDataSource(yamlFile);
Read-Write split Data source construction method
DataSource dataSource = OrchestrationMasterSlaveDataSourceFactory.createDataSource(yamlFile);