Quick answer

What this guide helps you do

Understand the Proxmox VE 9.2 baseline: Debian 13.5, kernel 7.0, QEMU 11, LXC 7, ZFS 2.4, HA changes, SDN fabrics and home-lab upgrade priorities.

Quick answer

Proxmox VE 9.2 is the current SmallGrid reference point for new Proxmox installations and newly verified Proxmox instructions.

Released on 21 May 2026, it is based on Debian 13.5 and uses Linux kernel 7.0 as the stable default, with QEMU 11.0, LXC 7.0 and ZFS 2.4. It also adds dynamic HA load balancing, native WireGuard and BGP SDN fabrics, GUI management of custom CPU models, and cluster-wide HA arm/disarm controls.

For a one-node home lab, the underlying platform and kernel changes matter more than the cluster features. Existing 8.x or early 9.x guides should be checked before reusing screenshots, repository instructions, kernel parameters or passthrough workarounds.

Current platform baseline

ComponentProxmox VE 9.2 baseline
Base distributionDebian 13.5 “Trixie”
Stable default kernelLinux 7.0
Virtual machinesQEMU 11.0
ContainersLXC 7.0
ZFS2.4
Default Ceph optionCeph Tentacle 20.2.1
Alternative Ceph optionCeph Squid 19.2.3

The official 9.2 ISO is version 9.2-1. Proxmox advises updating a new ISO installation to the latest available packages after repository configuration.

Check the official download page before installation because the ISO and checksum can change.

What matters on a small one-node server

Linux kernel 7.0

Kernel changes can alter:

  • hardware detection
  • network-interface behaviour
  • IOMMU grouping and passthrough
  • GPU drivers
  • storage-controller support
  • power management
  • USB behaviour

Do not carry an old kernel workaround forward automatically. First check whether the problem still exists on the current Proxmox kernel and whether the old parameter has side effects.

Before changing kernel parameters, record:

pveversion -v
uname -a
lspci -nnk
lsmod
journalctl -k -b --no-pager

QEMU 11 and LXC 7

New virtualisation and container versions can change device models, defaults and compatibility. Existing guests should be backed up before host upgrades, and important VMs or containers should be tested after the update.

For a normal home lab, verify:

  • every guest starts
  • guest networking works
  • storage mounts remain present
  • backup jobs complete
  • passed-through devices remain attached
  • application health checks settle

ZFS 2.4

ZFS updates are relevant only to hosts that actually use ZFS. Do not redesign a working ext4/LVM installation simply because 9.2 contains a newer ZFS release.

For ZFS hosts, review the current Proxmox storage documentation before changing pool features. Enabling newer on-disk features can affect rollback or import compatibility with older environments.

Dynamic HA load balancing

Proxmox 9.2 adds a dynamic mode to the cluster resource scheduler. It can use current node and guest resource utilisation to improve placement of guests managed by the HA stack while respecting HA rules.

This is primarily a multi-node cluster feature. It does not make a single-node home lab more available.

Home-lab relevance:

  • useful for serious multi-node clusters with uneven load
  • not a reason to build a cluster by itself
  • migrations still depend on compatible storage, networking and guest configuration
  • HA still needs enough healthy nodes and a sound quorum design

Do not enable automated movement without first understanding where guests may run and what network or storage they require.

WireGuard and BGP SDN fabrics

Proxmox 9.2 expands software-defined networking with native WireGuard and BGP fabric support, plus BGP/EVPN filtering through route maps and prefix lists.

This can be valuable for:

  • secure links between sites
  • routed lab networks
  • advanced multi-node clusters
  • controlled route distribution

It is not required for a simple server on one home LAN. Tailscale, a straightforward Linux bridge or an ordinary VLAN design may remain easier to operate.

Use the feature only when the networking goal is clear and local console recovery is available.

Custom CPU models in the web interface

Custom virtual CPU profiles can now be created, edited and removed from the Datacenter section of the web interface.

This can help when:

  • a cluster contains different CPU generations
  • live migration needs a controlled common feature set
  • a workload requires specific CPU flags
  • repeatable guest CPU configuration matters

For a single host, the default CPU choices are usually sufficient. Do not hide useful CPU features without a compatibility reason.

Record the previous VM CPU model before changing it and test boot, performance and application licensing where relevant.

HA arm and disarm maintenance control

Proxmox 9.2 can disarm and re-arm the HA Manager cluster-wide while preserving HA resource states. This is intended to prevent unwanted HA actions or fencing during planned maintenance.

This improves a real operational pain point for HA clusters, but it is not a substitute for:

  • a written maintenance plan
  • verified backups
  • quorum awareness
  • checking that every node returns
  • confirming HA state after maintenance

Single-node users without HA resources will see little direct benefit.

What SmallGrid guides need to assume now

New Proxmox guidance should distinguish:

Proxmox VE 8.x
Proxmox VE 9.0 or 9.1
Proxmox VE 9.2 with kernel 7.0

The highest-risk areas for stale instructions are:

  1. installation screenshots and ISO details
  2. Debian repository names and formats
  3. kernel command-line assumptions
  4. IOMMU and GPU passthrough workarounds
  5. LXC template and container behaviour
  6. ZFS feature assumptions
  7. network and SDN screenshots

SmallGrid’s current Proxmox installation guide already uses the 9.2-1 ISO baseline. Other guides that use version-independent GUI or shell concepts remain useful, but readers should compare the interface and installed kernel with the guide’s verification date.

Should you upgrade an existing 8.x host?

Do not perform an in-place major upgrade from 8.x using a short generic checklist.

Proxmox documents an 8-to-9 upgrade path, and 8.4 receives security updates and critical fixes through August 2026. A safe migration requires:

  • current 8.4 packages
  • verified guest backups stored away from the host
  • the official upgrade checker and upgrade guide
  • repository and Ceph planning
  • console access
  • enough time to resolve warnings before reboot

If the host uses Ceph, follow the documented multi-stage Ceph and Proxmox upgrade order exactly.

For an important household server, a clean 9.2 installation with guest restoration may be easier to reason about than an in-place major upgrade, but that depends on storage layout and recovery evidence.

Should you update from 9.0 or 9.1?

Proxmox states that a 9.1 installation can update to 9.2 through Apt or the GUI updater. Before doing so:

  1. update the existing 9.x host fully
  2. read the 9.2 roadmap and package notes
  3. back up important guests
  4. record the current kernel and pveversion -v
  5. check free space and failed services
  6. update through the configured Proxmox repositories
  7. reboot into the intended kernel
  8. verify guests, networking, storage and passthrough

Do not combine the host update with guest OS upgrades or application migrations.

New AMD hardware and installation issues

Kernel 7.0 changes the hardware baseline, but it does not prove that every new AMD chipset, motherboard, NIC or storage controller will work perfectly.

For a reproducible installation problem, record:

Motherboard and revision:
BIOS version:
CPU and chipset:
Network controller:
Storage controller:
Proxmox ISO:
Kernel:
Boot mode:
SVM and IOMMU settings:
Exact symptom:

Check firmware updates and official Proxmox bug or forum evidence before publishing a workaround. Avoid disabling broad security or IOMMU protections based on one anecdotal report.

Verification checklist

  • pveversion -v is recorded.
  • The running kernel is confirmed with uname -a.
  • Important guests have independent backups.
  • Repository configuration matches Proxmox 9 and Debian 13.
  • Guest networking and storage work after reboot.
  • Passthrough devices work, if used.
  • Backup jobs complete.
  • Failed services and kernel errors are reviewed.
  • Cluster and HA state are checked, where applicable.
  • Old kernel workarounds are retained only when still necessary.

Official sources

For a practical 9.2 storage task, continue with How to Add a Second Hard Drive to Proxmox VE 9.2 Safely. If the disk already contains files, use How to Mount an Existing Linux Drive in Proxmox instead.

Recap

Proxmox VE 9.2 is now the correct reference point for new SmallGrid Proxmox work.

For one-node home labs, focus on the Debian 13.5, kernel 7.0, QEMU 11, LXC 7 and ZFS 2.4 platform changes. The new HA and SDN features are significant for clusters, but they do not make a simple server safer by themselves.