Amazon Relational Database Service backups

Amazon RDS and AWS Backup

When you consider the options to back up your Amazon RDS instances and clusters, it's important to clarify which kind of backup you want to create and use. Several AWS resources, including Amazon RDS, offer their own native backup solutions.

Amazon RDS gives the option of making automated backups and manual backups. In Amazon RDS terminology, all recovery points created by AWS Backup, including those in a backup plan, are considering manual backups.

When you use AWS Backup to create a backup (recovery point) of an Amazon RDS instance, AWS Backup checks if you have previously used Amazon RDS to create an automated backup. If an automated backup exists, AWS Backup creates a incremental snapshot copy (copy-db-snapshot operation). If no backup exists, AWS Backup creates a snapshot of the instance you indicate, instead of a copy (create-db-snapshot operation).

The first snapshot made by AWS Backup, created by either operation, will result in 1 full snapshot. All subsequent copies of this will be incremental backups, as long as the full backup exists.

Important

When a AWS Backup backup plan is scheduled to create multiple daily snapshots of an Amazon RDS instance, and when one of those scheduled AWS Backup Start Backup window coincides with the Amazon RDS Backup window, the data lineage of the backups can branch off into non-identical backups, creating unplanned and conflicting backups. To prevent this, ensure your AWS Backup backup plan or Amazon RDS window do not coincide in their times.

Amazon RDS continuous backups and point in time restore

Continuous backups involve using AWS Backup to create a full backup of your Amazon RDS resource, then capturing all changes through a transaction log. You can achieve a greater granularity by rewinding to the point in time you desire to restore to instead of choosing a previous snapshot taken at fixed time intervals.

See continuous backups and PITR supported services and managing continuous backup settings for more information.

Amazon RDS Multi-Availability Zone backups

AWS Backup backs up and supports Amazon RDS for MySQL and for PostgreSQL Multi-AZ (Availability Zone) deployment options with one primary and two readable standby database instances.

Multi-Availability Zone backups are available in the following regions: Asia Pacific (Sydney) Region, Asia Pacific (Tokyo) Region, Europe (Ireland) Region, US East (Ohio) Region, US West (Oregon) Region, Europe (Stockholm) Region, Asia Pacific (Singapore) Region, US East (N. Virginia) Region, and Europe (Frankfurt) Region.

The Multi-AZ deployment option optimizes write transactions and is ideal when your workloads require additional read capacity, lower write transaction latency, more resilience from network jitter (which impacts the consistency of write transaction latency), and high availability and durability.

To create a Multi-AZ cluster, you can choose either MySQL or PostgreSQL as the engine type.

In the AWS Backup console, there are three deployment options:

• Multi-AZ DB cluster: Creates a DB cluster with a primary DB instances and two readable standby DB instances, which each DB instance in a different Availability Zone. Provides high availability, data redundancy, and increases capacity to server-ready workloads.

• Multi-AZ DB instance: Creates a primary DB instance and a standby DB instance in a different Availability Zone. This provides high availability and data redundancy, but the standby DB instance doesn’t support connections for read workloads.

• Single DB instance: Creates a single DB instance with no standby DB instances.

To create a backup for Amazon RDS, see Creating a backup for scheduling a backup as part of your backup plans or creating an on-demand backup.

Note

Point-in-Time Recovery (PITR) can support instances, but not clusters.

Copying a Multi-AZ DB cluster snapshot is not supported.

Differences between a multi-AZ cluster and an RDS instance

A backup in a single Availability Zone or in two Availability Zones is an RDS instance; a deployment and backup with three or more instances is a cluster, similar to Amazon Aurora, Amazon Neptune, and Amazon DocumentDB clusters.

The ARN (Amazon Resource Name) is rendered differently depending on if instance or cluster is used:

An RDS instance ARN: arn:aws:rds:region: account:db:name

An RDS Multi-Availability Cluster: arn:aws:rds:region:account :cluster:name

For more information, consult Multi-AZ DB cluster deployments in the Amazon RDS User Guide.

For more information on Creating a Multi-AZ DB cluster snapshot, see the Amazon RDS User Guide.