Backup, disaster recovery, and restoration of data and applications on ECS instances - Elastic Compute Service

Use snapshots, images, Cloud Backup, or high-availability architectures to back up and restore ECS data and applications.

Use cases

Solution	Scenario
Use a snapshot to back up and restore data	Create a snapshot for a disk or a snapshot-consistent group for multiple disks to capture a point-in-time backup. If data is lost, restore it from the snapshot. Note ECS snapshots incur storage fees. See Snapshot billing.
Use an image to back up and restore data	Create an image from an ECS instance to capture a full backup of its OS, configurations, software, and data. Note Retaining custom images incurs storage fees for their associated snapshots. If a custom image is derived from a paid image, creating ECS instances from it incurs image license fees. See Image billing.
Use the async replication feature to implement disaster recovery and restore data	Use async replication and replication pair-consistent groups to implement cross-region or cross-zone disaster recovery for disks. Note Async replication is billed by bandwidth (subscription) or by replicated data volume (pay-as-you-go).
Restore data when data disk partitions are lost	Data disk partitions may be lost during troubleshooting. Use tools such as fdisk to restore partitions and data. See Tools for restoring data on Linux and Troubleshoot issues where a data disk is not visible on a Windows instance.
Use Cloud Backup to back up and restore data and implement disaster recovery	Cloud Backup backs up and restores ECS files, directories, self-managed databases on ECS instances, and ECS instances (snapshot-based). It also supports cross-region and cross-zone disaster recovery with async replication. Note Cloud Backup operations incur fees. See Billing methods and billable items.

Back up and restore data with snapshots

ECS Snapshot is an agentless backup service that creates point-in-time backups of disks or disk sets. See Snapshot overview.

Create a snapshot for data backup

Create a snapshot manually or set up an automatic snapshot policy for system and data disks.
Restore data from a snapshot

If data is lost due to accidental operations or ransomware, roll back the disk using a snapshot to restore it to the snapshot state.

Back up and restore data with images

An image is a full backup of an ECS instance, including the operating system, software, and data. Create a custom image to back up an instance, then use that image to create new instances with the same environment.

Create an image for data backup

Create a custom image from an instance or from a snapshot to capture the OS, configurations, software, and data.
Restore data from an image

Create an instance from a custom or shared image or from a community image to restore the original instance's OS, applications, and data.

Implement disaster recovery with async replication

The async replication and replication pair-consistent group features replicate disk data across regions or zones using Elastic Block Storage (EBS). If the primary site fails, fail over to the secondary site. When the primary site recovers, restore data from the secondary site.

Restore lost data disk partitions

Data disk partitions may be lost during disk troubleshooting.

For Linux instances, use tools such as fdisk to restore partitions and data. See Best practices for data recovery on Linux instances.
For Windows instances, use Disk Management or commercial software to restore data. See Restore data in a Windows instance and Best practices for data restoration on Windows instances.

Back up and restore data with Cloud Backup

Cloud Backup is a centralized disaster recovery service on Alibaba Cloud. It supports backup and restoration of the following ECS objects:

ECS files or directories. See Overview.
Self-managed databases on ECS instances, such as MySQL, SQL Server, and Oracle. See Overview (ECS database backup).
ECS instances (snapshot-based). Clone an instance to quickly build development and test environments. See Feature overview.

Cloud Backup also supports cross-region and cross-zone disaster recovery with async replication, providing multi-instance consistency and cross-instance orchestration for application-level failover and fault recovery. See Feature overview.

Implement application disaster recovery with high-availability architectures

Switch from single-instance to multi-instance deployment with a single-zone or multi-zone high-availability architecture. If one instance fails or is overloaded, other instances maintain business continuity. See Deploy a highly available architecture.