Using DRS to Migrate Data from Oracle Database to GaussDB

Scenarios

This section describes how to use real-time synchronization of DRS to migrate data from an on-premises Oracle database to Huawei Cloud GaussDB in real time. Full+incremental synchronization can ensure that data is always in sync between the source Oracle database and the destination GaussDB instance. Full synchronization is used to synchronize data. Incremental synchronization is used to synchronize data between the source and destination databases in real time.

Step 1: Create a VPC and Security Group

Step 2: Create a GaussDB Instance

Step 3: Construct Data Before Migration

Step 4: Migrating the Database

Step 5: Verify Data After Migration

Problems to Resolve

• Enterprise workloads have been growing and evolving fast, and traditional databases lack the scalability needed to keep up. Enterprises need distributed databases.
• Building a traditional database means purchasing and installing servers, systems, databases, and other software. The O&M is expensive and difficult.
• Traditional databases have poor performance when it comes to handling complex queries.
• It is hard for traditional databases to smoothly synchronize data with no downtime.

Prerequisites

• You have registered with Huawei Cloud and completed account authentication.
• Your account balance is greater than or equal to $0 USD.
• In a testing scenario, you have set up an on-premises Oracle database.
• You have obtained the IP address, port number, username, and password of the Oracle database to be migrated.

Service Architecture

How Data Migration Works

The data migration process is completed using full and incremental synchronization, which includes the following operations:

1. In the full synchronization phase, schemas, including tables, primary keys, and unique keys, are synchronized first.
2. After schemas are synchronized, incremental data extraction is started to ensure that the incremental data generated during full data synchronization is completely extracted to the DRS instance.
3. A full migration task is started.
4. An incremental synchronization is automatically started after the full migration is complete. The replay starts from the position where the full synchronization starts.
5. A comparison task is started after the incremental replay is complete to check the data consistency. Real-time comparison is supported.
6. Workloads synchronization is started if the data is consistent between the source and destination databases.

Figure 1 Migration principle

Resource Planning

The resource planning in this section is just an example. You need to adjust it as needed.

Step 1: Create a VPC and Security Group

Create a VPC and security group for the GaussDB instance.

Step 2: Create a GaussDB Instance

Create a GaussDB instance as the destination database of the migration task.

1. Log in to the Huawei Cloud console.
2. Click in the upper left corner and select a region.
3. Click the service list icon on the left and choose Databases > GaussDB.
4. In the navigation pane on the left, choose GaussDB > Instances.
5. Click Buy DB Instance.
6. On the page shown in Figure 6, configure basic information about the instance, including the instance name, billing mode, edition type, DB engine version, instance type, transaction consistency, number of shards, number of coordinator nodes, and deployment AZ.
   Figure 6 Basic information
   

7. Select the instance specifications and storage space.
   Figure 7 Instance specifications
   

8. Select the VPC created in Creating a VPC and security group created in Creating a Security Group for the instance and configure the database port.
   Figure 8 Selecting a VPC and security group
   

9. Configure the password and other information.
   Figure 9 Configuring the password and other information
   

10. Click Next, confirm the information, and click Submit.
11. Go to the instance list.

    If status of the instance becomes Available, the instance has been created.

Step 3: Construct Data Before Migration

Before the migration, prepare some data types in the source database for verification after the migration is complete. The end-to-end test data in this section is for reference only.

The following table lists data types supported by DRS.

Perform the following steps to construct data in the source database:

1. Use a database connection tool to connect to the source Oracle database based on its IP address.
2. Construct data in the source database based on data types supported by DRS.
   1. Create a test user.

      create user test_info identified by xxx;

      test_info indicates the user created for the migration, and xxx indicates the password of the user.

   2. Assign permissions to the user.

      grant dba to test_info;

   3. Create a data table for the user.

      CREATE TABLE test_info.DATATYPELIST(

      ID INT,

      COL_01_CHAR______E CHAR(100),

      COL_02_NCHAR_____E NCHAR(100),

      COL_03_VARCHAR___E VARCHAR(1000),

      COL_04_VARCHAR2__E VARCHAR2(1000),

      COL_05_NVARCHAR2_E NVARCHAR2(1000),

      COL_06_NUMBER____E NUMBER(38,0),

      COL_07_FLOAT_____E FLOAT(126),

      COL_08_BFLOAT____E BINARY_FLOAT,

      COL_09_BDOUBLE___E BINARY_DOUBLE,

      COL_10_DATE______E DATE DEFAULT SYSTIMESTAMP,

      COL_11_TS________E TIMESTAMP(6),

      COL_12_TSTZ______E TIMESTAMP(6) WITH TIME ZONE,

      COL_13_TSLTZ_____E TIMESTAMP(6) WITH LOCAL TIME ZONE,

      COL_14_CLOB______E CLOB DEFAULT EMPTY_CLOB(),

      COL_15_BLOB______E BLOB DEFAULT EMPTY_BLOB(),

      COL_16_NCLOB_____E NCLOB DEFAULT EMPTY_CLOB(),

      COL_17_RAW_______E RAW(1000),

      COL_19_LONGRAW___E LONG RAW,

      COL_24_ROWID_____E ROWID,

      PRIMARY KEY(ID)

      );

   4. Insert two rows of data.

      insert into test_info.DATATYPELIST values(4,'huawei','xian','shanxi','zhongguo','shijie', 666,12.321,1.123,2.123,sysdate,sysdate,sysdate,sysdate,'hw','cb','df','FF','FF','AAAYEVAAJAAAACrAAA');

      insert into test_info.DATATYPELIST values(2,'Migrate-test','test1','test2','test3','test4', 666,12.321,1.123,2.123,sysdate,sysdate,sysdate,sysdate,'hw','cb','df','FF','FF','AAAYEVAAJAAAACrAAA');

   5. Commit the changes to the database.

      commit;

3. Create a database in the destination GaussDB instance.
   1. Log in to the Huawei Cloud console.
   2. Click in the upper left corner and select a region.
   3. Click the service list icon on the left and choose Databases > Data Admin Service.
   4. In the navigation pane on the left, choose Development Tool to go to the login list page.
   5. Click Add Login.
   6. On the displayed page, select the DB engine, source database, and target instance, enter the login username, password, and description (optional), and enable Show Executed SQL Statements.
   7. Click Test Connection to check whether the connection is successful.

      If a message is displayed indicating the connection is successful, continue with the operation. If a message is displayed indicating the connection failed and the failure cause is provided, make modifications according to the error message.

   8. Click OK.
   9. Locate the added record and click Log In in the Operation column.
      Figure 10 Logging in to an instance
      

   10. Choose SQL Operations > SQL Window on the top menu bar.

       

   11. Run the following statement to create an Oracle-compatible database:
       test_database_info indicates the database name. Replace it as required.

       CREATE DATABASE test_database_info DBCOMPATIBILITY 'ORA';

Step 4: Migrating the Database

Create a DRS instance and migrate data from the test_info database in the on-premises Oracle database to the test_database_info database in the GaussDB instance.

Performing a Pre-migration Check

Before creating a migration task, check the migration conditions to ensure smooth migration.

Before the migration, you need to obtain the notes on migration to the cloud.

Creating a Migration Task

1. Log in to the Huawei Cloud console.
2. Click in the upper left corner and select a region.

   Select the region where the destination instance is deployed.

3. Click the service list icon on the left and choose Databases > Data Replication Service.
4. In the navigation pane on the left, choose Data Synchronization Management. On the displayed page, click Create Synchronization Task.
5. Configure synchronization instance information.
   1. Select a region, billing mode, and project, and enter a task name.

      

   2. Configure synchronization instance details. Specifically, specify Data Flow, Source DB Engine, Destination DB Engine, Network Type, DRS Task Type, Destination DB Instance, Synchronization Instance Subnet, Synchronization Mode, Specify EIP (mandatory when Network Type is set to Public Network), Specifications, AZ, Enterprise Project, and Tags (optional).

      

      

   3. Click Create Now.
6. Configure the source and destination database information.
   1. Configure DNS Server as required. Specify connection information about the source database, including the IP address, port, username, and password.

      Click Test Connection.

      

   2. Enter the username and password of the destination database.

      Click Test Connection.

      

   3. Click Next. In the displayed box, read the message carefully and click Agree.

7. Configure the synchronization task.
   1. Select the databases and tables of the source database to be migrated. In this example, select the DATATYPELIST table from the test_info database.

      

   2. Locate the database and table, respectively, and click Edit to change the database name and table name as needed.

      

   3. On the displayed dialog box, enter a new name, for example, DATATYPELIST_After.

      The name cannot include special characters. Otherwise, an error will be reported during SQL statement execution after the migration.

      

   4. Confirm the settings and click Next.

      

8. Confirm advanced settings.

   The information on the Advanced Settings page is for confirmation only and cannot be modified. After confirming the information, click Next.

   

9. Process data.
   On this page, you can process the table to be migrated. For example, you can select the column to be migrated and change its name. In this example, change the column name COL_01_CHAR______E to new-line.

   1. Click Edit next to the table to be processed.

      

   2. Edit the COL_01_CHAR______E column.

      

   3. Enter the new name new-line and click Confirm.

      

   4. Click Next.
10. Perform a pre-check.
    1. After all settings are complete, perform a pre-check to ensure that the migration is successful.

       

    2. If any check item fails, review the cause and rectify the fault. Then, click Check Again.

       

    3. After all check items pass the pre-check, click Next.
11. Confirm the task.
    1. Check that all configured information is correct.

       

    2. Click Submit. In the display box, select I have read the precautions.
    3. Click Submit.

       

12. After the task is submitted, view and manage it.

    After the task is created, return to the task list to view the status of the created task.

    

Step 5: Verify Data After Migration

When the task status changes to Incremental, the full synchronization is complete. You can log in to GaussDB and view the data migration result.

1. Wait until the migration task status becomes Incremental.

   

2. Click the task name to go to the Basic Information page.
3. On the Synchronization Progress page, view the full synchronization result.

   As shown in the following figure, the DATATYPELIST table in the TEST_INFO database has been migrated to shard_0. Two rows of data were migrated successfully.

   

4. Verify data consistency.
   1. Choose Synchronization Comparison > Object-Level Comparison to view the database and table migration results.

      

   2. Choose Synchronization Comparison > Data-Level Comparison, click Create Comparison Task, and view the migration results of the rows in the table.

      

5. Connect to the test_database_info database in GaussDB using DAS.
6. Run the following statement to query the full synchronization result:

   SELECT * FROM test_info.datatypelist_after;

   After the schema in Oracle is migrated, it will be used as the schema in GaussDB. Therefore, it is required to add the schema in the query statement for exact query.

   As shown in the following figure, all data types in the table were successfully migrated and the data is correct.

   

7. Verify incremental synchronization.
   In full+incremental synchronization, after the full synchronization is complete, the data that is written to the source database after the task is created can still be synchronized to the destination database until the task is stopped. The following describes how to synchronize incremental data from the source database to the destination database:

   1. Use a database connection tool to connect to the source Oracle database based on its IP address.
   2. Run the following statement to insert a data record into the source database:
      Insert a data record whose ID is 1.

      insert into test_info.DATATYPELIST values(1,'Migrate-test','test1','test2','test3','test4', 666,12.321,1.123,2.123,sysdate,sysdate,sysdate,sysdate,'hw','cb','df','FF','FF','AAAYEVAAJAAAACrAAA');
      commit;

   3. Run the following statement in the destination database to query the result:

      SELECT * FROM test_info.datatypelist_after;

      As shown in the following figure, the new data inserted in the source database has been synchronized to the destination database in real time.

      

8. Stop the migration task.
   After data is completely migrated to the destination database, stop the synchronization task.

   1. Locate the task and click Stop in the Operation column.

      

   2. In the display box, click Yes.

      

9. After the migration is complete, test the GaussDB performance.

   For details, see Performance White Paper.