Solution Introduction

Context

With the rapid development of the information technology (IT), information systems are becoming ever important for critical services in a variety of industries. Service interruptions in storage systems may result in severe economic loss, brand image damage, and critical data loss, especially in the fields of communications, finance, medical care, e-commerce, logistics, and government. Therefore, service continuity is critical to the construction of storage systems.

Many industries build DR centers to protect production data and ensure service continuity of critical applications. In this construction mode, a production center corresponds to a DR center. The DR center does not provide services for external applications if the production center works properly. Once the production center encounters a disaster that leads to service interruptions and cannot be quickly recovered, the DR center is triggered to provide services. These DR systems are facing the following challenges:

• If the production center encounters a disaster such as a power failure, fire, flood, or earthquake, you must manually switch services from the production center to the DR center. Professional recovery measures and time-consuming debugging may be required. Services are interrupted for a long time.
• The DR center does not provide services and remains idle for most of the time, resulting in low resource utilization.

In this case, Huawei launches the HA solution to help customers resolve the preceding pain points.

Solution Description

The HA DR solution uses Huawei virtual intelligent storage devices and the HyperMetro array technology to write service data to two data centers in real time. Data in the two data centers is consistent in real time. When a device is faulty or even one data center is faulty, services are automatically switched over without data loss or service interruption, implementing intra-city DR protection with zero RPO.

The HA solution applies to the following scenarios:

• Partial or all services (including network, storage, and host services) in a center fail and need to be restored.
• Links between two data centers of a DR system are of high speed and low latency.

Deployment Modes

Deploy one UltraVR on an independent server in the HA DR center to implement DR management of data in the production center. The service system in the production center must be reachable to the storage management network.

HA DR Technology

The production center and intra-city DR center use HyperMetro storage DR.

Take the distributed deployment mode as an example. Figure 1 shows the HA networking.

Figure 1 Networking diagram of the HA DR technology (distributed deployment)

• Management network: includes DR management network and production management network. The two networks can be the same network or isolated from each other.
  • DR management network

    Communication network used by UltraVR to provide external services. The maintenance terminal and third-party systems access this network using the management IP address of the DR server.

  • Production management network

    Communication network for storage devices, hosts, and FusionCompute to provide external services. UltraVR, maintenance terminal, or third-party system uses the management IP address to access the network.

• Service network: communication network for read and write service I/Os between hosts or VMs and storage resources.
• Intra-city data center network: Both the production center and intra-city DR center can provide host services and carry the same services at the same time. The two data centers back up each other. If one data center fails, the other one automatically takes over services without data loss or service interruption.
• Quorum network: When the storage systems in the production center and intra-city DR center are faulty or the link between the storage systems is faulty, quorum network is used for service arbitration.

DR Management

UltraVR provides the following functions:

• Automatically detects the VMs running on UltraVR when the virtualization infrastructure (FusionCompute) is added.
• Implements DR protection for VMs by creating an HA protected group.