How to Back Up Virtual Machines and Containers

Virtualisation and containerisation have fundamentally changed how UK businesses deploy and manage their IT workloads. Virtual machines running on Hyper-V, VMware, or Azure host the majority of business-critical applications, while containers orchestrated through Docker and Kubernetes are increasingly used for modern microservices architectures, development environments, and cloud-native applications. But with this shift in how workloads are deployed comes a corresponding shift in how they must be protected.

Backing up virtual machines and containers is conceptually different from backing up traditional physical servers, and the strategies that worked well in the physical world do not always translate directly to virtualised and containerised environments. UK businesses that fail to adapt their backup strategies to these modern workload types risk discovering — too late — that their backup and recovery capabilities have significant gaps.

This guide explains the key principles, tools, and strategies for backing up both virtual machines and containers effectively, ensuring your business can recover from any data loss scenario — whether caused by hardware failure, ransomware, human error, or natural disaster.

Virtual Machine Backup Fundamentals

Virtual machines consist of two main components: the virtual disk files (containing the operating system, applications, and data) and the configuration files (defining the VM's hardware settings, network connections, and resource allocations). A complete VM backup must capture both components to enable full restoration.

Modern VM backup solutions use hypervisor-level snapshots and Changed Block Tracking (CBT) to create efficient backups without installing agents inside each VM. Rather than backing up every block on the virtual disk for each backup job, CBT identifies only the blocks that have changed since the last backup, dramatically reducing backup windows, network bandwidth consumption, and storage requirements.

The three primary backup approaches for VMs are image-level backups (capturing the entire VM as a single unit, enabling full VM restoration), file-level backups (backing up individual files and folders within the VM, enabling granular restoration), and application-aware backups (understanding the internal state of applications like SQL Server, Exchange, and Active Directory to ensure consistent, recoverable backups). The most robust backup strategies combine all three approaches.

Backing Up VMs on Different Platforms

Microsoft Hyper-V

For UK businesses running Hyper-V — whether on-premises Windows Server or Azure Stack HCI — several backup solutions integrate natively with the hypervisor. Veeam Backup and Replication is the market leader, offering comprehensive Hyper-V backup with CBT support, instant VM recovery, and flexible storage targets. Windows Server Backup provides a basic but free option included with Windows Server, suitable for smaller environments. Altaro VM Backup (now Hornetsecurity) offers an SME-focused solution with straightforward pricing and excellent Hyper-V integration.

VMware vSphere

VMware environments benefit from a mature backup ecosystem. Veeam again leads the market, with deep vSphere integration, Storage API for Data Protection (VADP) support, and advanced features like SureBackup for automated recovery verification. Nakivo Backup and Replication provides a cost-effective alternative with strong VMware support, while Commvault offers enterprise-grade protection for larger environments.

Azure Virtual Machines

Azure VMs benefit from Azure Backup — Microsoft's native backup service that provides agent-less VM backup with application-consistent snapshots, configurable retention policies, and geo-redundant storage options across UK data centres. Azure Backup is deeply integrated with the Azure platform, making it the natural choice for most UK businesses running Azure workloads. For organisations requiring cross-platform backup management (Azure plus on-premises), third-party solutions like Veeam Backup for Azure provide a unified management experience.

Container Backup: A Different Challenge

Containers present fundamentally different backup challenges compared to virtual machines. By design, containers are ephemeral — they are created, run their workload, and can be destroyed and recreated at any time. The container image itself (the blueprint from which containers are launched) is typically stored in a container registry and does not need traditional backup in the same way as a VM. What does need backing up is the persistent data associated with containerised applications — the databases, file stores, and configuration data that containers access through persistent volumes.

The distinction between stateless and stateful containers is crucial for backup planning. Stateless containers (such as web servers and API gateways) can be recreated from their image and configuration without data loss — their "backup" is essentially the container image in the registry and the orchestration configuration (Kubernetes manifests or Docker Compose files). Stateful containers (such as databases, message queues, and file storage services) maintain data in persistent volumes that must be backed up using dedicated strategies.

Kubernetes Backup Strategies

For UK businesses running Kubernetes — whether on Azure Kubernetes Service (AKS), Amazon EKS, or on-premises clusters — the backup strategy must encompass both the cluster state (namespaces, deployments, services, config maps, secrets) and the persistent volume data. Velero (formerly Heptio Ark) is the most widely used open-source tool for Kubernetes backup, providing the ability to back up and restore both cluster resources and persistent volumes. Kasten K10 by Veeam offers a more feature-rich commercial solution with a user-friendly dashboard, application-aware backup policies, and multi-cluster support.

The 3-2-1 Rule for Modern Workloads

The time-tested 3-2-1 backup rule remains relevant for virtual machines and containers: maintain at least three copies of your data, stored on at least two different types of media, with at least one copy held off-site. For UK businesses, this typically translates to the production data on your primary storage, a local backup on a dedicated backup repository (NAS, SAN, or dedicated backup server), and an off-site copy in a UK-based cloud storage service or a second geographic location.

For businesses subject to ransomware risk — which effectively means all UK businesses — consider extending this to a 3-2-1-1 strategy, where the additional "1" represents an immutable or air-gapped backup copy that cannot be modified or deleted by ransomware, even if attackers gain administrative access to your systems. Azure Immutable Blob Storage, AWS S3 Object Lock, and Veeam's hardened Linux repository all provide options for creating immutable backup copies.

Testing Your Recovery Capabilities

A backup that has never been tested is not a backup — it is a hope. Yet studies consistently show that the majority of UK businesses never perform full restore tests. They assume that because the backup job completes successfully each night, they will be able to recover when needed. This assumption is dangerously optimistic.

Establish a regular restore testing schedule. At minimum, perform a full VM restore test quarterly and a file-level restore test monthly. For containers, practise restoring a complete namespace from backup into a test cluster. Document the restore procedure, time how long each step takes, and verify that the restored systems function correctly. The goal is not just to confirm that the data can be restored, but to verify that your team can execute the recovery process under pressure and that recovery times meet your business requirements.

Veeam's SureBackup feature automates much of this process for VM backups, automatically restoring VMs into an isolated sandbox environment and running verification scripts to confirm they boot correctly and key services are running. For organisations with the resources, regular disaster recovery drills that simulate complete site failures and test end-to-end recovery across VMs, containers, and supporting infrastructure provide the highest level of confidence.