Unlock True Shared Storage and Seamless Live Migration for Proxmox VE with Everpure

Everpure FlashArray gives Proxmox VE true shared storage over Fibre Channel, iSCSI, NVMe/TCP or NFS, so VMs live-migrate between nodes and HA restarts them elsewhere without copying data. With ActiveCluster across two sites, Everpure’s validation brought protected VMs back at the second site in about 170 seconds with zero RPO.

Many organizations are revisiting their hypervisor strategies, and Proxmox Virtual Environment (Proxmox VE) is often landing on the shortlist. It is open source, runs KVM virtual machines and LXC containers side by side, has clustering and high availability built in, and is now running on more than 2.3 million servers worldwide.

As IT professionals increasingly adopt Proxmox VE for open source virtualization, supporting it allows Everpure to meet customer demand where they’re already deploying workloads. This directly responds to the platform’s rapidly growing user base and community momentum by giving sysadmins and engineers native flexibility, reliability, and support in their preferred environment. Proxmox has also recently expanded its Enterprise Support with , removing a deployment obstacle for some organizations.

Adopting a new hypervisor is straightforward compared to addressing underlying application dependencies: predictable low latency, shared storage supporting live migration, zero-downtime maintenance, and reliable disaster recovery. Avoid exchanging licensing challenges for complex storage management.

That’s the gap Everpure and Proxmox are working together to close. Everpure has welcomed Proxmox into its Technology Alliance Partner program, and together we’re collaborating on validation, reference guidance, and deeper integration. You don’t have to wait for any of that to get started, though. The solution is available today, running in production for numerous customers, and it’s fully documented.

Which storage protocols does Proxmox VE support on FlashArray?

Most teams first want to know whether it will work with their existing network. Odds are it will. Everpure has published connection guides and best practices for running Proxmox VE on FlashArray™ over all four common transports:

• Fibre Channel if you’ve got an FC fabric and want to keep it
• iSCSI for Ethernet shops that want familiar block storage
• NVMe/TCP when you want low latency over standard Ethernet, with separate guides for network design, persistent connections, performance tuning, and security
• NFS through FlashArray File Services

Whichever path you choose, FlashArray shows up as shared storage for the entire Proxmox cluster. That’s what makes Proxmox’s built-in features work the way you expect, so VMs and containers can live-migrate between nodes and HA can restart them elsewhere without copying data.

How do you migrate VMs to Proxmox VE with a fallback?

Migration is where most of the anxiety lives, and it’s where FlashArray does some of its most useful work.

If your VMs already sit on an NFS datastore on FlashArray, the Everpure guide walks through a non-destructive migration. You take a snapshot of the VM’s managed directory on the array, then copy it to a new directory. That copy is instant and space-efficient because no data is physically copied. Proxmox attaches the cloned disk, the VM boots, and downtime ends right there. From that point, you can move the disk onto Proxmox-managed storage while the VM keeps running.

The original VM files are never touched. If something looks off, you can power the source VM back on and try again. Having that fallback makes it a lot easier to get sign-off on the first batch of production moves.

There’s also a direct migration method, a guide for moving VMs, a prep checklist, post-migration steps, and a quick reference page with troubleshooting articles for the usual suspects (boot failures, import permissions, and slow virtual disk imports).

How fast does Proxmox VE recover with ActiveCluster? For teams with tight uptime requirements, Everpure validated ActiveCluster™ with Proxmox VE, and the solution validation is worth reading in full.

The lab used six Proxmox VE 9.1 nodes split across two sites, each site backed by its own FlashArray system. ActiveCluster synchronously replicates every write to both arrays before acknowledging it, so your RPO is zero.

Then the team held down the power button on every server at Site A.

About 170 seconds later, the protected VMs were running at Site B, providing an RTO of just under three minutes. Most of that time was the Proxmox fencing timer doing its job, making sure the failed nodes were truly gone before restarting anything. The restart itself took about 10 seconds. There was no storage failover to wait on because the volumes were already active at Site B the whole time.

The guide is also candid about what it takes to make this work. A two-site Proxmox cluster needs a QDevice, a small quorum daemon running on a third host outside both sites. Without it, the surviving site can’t reach quorum and simply waits. With it, the cluster keeps going, and setup is two commands. The guide also covers the 10ms round-trip requirement between arrays and how to tune the watchdog timeout if you want to trade a little caution for faster recovery.

This test was designed to validate recovery from a failed compute cluster. However, the ActiveCluster feature also fully supports storage array failure. If the FlashArray system at Site A had failed instead of the compute nodes, multipath on Site A hosts would have promoted the paths to the Site B FlashArray system automatically, also with no intervention required and with an RPO of zero. ActiveCluster handles both scenarios the same way: transparently.

Note that Everpure is actively working on supporting ActiveCluster with NVMe-oF over TCP and FC. If you’re interested in early access to this feature, please consult your Everpure account team.

How do you back up Proxmox VE on FlashArray?

Proxmox Backup Server (PBS) gives you incremental, deduplicated backups of your VMs and containers. Everpure has published guides for both installing PBS with FlashArray as its storage and for connecting PBS as a backup target for your cluster. Of course, you would not use the same FlashArray system as both source and target, and this gives you the option of using a capacity-optimized array such as FlashArray//C™ as your backup target. In addition, all FlashArray systems offer SafeMode™ Snapshots, which provide immutable, tamper-proof recovery points that remain protected even if administrative credentials are compromised or ransomware attacks backup data. Backup copies can be protected by SafeMode, helping to ensure recovery in the event of a cyber incident.

For enterprises that need the fastest possible system recovery, Everpure™ FlashBlade provides extreme bandwidth and recovery at scale and also offers SafeMode Snapshot protection.

What does FlashArray add to Proxmox VE?

When you step back, the value is pretty simple. Proxmox gives you an open, flexible hypervisor with no lock-in. FlashArray gives you the storage behavior enterprise workloads are used to, with always-on data reduction and thin provisioning at the array, non-disruptive upgrades, encryption at rest, and synchronous replication when you need it. You get to change the layer that was causing you pain and keep the layer that was working.

Is there a native Proxmox storage plugin for FlashArray?

Proxmox and Everpure engineering teams are working toward tighter, array-aware integration. The goal is to bring operations like volume provisioning, array-offloaded snapshots and clones, and data-reduction-aware capacity reporting right into Proxmox workflows, so admins can drive FlashArray from the tools they already use.

One of the capabilities we’re already working on is building a native storage plugin for Proxmox VE, bringing the performance, efficiency, and data services of Everpure FlashArray directly into the Proxmox management experience. The plugin will function mainly as a backend orchestrator, similar to Cinder in OpenStack or a Kubernetes CSI driver.

Many Proxmox deployments still rely on NFS, iSCSI with manual LUN management, or Ceph clusters that add operational overhead. The plugin will close that gap by letting administrators provision, attach, snapshot, and clone VM disks on FlashArray without leaving Proxmox or stitching together scripts.

Why it matters:

• Choice without compromise. Customers evaluating alternatives to their current hypervisor shouldn’t have to give up enterprise-grade storage to do it. Your array, your data services, your choice of platform.
• Simpler operations. Native integration replaces manual volume and host management with consistent workflow orchestration.
• Array-native efficiency. Space-efficient snapshots, clones, and data reduction happen on the array, so VM operations are fast and don’t consume host resources.
• One storage platform across hypervisors. Teams running a mix of environments can standardize on a single storage foundation as they migrate or diversify.

As organizations look for flexibility in their virtualization stack, Everpure is committed to meeting customers wherever they choose to run. The Proxmox plugin is a step toward that: open platform, enterprise storage.

Get started

Everything above is public here. Start with the quick start guide, pick your protocol, and read its best practices before you build out production. If you’re planning a move off your current hypervisor, talk to your Everpure account team. They can help you size the storage side and plan a migration that keeps a fallback in place.

FAQ

• Can Proxmox VE use shared storage with live migration on FlashArray? Yes. FlashArray appears as shared storage to the whole cluster over FC, iSCSI, NVMe/TCP or NFS, so VMs and containers live-migrate between nodes and HA restarts them elsewhere without copying data.
• How fast does Proxmox VE fail over between two sites with ActiveCluster? In Everpure’s validation, protected VMs were running at Site B about 170 seconds after Site A lost power, with zero RPO. Most of that time was the Proxmox fencing timer. A QDevice on a third host and under 10 ms round trip between arrays are required.
• How do you migrate VMs from an NFS datastore on FlashArray to Proxmox VE with a fallback? Snapshot the VM directory on the array and copy it; the copy is instant and space-efficient, the original files are never touched, and you can power the source VM back on if something looks off.
• How do you protect Proxmox VE backups from ransomware? Use Proxmox Backup Server on FlashArray, with a capacity-optimized target such as FlashArray//C, and SafeMode Snapshots for immutable recovery points.