Elastic Block Storage (EBS) provides cross-zone and cross-region disaster recovery for disks using asynchronous replication. If a production site fails, you can fail over to the disaster recovery site. This capability helps protect your systems from failures caused by regional disasters and ensures business availability and continuity.
Disaster recovery scenarios
Disk disaster recovery is used in two main scenarios: cross-zone disaster recovery and cross-region disaster recovery.
Cross-zone disaster recovery
If an application in a zone cannot recover quickly from force majeure events, such as a data center fire or power outage, or from device failures caused by software or hardware damage, you can use the cross-zone disaster recovery capability of async replication to recover from the failure and ensure business continuity.
Cross-region disaster recovery
If a production site fails due to a large-scale disaster, such as a tsunami or an earthquake, you can fail over your business systems to a disaster recovery site in a different region. Asynchronous replication provides cross-region disaster recovery for disk data, which ensures business continuity and protects your systems from failures caused by regional disasters.
Disaster recovery plans
Async replication
Async replication is suitable for disaster recovery for a single disk pair. This feature uses the data replication capability of block storage to asynchronously replicate data from a primary disk to a secondary disk in a different zone or region. If the primary disk fails, you can perform a failover to the secondary disk. You can then perform a reverse replication to complete the disaster recovery process.
Cross-zone disaster recovery
Cross-region disaster recovery
Replication pair-consistent group
A replication pair-consistent group is suitable for disaster recovery for multiple disk pairs. In disaster recovery scenarios where a business system involves multiple disks, you can use a replication pair-consistent group to manage the asynchronous replication of these disks in a unified manner. This feature ensures that data on all disks within the same replication group can be restored to the same point in time, which ensures data consistency for applications that use multiple disks.
The replication pair-consistent group feature asynchronously replicates data from primary disks at a production site to secondary disks at a disaster recovery site across zones or regions. If the production site fails, you can perform a failover to the disaster recovery site and then perform a reverse replication to complete the disaster recovery process.
Cross-zone disaster recovery
Cross-region disaster recovery
Limits
Specification limits
The following table describes the specification limits for async replication and replication pair-consistent groups.
Item | Description |
Number of replication relationships that can be created for a disk | 1 |
Number of replication pairs that can be added to a replication pair-consistent group | 17 |
Replication cycle | 15 minutes (Data is automatically replicated asynchronously every 15 minutes.) |
Replication speed | Data is replicated at a maximum speed of 100 MB/s. The speed may vary based on the system load. |
Primary disk type | ESSD or ESSD AutoPL disks. |
Secondary disk type | A disk of the same type, performance level, and capacity as the primary disk. |
Disk limits
When you use async replication and replication pair-consistent groups, the following limits apply to the primary and secondary disks.
The numbers in the following table correspond to the following limits:
①: After async replication is enabled, the secondary disk becomes read-only.
②: The recovery point objective (RPO) for asynchronous replication is greater than zero. Therefore, data consistency is not guaranteed between snapshots created for a primary disk and its secondary disk at the same time.
③: Data can be replicated only between two encrypted disks. You cannot replicate data between an encrypted disk and an unencrypted disk.
Limitation | Support for primary disks | Secondary disk support |
Cloud Disk Read and Write Operations | √ | × ① |
Delete a cloud disk | × | × |
Initialize a disk | × | × |
Resize a disk | × | × |
Attach a disk | √ | × |
Create a snapshot | √ | √ ② |
Roll back a disk from a snapshot | √ | × |
Change the disk type | × | × |
Modify the performance level | × | × |
Encrypted cloud disk | √ | √③ |
Enable multi-attach | × | × |
Migrate a disk with an instance | × | × |
Billing
The async replication feature is pay-as-you-go.
You are charged based on the total amount of replicated data.
The pay-as-you-go fee for async replication is calculated as follows: Unit price of data × Total amount of replicated data.For example, if 100 GB of data is replicated from Disk A in the China (Hangzhou) region to Disk B in the China (Shanghai) region and the unit price is CNY 0.5/GB, the total fee is 100 × 0.5 = CNY 50.
You are charged for cross-region replication. Cross-zone replication within the same region is free of charge.
If you perform disaster recovery drills, the disks created at the disaster recovery site are billed on a pay-as-you-go basis. For more information, see Pay-as-you-go.
RTC billing:
If you enable RTC for a cross-region async replication pair, an additional fee is charged based on the amount of replicated data. The unit price is CNY 0.1/GB..
RTC for cross-zone replication in the same region is free of charge.
After you enable RTC, if the Service-Level Agreement (SLA) for RTC is not met, you are compensated according to the monthly SLA. For more information, see Service Level Agreement for Replication Time Control (RTC) of Elastic Block Storage (EBS) Async Replication.
Terms
Before you implement disaster recovery for disks, familiarize yourself with the following basic concepts:
Basic Concepts | Description |
Asynchronous replication | Asynchronous replication periodically copies data from one disk to another across regions or zones. Due to the nature of asynchronous replication, the data on the two disks may not be completely consistent. |
Production site | A data center that hosts your running business. It can operate independently and directly supports your business operations. |
Disaster recovery site | A backup data center established in addition to the production site. If the production site fails, the disaster recovery site can take over to ensure business continuity. |
Primary disk | The disk that requires disaster recovery and backup. It is also called the source disk. After a reverse replication, the primary disk becomes the secondary disk. |
Secondary disk | The disk that stores backup data after asynchronous replication. It is also called the destination disk. After a reverse replication, the secondary disk becomes the primary disk. |
Primary site | The data center where the primary disk is located. After a reverse replication, the primary site becomes the secondary site. |
Secondary site | The data center where the secondary disk is located. After a reverse replication, the secondary site becomes the primary site. |
Replication time control (RTC) | Asynchronous replication supports RTC to provide an RPO <= 10 minutes SLA guarantee for data replication from the primary disk to the secondary disk. |
Recovery point objective (RPO) | The amount of data, measured in time, that might be lost if a disk fails. It is a data metric for the async replication feature. For example, the RPO for async replication is fixed at 15 minutes. This means that if the primary disk fails, incremental data from the 15 minutes before the failure may be lost during data restoration. |
Recovery time objective (RTO) | The period between when a primary disk fails and when it is restored to normal operation. It is a data metric for the async replication feature. For example, an RTO of 1 hour means that after a primary disk fails, data restoration is completed within 1 hour to return the disk to a normal state. |
Asynchronous replication relationship | The relationship formed by a primary disk, a secondary disk, and the asynchronous replication configuration. |
Replication pair | A pair of disks in a replication relationship. A replication pair-consistent group can contain multiple replication pairs. |
Failover | A sub-feature of async replication. The failover feature enables read and write permissions on the secondary disk. |
Reverse replication | A sub-feature of async replication. The reverse replication feature reverses the primary-secondary relationship in a replication pair. This allows data to be replicated from the original secondary disk to the original primary disk. |
Procedures
A disaster recovery disk group
Multi-group disaster recovery for cloud disks