Oracle Flashback Table and MySQL Snapshots - Oracle to Aurora MySQL Migration Playbook
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Oracle Flashback Table and MySQL Snapshots

Feature compatibility Amazon SCT / Amazon DMS automation level Amazon SCT action code index Key differences

Three star feature compatibility

N/A

N/A

Storage-level backup managed by Amazon RDS.

Oracle Usage

Oracle Flashback Table is a data protection feature used to undo changes to a table and rewind it to a previous state, not from the backup. While Flashback table operations are running, the affected tables are locked, but the rest of the database remains available.

If the structure of a table has been changed since the point of restore, the FLASHBACK will fail.

Make sure that the row movement is turned on.

The data to restore must be found in the undo, and the database administrator manages the size and retention.

You can restore a table to a System Change Number (SCN), restore point, or timestamp.

Examples

Flashback a table using SCN (query V$DATABASE to obtain the SCN).

SELECT CURRENT_SCN FROM V$DATABASE;
FLASHBACK TABLE employees TO SCN 3254648;

Flashback a table using a restore point (query V$RESTORE_POINT to obtain restore points).

SELECT NAME, SCN, TIME FROM V$RESTORE_POINT;
FLASHBACK TABLE employees TO RESTORE POINT employees_year_update;

Flashback a table using a timestamp (query V$PARAMETER to obtain the undo_retention value).

SELECT NAME, VALUE/60 MINUTES_RETAINED
FROM V$PARAMETER
WHERE NAME = 'undo_retention';
FLASHBACK TABLE employees TO
TIMESTAMP TO_TIMESTAMP('2017-09-21 09:30:00', 'YYYY-MM-DD HH:MI:SS');

For more information, see Backup and Recovery User Guide in the Oracle documentation.

MySQL Usage

Snapshots are the primary backup mechanism for Amazon Aurora databases. They are extremely fast and nonintrusive. You can take snapshots using the Amazon RDS Management Console or the Amazon CLI. Unlike RMAN, there is no need for incremental backups. You can choose to restore your database to the exact time when a snapshot was taken or to any other point in time.

Amazon Aurora provides the following types of backups:

  • Automated backups — Always enabled on Amazon Aurora. They do not impact database performance.

  • Manual backups — You can create a snapshot at any time. There is no performance impact when taking snapshots of an Aurora database. Restoring data from snapshots requires creation of a new instance. Up to 100 manual snapshots are supported for each database.

Examples

For examples, see MySQL Snapshots.

Summary

Description Oracle Amazon Aurora

Create a restore point

CREATE RESTORE POINT
  before_update GUARANTEE
  FLASHBACK DATABASE;
aws rds create-db-cluster-snapshot
  --db-cluster-snapshotidentifier Snapshot_name
  --db-cluster-identifier Cluster_Name

Configure flashback retention period

ALTER SYSTEM SET
  db_flashback_retention_target=2880;

Configure the Backup retention window setting using the Amazon management console or Amazon CLI.

Flashback table to a previous restore point

shutdown immediate;
startup mount;
flashback database to
  restore point before_update;

Create new cluster from a snapshot.

aws rds restore-db-cluster-from-snapshot
    --db-cluster-identifier NewCluster
    --snapshot-identifier SnapshotToRestore
    --engine aurora-mysql

Add new instance to the cluster.

aws rds create-db-instance
    --region useast-1
    --db-subnet-group default
    --engine aurora-mysql
    --db-cluster-identifier clustername-restore
    --db-instanceidentifier newinstance-nodeA
    --dbinstance-class db.r4.large

Use mysqldbexport and mysql to copy the table from the restored instance to the original instance.

Flashback table to a previous point in time

shutdown immediate;
startup mount;
FLASHBACK DATABASE TO TIME
  "TO_DATE ('01/01/2017','MM/DD/YY')";

Create a new cluster from a snapshot and provide a specific point in time.

aws rds restore-db-cluster-to-point-in-time
  --db-cluster-identifier clustername-restore
  --source-db-cluster-identifier clustername
  --restore-to-time 2017-09-19T23:45:00.000Z

Add a new instance to the cluster:

aws rds create-db-instance
  --region us-east-1
  --db-subnetgroup default
  --engine aurora-mysql
  --db-cluster-identifier clustername-restore
  --db-instance-identifier newinstance-nodeA
  --db-instance-class db.r4.large

Use mysqldbexport and mysql to copy the table from the restored instance to the original instance.

For more information, see mysqldump — A Database Backup Program in the MySQL documentation, rds in the CLI Command Reference and Restoring a DB instance to a specified time and Restoring from a DB snapshot in the Amazon RDS user guide.