[SOLVED] Clarification request – SAN FC shared storage, Proxmox snapshots and Veeam backup compatibility

K

katti

New Member
Dec 20, 2025
9
1
3
Hello,

I am currently running a Proxmox VE 9.1.4 cluster with two nodes and a QDevice for quorum. My shared storage is an HP 3PAR SAN connected via Fibre Channel (FC).

Current Configuration:

  • Storage Type: LVM (Shared)
  • Connection: Fibre Channel
  • Disk Format: RAW
  • Backup Solution: Veeam Backup for Proxmox VE (using the official plugin)
The Challenge:I am unable to perform live backups with Veeam

Questions:

  1. Is there a recommended way to enable snapshot support on a shared FC LUN without moving to Ceph?
  2. Would moving from LVM to a clustered file system (like OCFS2 or GFS2) declared as a "Shared Directory" in Proxmox be a supported way to use .qcow2 files and thus enable snapshots?
  3. Are there any specific storage plugins or configurations for HP 3PAR that allow thin-provisioning/snapshots in a shared environment?
Thank you for your help.
 
Hi @katti,
You do not need storage snapshots for Veeam backup of any type of storage pool in PVE.

1 Yes, you can have snapshots with FC SAN by using new tech-preview feature: https://forum.proxmox.com/threads/proxmox-ve-9-lvm-snapshot-feature-still-tech-preview.175210/

2 Using QCOW on any file system is a supported way to use QCOW. Using QCOW enables you to use snapshots. Using OCFS2 or GFS is not officially supported by PVE, as in - it is not tested.

3 I am not aware of any official 3PAR plugins. Its unlikely that one would be developed based on where in technology lifecycle the 3PAR storage is. You can use your SAN with above mentioned experimental feature. There may be still some wrinkles to work out there, i.e. https://forum.proxmox.com/threads/possible-bug-with-qemu-img-zeroinit-and-lvm.180171


Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox
 
  • Like
Reactions: katti, Johannes S and alexskysilk
katti said:
hello @bbgeek17

https://community.veeam.com/blogs-a...xi-sACEl-C6qpnR4fQ_aem_5C3vGQ0lI0K007TJIS2WKg

FYI
Click to expand...
Thank you for the link @katti.
It so happens that we, at Blockbridge, are somewhat familiar with how Veeam operates:
https://forums.veeam.com/proxmox-ve-f79/fyi-potential-data-corruption-issue-with-proxmox-t95796.html

The forum post you referenced does not spell it out, but the changes in PVE that were needed to implement block level snapshots required a new virtual block driver. This broke Veeam plugin until Veeam updated the plugin to accommodate for the change.

As far as we know, Veeam continues to have unique integration with VMware (not available for PVE) in its ability to initiate storage snapshots and backup data from snapshot/clone.

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox
 
Last edited:
  • Like
Reactions: katti and Johannes S
Thank you @LnxBil and @bbgeek17 , for your very relevant comment; we will definitely take it into account.
However, my goal is to increase the resilience of the infrastructure by replicating virtual machines from the production site to the replication site, without placing the production and disaster recovery servers in the same cluster.
The objective is to ensure a true disaster recovery capability on the replication site in the event of a total loss of the primary site.

Would anyone have an architecture design or real-world experience that fits this scenario?
 
katti said:
The objective is to ensure a true disaster recovery capability on the replication site in the event of a total loss of the primary site.
Click to expand...

katti said:
Would anyone have an architecture design or real-world experience that fits this scenario?
Click to expand...
Hi @katti, many people have designed their infrastructure with DR in mind. The correct solution always depends on RPO, RTO, and budget.

The key questions your business decision-makers need to answer are:
  • How much downtime can be tolerated?
  • How far behind the primary site can the DR site lag?
  • Can the business operate if the replication link or the DR site is unavailable?
  • Does replication need to be synchronous, or is limited asynchronous I/O acceptable?
  • How much money are they willing to spend?
Possible options include ZFS replication, metro-cluster designs, or vendor solutions from Dell, HPE, and others. Note that PVE ZFS replication requires the primary and DR systems to be part of the same PVE cluster, which introduces additional dependencies and limitations. Metro-style replication, on the other hand, is not natively integrated into PVE and requires custom design and integration on top of it.

Ultimately, this is not something that can be properly designed via a forum post. A proper solution requires multiple design iterations, business discussions, and time. There are several PVE partners who would likely be willing and able to assist with this process.

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox
 
  • Like
Reactions: katti
You must log in or register to reply here.
Top Bottom
Back