Synchronize data from PolarDB for PostgreSQL to SelectDB - Data Transmission Service

This topic describes how to use Data Transmission Service (DTS) to synchronize data from a PolarDB for PostgreSQL cluster to a SelectDB instance. This process is useful for large-scale data analytics.

Prerequisites

You have created a destination SelectDB instance. The destination instance must have more disk space than the amount used by the source PolarDB for PostgreSQL cluster. For more information, see Create an instance.
You have set the wal_level parameter of the source PolarDB for PostgreSQL cluster to logical. For more information, see Set cluster parameters.

Notes

Type	Description
Source database limits	Synchronization objects: All tables to be synchronized must have a primary key or a non-null unique index. Ensure that the table fields are unique. Otherwise, duplicate data may appear in the destination database. Tables without a primary key or a non-null unique index: We recommend that you select Schema Synchronization for the Synchronization Types parameter and duplicate for the Engine parameter in the Configurations for Databases, Tables, and Columns step when you configure an instance. Note During schema synchronization, DTS adds fields to the destination table. For more information, see Additional column information. If the source database has one or more long-running transactions and incremental data is synchronized in the data synchronization task, the write-ahead logging (WAL) logs generated before the long-running transactions in the source database are committed may be accumulated. As a result, the disk space of the source database may be insufficient. To ensure that the synchronization task runs as expected and to prevent logical replication interruptions caused by a primary/secondary switchover, the PolarDB for PostgreSQL cluster must support and enable Logical Replication Slot Failover. Note If the source PolarDB for PostgreSQL cluster does not support Logical Replication Slot Failover (for example, if the Database Engine of the cluster is PostgreSQL 14), an HA switchover in the source database may cause the synchronization instance to fail and become unrecoverable. Due to the limits of logical replication in the source database, if a single piece of data to be synchronized exceeds 256 MB after an incremental change, the synchronization instance may fail and cannot be recovered. You must reconfigure the synchronization instance. During initial schema synchronization and initial full data synchronization, do not perform Data Definition Language (DDL) operations that change the schema of databases or tables. Otherwise, the data synchronization task fails. Note During initial full data synchronization, DTS queries the source database. This creates metadata locks, which may block DDL operations on the source database.
Other limits	You can synchronize data to only tables that use the Unique Key model or Duplicate key model in ApsaraDB for SelectDB instances. Unique key model If you synchronize data to the destination table that uses the Unique key model, make sure all unique keys of the destination table already are contained in the source table and the object to be synchronized. Otherwise, this may cause data inconsistency. Duplicate key model If you synchronize data to the destination table that uses the Duplicate key model, and if one of the following operations occurs, duplicate data may exist in the destination database. In this case, you can deduplicate the data based on additional columns such as _is_deleted, _version, and _record_id. Retry operation occurred in a data synchronization instance. Retry operation was executed in a data synchronization instance. Two or more DML operations were executed for the same row of data after a data synchronization instance started. Note When the destination table is the table that uses the Duplicate key model, DTS converts UPDATE or DELETE statements to INSERT statements. You can specify only the `bucket_count` parameter in the Selected Objects section. Note The `bucket_count` parameter value must be a positive integer. Default value: auto. ApsaraDB for SelectDB instances support only database names and table names that start with letters. If the name of a database or table that you want to synchronize does not start with a letter, you must use the object name mapping feature to rename the database or table. If the name of the object that you want to synchronize, such as a database, a table, or a column, contains Chinese characters, you must use the object name mapping feature to rename the object. For example, you can change the name from Chinese to English. Otherwise, the task may fail. You cannot modify DDL operations on multiple columns at a time or modify DDL operations on a table consecutively. A single synchronization instance can synchronize only one database. To synchronize multiple databases, you must configure a separate synchronization instance for each database. Synchronization of TimescaleDB extension tables and tables with cross-schema inheritance is not supported. In the following three scenarios, you must run the `ALTER TABLE schema.table REPLICA IDENTITY FULL;` command on the tables to be synchronized before writing data to them. This ensures data consistency. During the execution of this command, do not perform table lock operations. Otherwise, the tables may be locked. If you skip the related check in the precheck, DTS automatically runs this command during the initialization of the instance. When the instance runs for the first time. When you select objects to synchronize at the schema level, and a new table is created in the schema or a table to be synchronized is rebuilt using the RENAME command. When you use the feature to modify synchronization objects. Note In the command, replace `schema` and `table` with the schema name and table name of the data to be synchronized. Perform this operation during off-peak hours. During initial full data synchronization, DTS consumes some read and write resources from the source and destination databases. This may increase the database load. Before you start data synchronization, evaluate the performance of the source and destination databases. Perform data synchronization during off-peak hours (for example, when the CPU load of both databases is below 30%). During data synchronization, do not add backend nodes to the ApsaraDB for SelectDB database. Otherwise, the task fails. You can restart the data synchronization instance to resume the data synchronization task that failed. In a multi-table merge scenario (synchronizing data from multiple source tables to a single destination table), make sure the schemas of the source tables are identical. Otherwise, data inconsistency or task failure may occur. During data synchronization, do not create clusters in the destination ApsaraDB for SelectDB instance. Otherwise, the data synchronization task fails. You can restart the data synchronization instance to resume the data synchronization task that failed. DTS does not check the validity of metadata, such as sequences. You must manually check the validity of metadata. DTS creates the following temporary tables in the source database to obtain information such as DDL statements for incremental data, the schema of incremental tables, and heartbeat data. Do not delete these temporary tables during synchronization. Otherwise, the DTS task may become abnormal. The temporary tables are automatically deleted after the DTS instance is released. `public.dts_pg_class`, `public.dts_pg_attribute`, `public.dts_pg_type`, `public.dts_pg_enum`, `public.dts_postgres_heartbeat`, `public.dts_ddl_command`, `public.dts_args_session`, and `public.aliyun_dts_instance`. During data synchronization, DTS creates a replication slot with the prefix `dts_sync_` in the source database to replicate data. Using this replication slot, DTS can obtain incremental logs from the source database within the last 15 minutes. When the data synchronization fails or the instance is released, DTS attempts to clean up the replication slot. Note If you change the password of the source database account used by the task or delete the DTS IP address whitelist from the source database during synchronization, the replication slot cannot be automatically cleaned up. In this case, you must manually clean up the replication slot in the source database to prevent it from accumulating and occupying disk space, which can make the source database unavailable. If a primary/secondary failover occurs in the source database, you must log on to the secondary database to perform the cleanup. When synchronizing partitioned tables, include both the parent table and its child partitions as synchronization objects. Otherwise, data in the partitioned table may become inconsistent. Note The parent table of a PostgreSQL partitioned table does not store data directly. All data is stored in the child partitions. The synchronization task must include the parent table and all its child partitions. Otherwise, data from the child partitions may be missed, leading to data inconsistency between the source and destination. During incremental synchronization, DTS uses a batch synchronization policy to reduce the load on the destination. By default, for a single synchronization object, DTS writes data at most once every 5 seconds. This may cause a normal synchronization latency, usually within 10 seconds. To reduce this latency, modify the `selectdb.reservoir.timeout.milliseconds` parameter of the DTS instance in the console. This adjusts the batching time. The allowed range is 1,000 to 10,000 milliseconds. Note A lower batching time increases the write frequency of DTS. This may increase the load and write response time (RT) of the destination, which in turn increases the DTS synchronization latency. Adjust the batching time based on the load of the destination. If an instance fails, DTS helpdesk will try to recover the instance within 8 hours. During the recovery process, operations such as restarting the instance or adjusting its parameters may be performed. Note When parameters are adjusted, only the parameters of the DTS instance are modified. The parameters in the database are not modified. The parameters that may be modified include but are not limited to those described in Modify instance parameters.

Billing

Synchronization type	Task configuration fee
Schema synchronization and full data synchronization	Free of charge.
Incremental data synchronization	Charged. For more information, see Billing overview.

Supported SQL operations

Operation type	SQL statement
DML	INSERT, UPDATE, DELETE
DDL	ADD COLUMN, DROP COLUMN

Permissions for database accounts

Database	Required permissions	How to create an account and grant permissions
Source PolarDB for PostgreSQL cluster	A privileged account that owns the database to be synchronized.	Create a database account and Database Management.
Destination SelectDB instance	Cluster access permissions (Usage_priv) and database access permissions (Select_priv, Load_priv, Alter_priv, Create_priv, and Drop_priv).	Cluster Permission Management and Basic Permission Management.

Procedure

Use one of the following methods to go to the Data Synchronization page and select the region in which the data synchronization instance resides.
DTS console
1. Log on to the DTS console.
2. In the left-side navigation pane, click Data Synchronization.
3. In the upper-left corner of the page, select the region in which the data synchronization task resides.
DMS console
Note
The actual operations may vary based on the mode and layout of the DMS console. For more information, see Simple mode and Customize the layout and style of the DMS console.
1. Log on to the DMS console.
2. In the top navigation bar, move the pointer over Data + AI and choose DTS (DTS) > Data Synchronization.
3. From the drop-down list to the right of Data Synchronization Tasks, select the region in which the data synchronization instance resides.
Click Create Task to go to the task configuration page.

Configure the source and destination databases. The following table describes the parameters.

Category	Configuration	Description
None	Task Name	The name of the DTS task. DTS automatically generates a task name. We recommend that you specify a descriptive name that makes it easy to identify the task. You do not need to specify a unique task name.
Source Database	Select Existing Connection	If you use a database instance that is registered with DTS, select the instance from the drop-down list. DTS automatically populates the following database parameters for the instance. For more information, see Manage database connections. Note In the DMS console, you can select the database instance from the Select a DMS database instance drop-down list. If you fail to register the instance with DTS, or you do not need to use the instance that is registered with DTS, you must configure the following database information.
	Database Type	Select PolarDB for PostgreSQL.
	Access Method	Select Alibaba Cloud Instance.
	Instance Region	Select the region where the source PolarDB for PostgreSQL cluster resides.
	Replicate Data Across Alibaba Cloud Accounts	In this example, a database of the current Alibaba Cloud account is used. Select No.
	Instance ID	Select the ID of the source PolarDB for PostgreSQL cluster.
	Database Name	Enter the name of the database that contains the objects to be synchronized in the source PolarDB for PostgreSQL cluster.
	Database Account	Enter the database account of the source PolarDB for PostgreSQL cluster. For permission requirements, see Permissions for database accounts.
	Database Password	The password that is used to access the database.
Destination Database	Select Existing Connection	If you use a database instance that is registered with DTS, select the instance from the drop-down list. DTS automatically populates the following database parameters for the instance. For more information, see Manage database connections. Note In the DMS console, you can select the database instance from the Select a DMS database instance drop-down list. If you fail to register the instance with DTS, or you do not need to use the instance that is registered with DTS, you must configure the following database information.
	Database Type	Select SelectDB.
	Access Method	Select Alibaba Cloud Instance.
	Instance Region	Select the region where the destination SelectDB instance resides.
	Replicate Data Across Alibaba Cloud Accounts	In this example, a database of the current Alibaba Cloud account is used. Select No.
	Instance ID	Select the ID of the destination SelectDB instance.
	Database Account	Enter the database account of the destination SelectDB instance. For permission requirements, see Permissions for database accounts.
	Database Password	The password that is used to access the database.

In the lower part of the page, click Test Connectivity and Proceed.
Note
Make sure that the CIDR blocks of DTS servers can be automatically or manually added to the security settings of the source and destination databases to allow access from DTS servers. For more information, see Add DTS server IP addresses to a whitelist.

Configure the objects to be synchronized.

In the Configure Objects step, configure the objects that you want to synchronize.

Configuration	Description
Synchronization Types	The synchronization types. By default, Incremental Data Synchronization is selected. You must also select Schema Synchronization and Full Data Synchronization. After the precheck is complete, DTS synchronizes the historical data of the selected objects from the source database to the destination cluster. The historical data is the basis for subsequent incremental synchronization. Important When data is synchronized from a PolarDB for PostgreSQL cluster to SelectDB, data types are converted. If you do not select Schema Synchronization, you must create tables in the destination SelectDB instance using the corresponding Unique or Duplicate model. For more information, see Data type mapping, Additional column information, and Data model.
Processing Mode of Conflicting Tables	Precheck and Report Errors: DTS checks whether a table with the same name exists in the destination database. If a table with the same name does not exist, the precheck is successful. If a table with the same name exists, the precheck fails and the data synchronization task does not start. Note If you cannot delete or rename the table with the same name in the destination database, you can use the object name mapping feature. For more information, see Map schema, table, and column names. Ignore Errors and Proceed: DTS skips the check for tables with the same name in the destination database. Warning Selecting Ignore Errors and Proceed may cause data inconsistency and pose risks to your business. For example: If the table schemas are the same and a record in the destination database has the same primary key or unique key value as a record in the source database, the record from the source database overwrites the record in the destination database. If the table schemas are different, data initialization may fail, only some columns may be synchronized, or the entire synchronization task may fail. Use this option with caution.
Capitalization of Object Names in Destination Instance	The capitalization of database names, table names, and column names in the destination instance. By default, DTS default policy is selected. You can select other options to ensure that the capitalization of object names is consistent with that in the source or destination database. For more information, see Specify the capitalization of object names in the destination instance.
Source Objects	Select one or more objects from the Source Objects section and click the icon to add the objects to the Selected Objects section. Note You can select objects at the schema, table, and column levels.
Selected Objects	To configure the name of an object to be synchronized in the destination database or specify an object that receives data in the destination database, right-click the object in the Selected Objects section. For more information, see Map object names. To remove a selected object, click the object in the Selected Objects section and then click the icon to move the object to the Source Objects section. If you selected Schema Synchronization for Synchronization Types, selected objects at the table level, and need to set the number of buckets (the `bucket_count` parameter), right-click the table in the Selected Objects box. In the Parameter Settings area, set Enable Parameter Settings to Yes. Set the Value as required, and then click OK. Note If you use the object name mapping feature to rename an object, other objects that are dependent on the object may fail to be synchronized. To specify WHERE conditions to filter data, right-click a table in the Selected Objects section. In the dialog box that appears, specify the conditions. For more information, see Specify filter conditions. To select SQL operations for incremental synchronization, right-click an object in the Selected Objects section. In the dialog box that appears, select the SQL operations that you want to synchronize.

Click Next: Advanced Settings to configure advanced settings.

Configuration	Description
Dedicated Cluster for Task Scheduling	By default, DTS schedules the task to the shared cluster if you do not specify a dedicated cluster. If you want to improve the stability of data synchronization instances, purchase a dedicated cluster. For more information, see What is a DTS dedicated cluster.
Retry Time for Failed Connections	The retry time range for failed connections. If the source or destination database fails to be connected after the data synchronization task is started, DTS immediately retries a connection within the time range. Valid values: 10 to 1440. Unit: minutes. Default value: 720. We recommend that you set this parameter to a value greater than 30. If DTS reconnects to the source and destination databases within the specified time range, DTS resumes the data synchronization task. Otherwise, the data synchronization task fails. Note If you specify different retry time ranges for multiple data synchronization tasks that have the same source or destination database, the shortest retry time range takes precedence. When DTS retries a connection, you are charged for the DTS instance. We recommend that you specify the retry time range based on your business requirements. You can also release the DTS instance at your earliest opportunity after the source and destination instances are released.
Retry Time for Other Issues	The retry time range for other issues. For example, if the DDL or DML operations fail to be performed after the data synchronization task is started, DTS immediately retries the operations within the time range. Valid values: 1 to 1440. Unit: minutes. Default value: 10. We recommend that you set this parameter to a value greater than 10. If the failed operations are successfully performed within the specified time range, DTS resumes the data synchronization task. Otherwise, the data synchronization task fails. Important The value of the Retry Time for Other Issues parameter must be smaller than the value of the Retry Time for Failed Connections parameter.
Enable Throttling for Full Data Synchronization	During full data synchronization, DTS uses the read and write resources of the source and destination databases. This may increase the load on the database servers. You can configure the Queries per second (QPS) to the source database, RPS of Full Data Migration, and Data migration speed for full migration (MB/s) parameters for full data synchronization tasks to reduce the load on the destination database server. Note You can configure this parameter only if Full Data Synchronization is selected for the Synchronization Types parameter.
Enable Throttling for Incremental Data Synchronization	Specifies whether to enable throttling for incremental data synchronization. You can enable throttling for incremental data synchronization based on your business requirements. To configure throttling, you must configure the RPS of Incremental Data Synchronization and Data synchronization speed for incremental synchronization (MB/s) parameters. This reduces the load on the destination database server.
Environment Tag	As needed, you can select an environment tag to identify the instance. No selection is needed for this example.
Configure ETL	Specifies whether to enable the extract, transform, and load (ETL) feature. For more information, see What is ETL? Valid values: Yes: configures the ETL feature. You can enter data processing statements in the code editor. For more information, see Configure ETL in a data migration or data synchronization task. No: does not configure the ETL feature.
Monitoring and Alerting	Specifies whether to configure alerting for the data synchronization instance. If the task fails or the synchronization latency exceeds the specified threshold, alert contacts will receive notifications. Valid values: No: does not enable alerting. Yes: configures alerting. In this case, you must also configure the alert threshold and alert notification settings. For more information, see the "Configure monitoring and alerting when you create a DTS task" section of the Configure monitoring and alerting topic.

Optional: Click Next: Configure Database and Table Fields to set the Primary Key Column, Distribution Key, and Engine for the tables in the destination database.
Note
- This step is available only if you select Schema Synchronization for Synchronization Types. You can set Definition Status to All to make modifications.
- You can select multiple columns as a composite primary key in the Primary Key Column field. You must select one or more columns from the Primary Key Column to serve as the Distribution Key.
- For tables that have neither a primary key nor a UNIQUE constraint, you must set Engine to duplicate. Otherwise, the synchronization may fail or data may be lost.

Save the task settings and run a precheck.
- To view the parameters to be specified when you call the relevant API operation to configure the DTS task, move the pointer over Next: Save Task Settings and Precheck and click Preview OpenAPI parameters.
- If you do not need to view or have viewed the parameters, click Next: Save Task Settings and Precheck in the lower part of the page.
Note
- Before you can start the data synchronization task, DTS performs a precheck. You can start the data synchronization task only after the task passes the precheck.
- If the data synchronization task fails the precheck, click View Details next to each failed item. After you analyze the causes based on the check results, troubleshoot the issues. Then, rerun the precheck.
- If an alert is triggered for an item during the precheck:
  If an alert item cannot be ignored, click View Details next to the failed item and troubleshoot the issue. Then, run a precheck again.
  If an alert item can be ignored, click Confirm Alert Details. In the View Details dialog box, click Ignore. In the message that appears, click OK. Then, click Precheck Again to run a precheck again. If you ignore the alert item, data inconsistency may occur, and your business may be exposed to potential risks.

Purchase the instance.

Wait until the Success Rate becomes 100%. Then, click Next: Purchase Instance.

On the buy page, configure the Billing Method and Instance Class parameters for the data synchronization task. The following table describes the parameters.

Section	Parameter	Description
New Instance Class	Billing Method	Subscription: You pay for a subscription when you create a data synchronization instance. The subscription billing method is more cost-effective than the pay-as-you-go billing method for long-term use. Pay-as-you-go: A pay-as-you-go instance is billed on an hourly basis. The pay-as-you-go billing method is suitable for short-term use. If you no longer require a pay-as-you-go data synchronization instance, you can release the instance to reduce costs.
	Resource Group Settings	The resource group to which the data synchronization instance belongs. Default value: default resource group. For more information, see What is Resource Management?
	Instance Class	DTS provides instance classes that vary in synchronization speed. You can select an instance class based on your business requirements. For more information, see Instance classes of data synchronization instances.
	Subscription Duration	If you select the subscription billing method, specify the subscription duration and the number of data synchronization instances that you want to create. The subscription duration can be one to nine months, one year, two years, three years, or five years. Note This parameter is available only if you select the Subscription billing method.

Read and select Data Transmission Service (Pay-as-you-go) Service Terms.
Click Buy and Start. In the dialog box that appears, click OK.
You can view the progress of the task in the task list.

Data type mapping

Category	PolarDB for PostgreSQL cluster data type	SelectDB instance data type
NUMERIC	SMALLINT	SMALLINT
	INTEGER	INT
	BIGINT	BIGINT
	DECIMAL	DECIMAL
	NUMERIC	DECIMAL
	REAL	DOUBLE
	DOUBLE	DOUBLE
	SMALLSERIAL	SMALLINT
	SERIAL	INT
	BIGSERIAL	BIGINT
MONETARY	MONEY	STRING
CHARACTER	CHAR(n) VARCHAR(n)	VARCHAR Important To prevent data loss, data of the CHAR(n) and VARCHAR(n) types is converted to VARCHAR(4*n) after it is synchronized to the SelectDB instance. If the data length is not specified, the SelectDB default value VARCHAR(65533) is used. If the data length exceeds 65533, the data is converted to STRING after being synchronized to SelectDB.
CHARACTER	TEXT	STRING
BINARY	BYTEA	STRING
DATE AND TIME	TIMESTAMP [(P)] [WITHOUT TIME ZONE]	DATETIMEV2
	TIMESTAMP [(P)] WITH TIME ZONE	DATETIMEV2
	DATE	DATEV2
	TIME [(P)] [WITHOUT TIME ZONE]	VARCHAR(50)
	TIME [(P)] WITH TIME ZONE	VARCHAR(50)
	INTERVAL [FIELDS] [(P)]	STRING
BOOLEAN	BOOLEAN	BOOLEAN
GEOMETRIC	POINT LINE LSEG BOX PATH POLYGON CIRCLE	STRING
NETWORK ADDRESS	CIDR INET MACADDR MACADDR8	STRING
TEXT SEARCH	TSVECTOR	STRING
XML	XML	STRING
JSON	JSON	JSON

Additional column information

Note

The following table describes the additional columns that DTS automatically adds or that you must add to the destination Duplicate model table.

Name	Data type	Default value	Description
_is_deleted	Int	0	Indicates whether the data is deleted. Insert: The value is 0. Update: The value is 0. Delete: The value is 1.
_version	Bigint	0	Full data synchronization: The value is 0. Incremental data synchronization: The corresponding timestamp (in seconds) from the source database's binary log.
_record_id	Bigint	0	Full data synchronization: The value is 0. Incremental data synchronization: The record ID from the incremental log, which is the unique identifier for that log entry. Note The ID value is unique and increments.

Prerequisites

Notes

Unique key model

Duplicate key model