This topic describes the limits of the Elastic Compute Service (ECS) disaster recovery feature on operating systems, platforms, databases, and applications.
|Windows Server||2008 R2, 2012, 2012 R2, and 2016|
Notice You must make sure that the /boot partition and the / partition reside on the same disk. If the partitions do not reside on the same disk, move the partitions to the same disk, and then register the ECS instance for which you want to enable ECS disaster recovery.
ECS disaster recovery is implemented based on the disk-level data replication technology and is independent of the underlying platform. The following table describes the platforms that support ECS disaster recovery. If you have questions, submit a ticket.
|Physical machine||Full support|
|vSphere||5.5, 6.0, 6.5, 6.7, and 7.0|
Databases and applications
You can apply the replication technology of ECS disaster recovery to all types of databases and applications.
In most cases, automated scripts are required for various applications to ensure consistency among updates. You can use the tools and scripts that are provided by Alibaba Cloud to implement ECS disaster recovery. This ensures smooth recovery of applications.
ECS disaster recovery also has the following limits:
- If the size of a physical volume where the system disk of an ECS instance resides exceeds 2 TB, you cannot perform a full restoration on the ECS instance.
- A single physical volume of a data disk cannot exceed 32 TB.
- Disk write limits:
If the average IO size is 4 KB, the maximum disk write speed is about 10 MB/s. If the average IO size is 64 KB, the maximum disk write speed is about 30 MB/s.
The maximum disk write speed is 10 MB/s.
If the IO size is smaller or the disk write volume is larger, the Recovery Point Objective (RPO) is not affected, but the Recovery Time Objective (RTO) is prolonged.
When you perform cross-region disaster recovery, the total throughput supported by a single site pair is 400 Mbit/s. If the data volume exceeds this limit, both the RPO and RTO may be prolonged. When you design a disaster recovery solution, you must evaluate the business situation of protected servers and estimate the amount of data written to disks.