The data migration mode migrates data between data stores. This mode is typically used for one-time migrations that require minimized downtimes. The data migration mode also provides transform, and load (ETL) features such as object name mapping and data filtering.
Data migration phases
A data migration task consists of several phases, namely schema migration, full data migration, and incremental data migration. You can select the required phases when creating a data migration task. However, to conduct a minimized-downtime migration, you must select all the three phases.
In the schema migration phase, Data Transmission Service (DTS) migrates schema objects from the source database to the target database, including tables, views, triggers, and stored procedures.
To migrate schema objects across heterogeneous databases, DTS converts schema objects into data types that work with the target database. For example, it converts the NUMBER data type in Oracle databases to the DECIMAL data type in MySQL databases.
|Full data migration||In the full data migration phase, DTS replicates the existing data from the source
database to the target database. If you select only schema migration and full data
migration for the task, updates that occur during the full data migration phase will
not be captured and replicated to the target database.
Note To ensure data consistency, we recommend that you suspend all updates to the source database during the full data migration phase. To migrate data with minimized downtimes, you must select schema migration, full data migration, and incremental data migration when creating the data migration task.
|Incremental data migration||DTS tracks the transaction log of the source database to capture the changes that
occur during the full data migration phase. After the full data migration phase, DTS
re-applies these data changes in the target database.
Note The incremental data migration phase is an ongoing replication process. Therefore, a migration task with incremental data migration enabled does not automatically stop. You must manually stop the migration task.
Data migration supports the following ETL features:
- Object name mapping: You can configure name mappings between source and target columns, tables, or databases that have different names.
- Data filtering: You can use SQL-like queries to filter the source table to migrate only records that match certain conditions. For example, you can specify a time range to migrate only the latest data.
The creator of a migration task can choose to receive alerts about errors that occur during migration and take immediate action.
To minimize the impact on your normal operations, you can set transmission rates for the full data migration phase of a data migration task so that the data migration workloads do not overwhelm your source database.
Migration task lifecycle
To fulfill a data replication process in any of the three modes, you must create a data replication task. A data replication task holds all the configurations of the replication process, such as the source database, target database, selected migration phases, and objects to be migrated.
The following table describes the statuses of a data migration task.
|Task status||Description||Available operations|
|Not Started||The migration task is configured but the precheck is not performed.||Perform the precheck and delete the migration task|
|Prechecking||A precheck is being performed.||Delete the migration task|
|Precheck Passed||The migration task has passed the precheck but is not started.||Start or delete the migration task|
|Migrating||Data is being migrated.||Pause, stop, or delete the migration task|
|Migration Failed||An error occurred. You can determine the phase during which the error occurred according to the progress of the migration task.||Delete the migration task|
|Paused||The migration task is paused.||Resume or delete the migration task|
|Completed||The migration task is completed, or you have clicked End to manually stop the migration task.||Delete the migration task|
A migration task in the Migration Failed status will be retried within seven days of the failure. To discard the data migration, you must manually stop or delete the migration task.