Tablestore provides the zone-redundant storage (ZRS) feature for disaster recovery of instance data in data centers. You can store your data in Tablestore instances that support ZRS. This way, data consistency can be ensured even if a data center becomes unavailable due to a network interruption, power failure, or disaster. The ZRS feature ensures high data availability and disaster recovery.
Redundancy types
Tablestore provides two redundancy types: locally redundant storage (LRS) and ZRS. If your business requires higher availability, we recommend that you store your data in a region that supports ZRS.
LRS
LRS stores multiple copies of your data on multiple devices of different facilities in the same zone. LRS ensures data durability and availability in case of a hardware failure.
LRS stores copies of data in a specific zone. If the zone becomes unavailable, data in the zone is inaccessible.
ZRS
ZRS stores multiple copies of your data in multiple zones of the same region. If one of the zones is unavailable, your data is still accessible.
ZRS provides disaster recovery at the data center level. If a data center becomes unavailable due to a network interruption, power failure, or disaster, Tablestore can still ensure data consistency. The entire fault handling process is user-imperceptible and can be performed without business interruption and data loss. It can also meet the requirements of key business systems for a recovery point objective (RPO) of 0 and a recovery time objective (RTO) of 0.
Differences between LRS and ZRS
LRS and ZRS differ in terms of data durability, service availability, applicable instance types, and support for search indexes. The following table describes the differences between LRS and ZRS.
Item | LRS | ZRS |
Data durability | 99.999999999% (11 nines) | 99.9999999999% (12 nines) |
Service availability | 99.9% | 99.99% |
Instance types |
|
|
Support for search indexes | Search indexes in the following regions support LRS: China (Qingdao), China (Guangzhou), China (Chengdu), Malaysia (Kuala Lumpur), UK (London), US (Silicon Valley), US (Virginia), Thailand (Bangkok), SAU (Riyadh - Partner Region), and Philippines (Manila). | Search indexes in the following regions support ZRS: China (Beijing), China (Shanghai), China (Hangzhou), China East 1 Finance, China (Shenzhen), China (Zhangjiakou), China (Ulanqab), China (Hong Kong), Japan (Tokyo), Indonesia (Jakarta), Singapore, and Germany (Frankfurt). |
Regions that support ZRS for instances and search indexes and regions in which the redundancy type of existing instances and search indexes is automatically upgraded to ZRS
The following table describes the regions that support ZRS for instances and search indexes and the regions in which the redundancy type of existing instances and search indexes is automatically upgraded to ZRS.
For more information about instances, see Instances.
Resource | Region that supports ZRS | Region that supports automatic upgrade to ZRS for existing resources |
High-performance instances | China (Hangzhou), China (Shanghai), China (Beijing), China (Zhangjiakou), China (Ulanqab), China (Shenzhen), China (Hong Kong), and Singapore | China (Hangzhou), China (Shanghai), China (Hong Kong), and Singapore |
Capacity instances | China (Hong Kong) | N/A |
Search indexes | China (Hangzhou), China East 1 Finance, China (Shanghai), China (Beijing), China (Zhangjiakou), China (Ulanqab), China (Shenzhen), China (Hong Kong), Japan (Tokyo), Indonesia (Jakarta), Singapore, and Germany (Frankfurt) | China (Hangzhou), China East 1 Finance, China (Shanghai), China (Beijing), China (Zhangjiakou), China (Ulanqab), China (Shenzhen), China (Hong Kong), Japan (Tokyo), Indonesia (Jakarta), Singapore, and Germany (Frankfurt) |
Redundancy type upgrade for existing instances
Tablestore automatically upgrades the redundancy type of existing instances and search indexes in the following regions from LRS to ZRS. For more information about the upgrade, see the upgrade announcements of Tablestore.
High-performance instances in the China (Hangzhou), China (Shanghai), China (Hong Kong), and Singapore regions
Search indexes in the China (Hangzhou), China East 1 Finance, China (Shanghai), China (Beijing), China (Zhangjiakou), China (Ulanqab), China (Shenzhen), China (Hong Kong), Japan (Tokyo), Indonesia (Jakarta), Singapore, and Germany (Frankfurt) regions
If you have concerns about the performance issues that may occur due to the automatic upgrade, submit a ticket.
Usage notes
If you create an instance whose type and region support ZRS, the default redundancy type of the instance is ZRS.
An instance can have only one redundancy type. The redundancy type that is displayed in the Tablestore console prevails.
The redundancy type of a search index is independent of the redundancy type of the instance to which the search index belongs. The redundancy type of a search index is determined by the redundancy type that is supported by the region for search indexes.
If you create a search index in a region that supports ZRS for search indexes, the default redundancy type of the search index is ZRS.
Create a ZRS instance
Log on to the Tablestore console.
In the top navigation bar, select a resource group and a region that supports ZRS for instances.
The instance types that the region supports are displayed in the console.
Click Create Instance.
On the CU Mode (Formerly On-demand Mode) tab of the Billing Method dialog box, specify an instance name, set the Instance Type parameter to High-performance, and then enter a description in the Instance Description field based on your business requirements.
Click OK.
After you create a high-performance instance, you can view the status of the instance on the Overview page. If the value of the Redundancy Type column is ZRS, the redundancy type of the instance is ZRS.
Billing rules
The ZRS feature of an instance does not affect the existing billing rules of the instance.
Other operations
To prevent operations such as accidental deletion and malicious tampering from affecting the availability of important data, you can use the data backup feature to back up the table data in an instance. For more information, see Overview.