Help Center/ Distributed Cache Service/ User Guide (Kuala Lumpur Region)/ Service Overview/ DCS Instance Types/ Master/Standby Redis
Updated on 2025-11-14 GMT+08:00
View PDF
Share

Master/Standby Redis

This section describes master/standby DCS Redis instances.

The major version of a DCS Redis 4.0 or later instance cannot be upgraded. For example, a master/standby DCS Redis 4.0 instance cannot be upgraded to a master/standby DCS Redis 5.0 one. If your service requires the features of higher Redis versions, create a DCS Redis instance of a higher version and then migrate data from the old instance to the new one.

Features

Master/Standby DCS instances have higher availability and reliability than single-node DCS instances.

Master/Standby DCS instances have the following features:

  1. Data persistence and high reliability

    By default, data persistence is enabled by both the master and the standby node of a master/standby instance.

    The standby node of a DCS Redis instance is invisible to you. Only the master node provides data read/write operations.

  2. Data synchronization

    Data in the master and standby nodes is kept consistent through incremental synchronization.

    After recovering from a network exception or node fault, master/standby instances perform a full synchronization to ensure data consistency.

  3. Automatic master/standby switchover

    If the master node becomes faulty, the instance is disconnected and unavailable for several seconds. The standby node takes over within 30 seconds without manual operations to resume stable services.

    • Disconnections and unavailability occur during the failover. The service client should be able to reconnect or retry.
    • After a master/standby switchover is complete, the last master node (already switched to the replica) is not recovered immediately and services will fail to access it. In this case, configure a Redis SDK by referring to Access in Different Languages.
  4. DR policies

    Each master/standby or cluster DCS instance can be deployed across AZs with physically isolated power supplies and networks. Applications can also be deployed across AZs to achieve high availability for both data and applications.

Architecture of Master/Standby DCS Redis 4.0/5.0/6.0/7.0 Instances

The following figure shows the architecture of a master/standby DCS Redis 4.0/5.0/6.0/7.0 instance.

Figure 1 Architecture of a master/standby DCS Redis 4.0/5.0/6.0/7.0 instance

Architecture description:

  1. Master/Standby DCS Redis 4.0 and later instances support Sentinels. Sentinels monitor the running status of the master and standby nodes. If the master node becomes faulty, a failover will be performed.

    Sentinels are invisible to you and is used only in the service. For details about Sentinel, see Sentinel Principle

  2. A standby node has the same specifications as a master node. A master/standby instance consists of a pair of master and standby nodes by default.
  3. To access a DCS Redis 4.0 or later instance, you can customize the port. If no port is specified, the default port 6379 will be used. In the architecture diagram, port 6379 is used. If you have customized a port, replace 6379 with the actual port.
Parent topic: DCS Instance Types