Agentless Data Protection for Huawei FusionCompute

There aren’t many solutions that support agentless data protection for Huawei FusionCompute environments. However, we have just the right solution for administrators who would like to provide a reliable backup for Huawei’s virtual environment. With Storware vProtect 4.3 – backup and recovery solution – you get multiple scenarios to integrate this platform with one of several enterprise-grade backup providers or use existing storage.

FusionCompute architecture is very similar to any other standard virtualization solution. It consists of a Virtual Resource Manager (VRM) and one or more Compute Node Agents (CNAs). VRM provides a web-based management interface for servers and virtual machines. Servers can be grouped into clusters to provide high availability.

Storware vProtect 4.3 introduces a technical preview of agentless backup for virtual machines running on this platform. It uses snapshots of the running VM to capture its state and exports data directly from the hypervisor. Since FusionCompute provides CBT API, vProtect can easily export just the data that has changed. With this strategy, you save the time needed to perform a backup and storage space.

With Storware vProtect, it’s possible to install multiple nodes (data movers) to increase overall data transfer throughput – the granularity is the hypervisor. It means that you can use the same number of data movers as you have hosts in your environment (but this would be rather an extreme case).

vProtect also supports snapshot management for this Huawei virtualization. On top of regular snapshots used for backups, you can periodically create additional snapshots in between for quick recovery just by reverting the VM to its state. There is no need to export data during backup or import it back during the restore process with this option.

Huawei FusionCompute Backup Strategy

The entire backup process starts with a snapshot created with the VRM API. The tools installed in the VM are required for CBT to run, so depending on its snapshot, CBT or standard type is created. The next step is to download the virtual machine metadata – also provided by the VRM API.

The next step is critical for high backup efficiency. The transfer needs to go from the hypervisor (CNA) – not the manager, so vProtect prepares transfer using VRMs API and creates appropriate backup resources. This allows vProtect to locate data on the hypervisor and initiate the transfer. Once the transfer completes, vProtect cleans up all backup resources.

With incremental backups, there are currently several restrictions. As CBT is required for incremental work – this implies the requirement of guest tools to be installed inside VM. If they are not present, a full backup will be initiated instead of incremental. However, assuming everything is properly configured, vProtect follows a similar approach to the regular full backup. An additional step before the actual transfer is fetching CBT information from the hypervisor and reading just the blocks reported as changed.

The restore process is quite symmetric to the backup – first, we need to recreate the VM using its metadata. Once vProtect has information about disks that have been created, it initiates the transfer of the data back (also directly to the hypervisor, not the manager). Learn more about backup and recovery for Huawei virtualization at vProtect GitBook.

Wrap up 

With Storware vProtect, you can now protect another virtualization platform with the same solution. Not only it provides both full and incremental backups, but also snapshot management capabilities. And all of these features can be used together with any backup provider currently supported by vProtect, which makes multiple implementation scenarios possible.

Remember that vProtect supports 20 different virtualization platforms and storage providers under one license! We believe that the flexibility and versatility of the software facilitate business development – TRY NOW.