What is FusionOne Compute?
FusionOne Compute is an enterprise-grade virtualization and cloud computing platform designed to consolidate physical servers into flexible virtual machines (VMs). Built on a KVM-based hypervisor, it provides centralized management of compute, storage, and network resources through a unified web console and RESTful APIs. According to Huawei's official product documentation, FusionOne Compute supports features such as live migration, high availability (HA), dynamic resource scheduling, and thin provisioning.
The platform is typically deployed in three layers:
Compute nodes (CNA) host virtual machines and provide CPU, memory, and local storage resources.
Management nodes (VRM) control the cluster, handle user requests, and maintain configuration databases.
Storage resources can be local disks, SAN/NAS devices, or Huawei’s distributed storage (FusionStorage) for shared VM storage.
Why Do You Need to Back Up FusionOne Compute
Virtualization does not eliminate the risk of data loss. VMs on FusionOne Compute can be affected by:
- Logical errors such as accidental deletion, database corruption, or faulty application updates.
- Hardware failures on compute nodes or underlying storage arrays.
- Ransomware attacks that encrypt VM disks and demand payment.
- User mistakes during routine maintenance, such as deleting the wrong VM or snapshot.
A reliable backup strategy provides point-in-time recovery, allowing you to restore an entire VM, individual virtual disks, or even single files without rebuilding the environment from scratch. For organizations running FusionOne Compute in production, backup is also required for compliance with regulations like GDPR, HIPAA, or industry-specific data retention rules.
Key Challenges of FusionOne Compute Backup
Backing up FusionOne Compute is not a simple "copy the VM folder" task. Several platform-specific challenges must be addressed:
Dynamic resource scheduling and live migration
VMs can move between compute nodes automatically based on load or maintenance operations. A backup job that assumes a VM resides on a fixed host may fail when the VM migrates mid‑backup.
Distributed storage complexity
When FusionOne Compute uses Huawei FusionStorage or other distributed storage, VM data is spread across multiple nodes and protected by replication or erasure coding. Native
snapshots may consume significant storage capacity and impact performance if not managed carefully.
Lack of built‑in application‑consistent snapshots
Although FusionOne Compute supports VM snapshots, these are crash‑consistent by default. Applications such as Microsoft SQL Server, Oracle, or Exchange require VSS or pre‑freeze scripts to flush transactions before the snapshot is taken. Without application awareness, restored databases may be in an inconsistent state.
Backup window and performance impact
Because FusionOne Compute shares compute and storage resources across all VMs, a poorly designed backup job can saturate the storage network and degrade production performance. Incremental backups with changed block tracking (CBT) are essential to minimize I/O load.
Three Common Methods to Protect FusionOne Compute
Administrators typically choose one of three approaches to protect FusionOne Compute VMs. Each has strengths and limitations.
Method 1: Native Snapshots
FusionOne Compute includes a snapshot feature that captures the state of a VM at a specific point in time. Snapshots are fast to create and can be used for short-term rollback.
How it works:
The hypervisor creates a delta disk for each VM disk. New writes go to the delta file while the original disk remains read-only. You can revert to the snapshot to discard all changes made after the snapshot point.
Pros:
- Instantaneous creation with minimal overhead.
- No additional software required.
- Useful for testing patches or configuration changes.
Cons:
- Snapshots are stored in the same storage pool as production data. If the storage array fails, snapshots are lost too.
- Long snapshot chains degrade VM performance significantly.
- Snapshots are not backups. They do not protect against storage hardware failure, accidental deletion of the entire VM, or ransomware that deletes snapshots.
- No application-level consistency unless you manually quiesce the application.
Best for: Short-term rollback before a risky operation, not for disaster recovery or long-term retention.
Method 2: Export VMs
FusionOne Compute allows administrators to export a VM as an OVF/OVA package or a set of disk image files. Exported VMs can be stored on another server or archived to object storage.
How it works:
You select a VM, choose the export format, and specify a destination. The hypervisor reads all disk and configuration data and writes them to a single archive file.
Pros:
- Creates an independent copy that can be stored off-site.
- No extra licensing cost.
- Can be used to migrate VMs between different FusionOne Compute clusters.
Cons:
- Exporting a running VM requires a snapshot or downtime.
- Full exports are slow for large VMs and consume high network bandwidth.
- No incremental export option means repeated full copies waste time and storage.
- Manual process with no scheduling, retention, or verification capabilities.
- Restoring a single file or application object is impossible without first restoring the entire VM.
Best for: One-time migrations or ad hoc archival, but not for automated backup.
Method 3: Third-Party Backup Tool
Purpose-built backup software like Vinchin Backup & Recovery integrates with FusionOne Compute’s management API to deliver agentless, policy-driven data protection.
How it works:
The backup server connects to the FusionOne Compute management node (VRM), discovers all VMs, and triggers hypervisor-level snapshots through the API. It then reads only the changed data blocks and stores them in a deduplicated backup repository. Application-consistent snapshots are achieved by invoking in-guest scripts or VSS before the snapshot.
Pros:
- Fully automated with scheduling, retention, and reporting.
- Agentless architecture eliminates the need to install backup agents inside each VM.
- Incremental forever backups with CBT drastically reduce backup windows.
- Supports instant VM recovery, file-level restore, and cross-platform recovery.
- Centralized management for large-scale FusionOne Compute environments.
Cons:
- Requires an additional license.
- Initial deployment and configuration take time.
- Must ensure the third-party tool supports the exact FusionOne Compute version and API.
Best for: Production environments where RPO and RTO matter, and where automation and recoverability are critical.
Common FusionOne Compute Backup Selection Mistakes
Even with the right tools, many organizations make avoidable mistakes when designing their FusionOne Compute backup architecture. Avoid these common pitfalls:
Relying solely on snapshots
Snapshots are not backups. If the primary storage array dies, every snapshot on that array dies with it. Ransomware also actively deletes snapshots. A proper backup copy must exist on a separate medium or off‑site location.
Ignoring application consistency
A crash‑consistent snapshot may allow the VM to boot, but databases and transactional applications may fail to mount or require lengthy recovery. For SQL Server, Oracle, Exchange, or any application with a write cache, use application‑aware backup.
Not testing restores
A backup is only as good as its restore test. Many teams never test recovering a VM until a real disaster occurs. Schedule quarterly restore drills to validate backup integrity and measure RTO.
Backing up only critical VMs
Some administrators protect only the VMs they think are important. A single unprotected VM can cause a compliance violation or a business interruption. Use automated discovery to ensure every VM is included, and use tags or policies to define protection tiers.
Underestimating storage and network impact
Backup jobs that run during business hours can saturate the storage network and degrade FusionOne Compute performance. Use incremental backups with CBT, throttle bandwidth if necessary, and schedule backup windows during low activity.
Choosing a generic backup product without FusionOne Compute API support
Not all backup software understands FusionOne Compute's object model or management API. A tool that tries to back up VMs at the storage layer may miss VM configuration or fail after live migration. Verify the vendor explicitly lists FusionOne Compute support and version compatibility.
Core Requirements of FusionOne Compute Backup Software
When evaluating backup software for FusionOne Compute, look for the following capabilities to ensure reliable protection and fast recovery.
Agentless VM-level backup
The solution should communicate directly with FusionOne Compute’s management node (VRM) via REST API. No backup agent should be required inside each VM. Agentless backup simplifies deployment, reduces maintenance, and avoids performance overhead inside guest operating systems.
Hypervisor-level snapshot integration
The backup software must trigger FusionOne Compute snapshots through the native API, not through third-party storage snapshots alone. This ensures that the VM configuration, disks, and memory state are captured together. After the backup, the tool must automatically delete temporary snapshots to avoid performance degradation.
Changed block tracking (CBT)
Incremental backups are essential for large environments. CBT records which disk blocks have changed since the last backup, allowing the software to read only those blocks. This reduces backup time by up to 90% and drastically lowers storage and network load on the FusionOne Compute cluster.
Application-consistent backup
For VMs running databases, email servers, or ERP systems, the backup tool must support application-aware processing. This is typically achieved by installing a lightweight runtime inside the VM that coordinates with VSS on Windows or pre-freeze/post-thaw scripts on Linux. The runtime flushes application transactions before the hypervisor snapshot is taken.
Independent backup repository with deduplication and compression
The backup data must be stored on a separate storage platform, not on the same FusionOne Compute storage pool. Deduplication and compression reduce storage consumption and bandwidth requirements. Support for object storage (S3-compatible) or tape enables long-term retention and off-site copies.
Instant recovery and granular restore
To minimize downtime, the backup solution should support instant VM recovery, which mounts the backup image directly on the FusionOne Compute host and starts the VM in minutes while data is copied in the background. File-level recovery allows restoring individual files or folders without restoring the entire VM, saving time for common user requests.
Multi-tenant and policy-based management
In enterprise or cloud service provider environments, administrators need to assign different backup policies, retention rules, and storage targets to different departments or tenants. Role-based access control (RBAC) and self-service restore portals are increasingly important.
Comprehensive reporting and alerting
Automated backup reports, success/failure alerts, and capacity forecasts help administrators prove compliance and identify issues before they cause data loss. Integration with monitoring systems like SNMP or email notification is a plus.
How Vinchin Addresses FusionOne Compute Backup Challenges
Vinchin Backup & Recovery is a purpose-built data protection solution that directly addresses the specific challenges of FusionOne Compute environments. Unlike generic backup tools, Vinchin integrates deeply with FusionOne Compute's management plane and hypervisor to deliver reliable, high-performance backup and recovery.
Deep FusionOne Compute API integration
Vinchin connects to FusionOne Compute's VRM using the native REST API. It automatically discovers all VMs, clusters, and storage resources. This integration ensures that live migration does not break backup jobs because Vinchin tracks VM identity and location dynamically through the management plane.
Agentless architecture with optional application awareness
Vinchin performs VM-level backup without installing persistent agents inside each VM. For application consistency, it uses a lightweight runtime that is automatically pushed to the VM during the backup job and removed afterward. This runtime invokes Microsoft VSS or Linux pre-freeze scripts to quiesce databases before the hypervisor snapshot.
CBT-based incremental forever backup
Vinchin uses changed block tracking (CBT) to perform incremental backups after the initial full backup. The result is a short backup window, minimal load on the FusionOne Compute storage network, and significant storage savings in the backup repository. Incremental data is synthesized into a new full backup according to policy, so restores remain fast without periodic full backups.
Independent, deduplicated backup repository
Vinchin stores backup data on separate storage, such as local disks, NFS/SMB shares, or S3-compatible object storage. Global deduplication and compression reduce storage footprint by up to 70%. This separation ensures that a failure of the primary FusionOne Compute storage array does not destroy the backups.
Instant VM recovery and granular restore
Vinchin's instant recovery technology mounts the backup image as an NFS datastore on the FusionOne Compute host, allowing a VM to boot in under 15 seconds while data is restored in the background. For common requests, Vinchin provides file-level recovery from both Windows and Linux VM backups, plus granular recovery for Microsoft SQL Server, Exchange, and Active Directory.
Automated DR and verification
Vinchin can replicate backups to a secondary site or cloud object storage. It also includes automated recovery verification, which boots the restored VM in an isolated environment and captures a screenshot to confirm that the operating system and applications start correctly. This eliminates the “backup exists but can’t restore” problem.
FAQ
Q1: Can I use FusionOne Compute native snapshots as a formal backup strategy?
A1: No. Native snapshots only support short-term rollbacks and are stored on the same production storage. They cannot defend against storage failures, accidental deletion, or ransomware attacks. For reliable data protection, you need to replicate data to an independent backup repository with professional backup tools.
Q2: Does FusionOne Compute support standard CBT for incremental backup?
A2: FusionOne Compute lacks standard CBT interfaces. Vinchin adopts hypervisor-level block and metadata analysis to capture incremental data accurately after the first full backup. This avoids full disk scans, effectively lowering backup time, bandwidth and storage overhead.
Q3: How to achieve application-consistent backup for SQL Server VMs on FusionOne Compute?
A3: Deploy Vinchin’s lightweight in-guest runtime, which integrates with Windows VSS to flush SQL Server transactions and create consistent snapshots automatically. It resumes database services after backup completion with no complex configuration, ensuring fully recoverable database data.
Q4: Does Vinchin support cross-platform recovery for FusionOne Compute VMs?
A4: Yes. Vinchin supports cross-hypervisor recovery, enabling restoration of FusionOne Compute VMs to VMware vSphere, Hyper-V and other KVM platforms. It facilitates flexible migration and DR testing while eliminating vendor lock-in.
Conclusion
Protecting FusionOne Compute virtual machines requires more than occasional snapshots or manual exports. The platform's dynamic scheduling, distributed storage, and lack of native application-aware backup create specific challenges that generic tools cannot address. A robust backup strategy must include agentless VM-level backup, hypervisor API integration, changed block tracking, application consistency, and an independent repository.
Vinchin Backup & Recovery meets all these requirements by tightly integrating with FusionOne Compute’s management plane and hypervisor. It automates backup scheduling, reduces storage and network impact with CBT and deduplication, and provides fast, granular recovery options including instant VM restore and file-level recovery. By avoiding common selection mistakes and deploying a purpose-built solution, you can ensure that your FusionOne Compute environment remains recoverable, compliant, and resilient against data loss.
