Use Limits
Last updated: 2025-03-21 15:13:02
Limits on Region and AZ
The global replication feature can replicate data in the same AZ or across AZs between any Tencent Cloud regions no matter where instances in the replication group are deployed. Currently, only the following regions and AZs support global replication, and AZs of instances in a global replication group cannot be adjusted.
Region | AZ |
Hong Kong (China) | Hong Kong Zones 2 and 3 |
Chengdu | Chengdu Zone 1 |
Silicon Valley | Silicon Valley Zone 2 |
Virginia | Virginia Zone 2 |
Shanghai | Shanghai Zones 4, 5, 6, and 7 |
Beijing | Beijing Zones 5 and 7 |
Guangzhou | Guangzhou Zones 4, 5, 6, and 7 |
Nanjing | Nanjing Zones 2 and 3 |
Singapore | Singapore Zones 2 and 4 |
Frankfurt | Frankfurt Zone 2 |
Bangkok | Bangkok Zone 2 |
São Paulo | São Paulo Zone 1 |
Limits on the version and architecture of global replication group instances
Global replication supports instances running on 4.0 Standard Architecture, 4.0 Cluster Architecture, 5.0 Standard Architecture, and 5.0 Cluster Architecture.
The architectures of instances in a replication group cannot be changed; for example, you cannot change an instance from Cluster Architecture to Standard Architecture.
The version and architecture of instances to be added to a replication group must be the same as those of the master instance specified during group creation.
Limits on the specifications of global replication group instances
It is recommended to set the number of shards for replication group instances to a power of 2, with a maximum value of 64, such as 8, 16, 32, or 64. Otherwise, uneven shard distribution may occur.
When creating a replication group, you must specify a master instance with at least two replicas for the group.
Currently, you can add up to four instances in a global replication group in the following deployment schemes: one master and three read-only instances, four master instances, or two master and two read-only instances.
The specifications of instances to be added to a replication group must be the same as those of existing instances in the group, and their memory capacity must be greater than or equal to the used capacity of the master instance specified during group creation.
When you change the specifications, all instances in a replication group must have the same specifications; otherwise, performance or capacity problems may occur.
Limits on the parameter settings
The
maxmemory-policy parameter of instances in a global replication group must be set to noeviction.Limits on Command Sync
The FLUSHDB or FLUSHALL command will be synced to all instances in the replication group. Therefore, run them with caution.
The Pub and Sub command groups will not be synced. We recommend that you use the
Stream data structure to replicate notification messages across regions.When the RESTORE command is synced, if the target instance has the same key, it will not be executed.
Limits on the sync granularity
Currently, syncing is performed at the instance granularity, that is, all instance data will be synced. You cannot choose to sync partial instance data.
Was this page helpful?
You can also Contact Sales or Submit a Ticket for help.
Yes
No