Replication of a VM between two independent Proxmox clusters (LVM storage)

katti

New Member
Dec 20, 2025
10
1
3
Hello,

I am facing a design question regarding VM replication between two independent Proxmox clusters.

Environment
  • Two separate Proxmox clusters (Cluster A and Cluster B)
  • No shared cluster configuration
  • Both clusters use LVM storage (no ZFS)
  • Inter-site VPN link
  • Proxmox Backup Server available
Requirement

The objective is to ensure activity continuity by having a VM from Cluster A available on Cluster B with minimal delay.
The two clusters are fully independent, and storage is LVM on both sides.
I would like to understand what is the correct and supported way to replicate (or keep synchronized) a VM from one cluster to another in this context.
Are there any recommended approaches or technical constraints to be aware of when both sides use LVM?


Thank you for your guidance.
 
Hi @katti , currently there is no PVE mechanism that would synchronously or asynchronously replicate your volumes between the site when LVM is involved.
You either need to move replication down the stack (into the underlying storage /SAN) or up the stack (special sync application or application layer sync).


Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox
 
  • Like
Reactions: katti
@katti
Depending on your requirements on a how long the services can be unavailable and on how much fresh data can be lost, a workaround could be frequent backups to PBS. And restoring to the other cluster when needed).

Thanks to "dirty bitmaps" ( https://pbs.proxmox.com/docs/technical-overview.html#fixed-sized-chunks ), backups don't necessarily take much time. (Note that dirty bitmaps are reset if the VM was rebooted after the previous backup).
 
Last edited:
  • Like
Reactions: Johannes S and UdoB
Late to this, but in case it helps you or anyone landing here from search.

As bbgeek17 said, PVE has no built in way to replicate LVM volumes between two independent clusters. Storage replication (pvesr) only works with ZFS. With LVM on both sides, your realistic options are:

1. PBS backup and restore. Back up the VM frequently from cluster A, and restore on cluster B when needed. Supported and simple. Dirty bitmaps keep incremental backups fast, but they reset when the VM restarts. Your RPO is the backup interval, and your RTO is the restore time, which grows with disk size. Live restore can start the VM before the restore finishes, at the cost of slower reads at first.

2. Storage level replication, if your SAN supports it. Fast, but you script the failover yourself: present the volumes on cluster B, recreate the VM configs, adjust IPs and start in the right order.

3. Block level replication underneath the hypervisor. Disclosure: I work on one, KVMDR, which does near sync replication between independent Proxmox clusters over a VPN, with orchestrated failover, failback and non disruptive DR tests. Happy to share details if useful.

Two constraints to check whatever you choose: bandwidth on the VPN link against your daily change rate, and IP addressing on cluster B if the subnets differ. And test the failover regularly. The plan you have never run is the one that fails.
 
  • Like
Reactions: Johannes S