How VM replication works in Proxmox VE with ZFS and how to set up HA

Proxmox VE has built-in replication for virtual machines and containers based on ZFS storage, which keeps data available across cluster nodes. Replication keeps a recent copy of a VM on another node, so if the primary node fails, services can resume with little disruption.

Updated September 2026 · Mr.PlanB

Proxmox Backup Strategy Data Protection

Sections

Action Points

Guidance Blocks

Overview

Replication has a few prerequisites. At least two nodes must belong to the same cluster, and the VM disks must be on local ZFS storage. The source and target nodes need ZFS pools with identical names, and the target node needs enough storage capacity.

Proxmox implements replication with ZFS snapshots. The initial synchronization sends a full copy of the VM disk, and later runs transfer only the incremental changes, which cuts network traffic considerably. The pvesr command-line tool manages replication tasks and lets administrators set schedules and targets.

To configure replication:

1. Create matching ZFS storage pools on all cluster nodes. 2. Make sure the storage names are identical on every node. 3. Select the VM, open the Replication tab, and add a replication job with the target node and interval. 4. Watch the replication status in the Replication tab logs.

You can combine replication with High Availability settings. If a node fails, HA can restart the VM automatically on the node that holds the replica.

Replication improves redundancy and shortens migration time, but it does not replace backups. It mirrors every change, including corruption and accidental deletion, so you still need a separate backup strategy for full data protection.

FAQs