Inquiry Regarding VM Migration Behavior with Proxmox VE and Veeam Backup & Replication Hacer clic para indicarle a Gmail que esta conversación es impo

[alejandro85]

Dear Proxmox Support Team
My name is Alejandro, and I am currently evaluating Proxmox VE in my homelab environment. I am writing from Argentina, and I apologize in advance for any language mistakes.

I am testing Proxmox VE 9.1.15 (latest updates installed) together with Veeam Backup & Replication versions 12.3.2.4165 and 13.0.1.1071.

During my testing, I observed the following behavior:

In a two-node cluster, when a VM is migrated from Node A to Node B and a backup job is executed immediately afterward, the backup fails because the VM is reported as not existing on the expected node.

From my research (forums and documentation), I understand that this may be related to how Veeam performs inventory updates — using periodic polling (every 15–20 minutes) rather than push notifications, unlike VMware vCenter.

My question is whether this behavior is expected, and if there are any plans to improve synchronization in future releases. This is particularly important when considering production environments that rely on VM live migration or HA.

In my company, we are currently evaluating a potential migration from VMware to Proxmox, and understanding this behavior is an important factor in our decision process.

Thank you very much for your time and support.

 

[Bu66as]

Hallo @alejandro85,

This is expected behavior, but it's a Veeam limitation, not a Proxmox issue. The Proxmox API (/api2/json/cluster/resources) always returns the correct current node for every VM in real-time — there is no delay on the Proxmox side after a migration.

Veeam's Proxmox integration relies on periodic inventory polling rather than querying the live cluster state before each backup job. That's why the backup fails right after a migration — Veeam still has the old node cached.

Workarounds:

• Trigger a manual rescan/inventory update in Veeam before running backup jobs after migrations
• Schedule backup jobs with enough delay after planned maintenance windows
• Use Veeam's pre-job script to force a rescan

Alternative to consider: Proxmox Backup Server (PBS) is fully integrated with PVE and inherently aware of VM migrations — it always backs up from the correct node without any polling delay. For a VMware-to-Proxmox migration, PBS is the natural replacement for the backup layer as well.

Whether Veeam improves their polling mechanism is entirely on Veeam's roadmap — I'd recommend opening a feature request with them directly.

If your company is evaluating a production migration from VMware to Proxmox, B1 Systems (official Proxmox partner) offers consulting and support specifically for this kind of transition.

 

[alejandro85]

Ohh thank you mr. @Bu66as I'll keep on mind!!
Regards!!