Proxmox Backup Server
Deduplicated backup for virtual machines and containers
What is Proxmox Backup Server?
Proxmox Backup Server stores incremental, deduplicated backups of virtual machines and containers with verification and encryption built in. It integrates directly with Proxmox VE but works with any client supporting its protocol.
Best for
Backing up a virtualisation host properly
Why choose Proxmox Backup Server
Proxmox Backup Server solves backup for virtualisation properly, which is harder than it looks. It takes incremental, deduplicated backups of virtual machines and containers with changed-block tracking, so a nightly backup of a large VM moves only what actually changed rather than the whole disk image. Verification of backup content is built in, which means you can check that a snapshot is restorable rather than hoping. Encryption is available client-side, so the backup target never sees readable data, and the mechanism that prevents backups from being modified after they are written means ransomware on a virtualisation host cannot quietly corrupt the archive. Integration with Proxmox VE is direct, but it works with any client speaking its protocol.
Replaces
- Veeam
- Acronis
- Nakivo
Key features
- Incremental forever with chunk-level deduplication
- Backup verification and garbage collection
- Client-side encryption
- Native Proxmox VE integration
What to watch out for
It is designed around Proxmox VE and gives the best experience there, with other clients supported but clearly second-class. The deduplication is per-datastore and relies on the datastore being a real filesystem that can handle its requirements — putting it on a filesystem that misbehaves produces corruption that is painful to diagnose. Verification and garbage collection are separate scheduled jobs, and skipping them means deduplication data grows without bound or unverified backups accumulate. Restoring a VM requires understanding the difference between restoring to the original and restoring over a running machine, and getting it wrong can overwrite the wrong thing. Memory and IOPS requirements are real once several VMs are involved.
How to deploy
- Install on bare metal or in a VM
- Create a datastore on dedicated storage
- Add as a storage target in Proxmox VE
Getting started
Install it on hardware with directly attached, reliable storage rather than a network share or an unusual filesystem, because the datastore is the component most likely to cause trouble. Create the datastore, add it to the Proxmox VE storage list, and schedule one VM's backup before doing anything else — then verify the task completes and the snapshot is listed. Set up verification and garbage collection as scheduled jobs from the start, since both are easy to postpone indefinitely. Configure retention so it matches your actual recovery window, and enable encryption before the first backup rather than after, because changing it later means re-running everything. Restore a test VM to a different ID and boot it, because that is the workflow you are paying for.
Typical setup
Installed on hardware with directly attached, reliable storage rather than a network share, since the datastore is the component most likely to cause corruption if the filesystem misbehaves. The datastore is registered with the virtualisation host and one virtual machine's backup is scheduled and verified before anything else is added. Verification and garbage collection run as scheduled jobs from the start rather than being deferred. Retention reflects the actual recovery window, encryption is enabled before the first backup, and a test virtual machine has been restored to a different identifier and booted successfully.
Who should look elsewhere
If you do not run virtualisation, this is not the tool for you — its strengths are entirely about VM and container images and it is poor value for backing up a couple of directories. Avoid it on hardware with unreliable or non-standard storage, since the datastore is unforgiving. And if your VMs are disposable and can be rebuilt from configuration management, elaborate backup infrastructure is solving a problem you have already designed away.
Project health
- GitHub stars: 255
- Status: actively developed
Figures pulled from the GitHub API and refreshed periodically.