Category: Proxmox

Proxmox – Network Bond Setup (Active-Standby)

February 25, 2026 / bytesmith17

This guide explains how to configure an active-backup (active-standby) network bond in Proxmox using two physical NICs. This setup provides network redundancy without requiring LACP or special switch configuration.

In active-backup mode, only one interface is active at a time. If the primary NIC fails, the secondary automatically takes over.

⚠️ This guide assumes Proxmox VE 7.x or 8.x using ifupdown2 (default on modern installs).

1. Edit the Network Interfaces File

Open the Proxmox network configuration file:

nano /etc/network/interfaces

2. Configure the Bond Interface (bond0)

Add or modify your configuration as follows:

auto lo
iface lo inet loopback# Bonding interface
auto bond0
iface bond0 inet manual
    bond-slaves eno1 eno2
    bond-mode active-backup
    bond-miimon 100
    bond-primary eno1

Explanation

bond-slaves eno1 eno2
Replace with your actual NIC names (check using ip link)
bond-mode active-backup
Enables failover mode (no LACP required)
bond-miimon 100
Checks link state every 100ms
bond-primary eno1
Optional — sets preferred primary interface

3. Create the Proxmox Bridge (vmbr0)

Attach the bond to a Proxmox bridge for VM and container networking:

# Virtual bridge using the bonded interface
auto vmbr0
iface vmbr0 inet static
    address 192.168.1.100/24
    gateway 192.168.1.1
    bridge-ports bond0
    bridge-stp off
    bridge-fd 0

Replace:

192.168.1.100/24 → Your Proxmox host IP
192.168.1.1 → Your gateway

4. Apply the Network Configuration

Instead of rebooting, safely reload networking:

ifreload -a

⚠️ If working remotely over SSH, be cautious — misconfiguration may disconnect you.

5. Verify the Bond Status

Check bond details:

cat /proc/net/bonding/bond0

You should see:

Bonding Mode: active-backup
Currently Active Slave: eno1 (or eno2 if failover occurred)
MII Status: up

To test failover:

Unplug the primary NIC and confirm the secondary becomes active.

How Active-Backup Bonding Works

Only one NIC carries traffic at a time
Automatic failover if link drops
No switch configuration required
Ideal for homelabs and unmanaged switches

This is different from LACP (802.3ad), which requires switch support and provides load balancing.

Optional: Identify Your Network Interfaces

If unsure of interface names:

ip link

Common names include:

eno1
eno2
enp3s0
enp4s0

Use those values in bond-slaves.

Troubleshooting Tips

If the bond does not come up:

Ensure both NICs are connected
Confirm correct interface names
Check for indentation errors in /etc/network/interfaces
Restart networking if necessary:

systemctl restart networking

Tested On

Proxmox VE 8.x
Dual NIC Intel-based systems
Standard home / unmanaged switches

Proxmox – Enable Intel iGPU Hardware Acceleration

February 25, 2026 / bytesmith17

This guide explains how to enable Intel integrated GPU (iGPU) hardware acceleration on a Proxmox host and pass it through to an LXC container (such as Emby) using VAAPI.

Tested on:

Proxmox VE 7.x / 8.x
Intel N100 / modern Intel iGPUs
Debian-based LXC containers

1. Install Intel GPU Drivers on the Proxmox Host

Proxmox already includes the kernel driver. You only need the user-space packages.

Update the Host

apt update && apt full-upgrade -y

Install Intel Media Packages

apt install -y \
  intel-media-va-driver \
  i965-va-driver \
  vainfo \
  intel-gpu-tools

On newer Proxmox kernels, intel-media-va-driver (iHD driver) is the important package.
i965-va-driver can remain installed for compatibility.

2. Verify the iGPU Is Detected

Check that the DRM device exists:

ls -l /dev/dri

You should see:

/dev/dri/card0
/dev/dri/renderD128

If you do not see these, check BIOS settings:

iGPU enabled
Primary display set to Auto or iGPU
No headless-disable options enabled

3. Test VAAPI on the Host

Run:

vainfo

Expected output should reference:

Intel iHD driver
H.264
HEVC
VP9
AV1 decode (supported on N100)

If vainfo works → the host configuration is complete.

4. Prepare the LXC Container for GPU Access

Edit your container configuration:

nano /etc/pve/lxc/<CTID>.conf

Add the following lines:

# Allow GPU devices
lxc.cgroup2.devices.allow: c 226:* rwm# Bind mount DRM devices
lxc.mount.entry: /dev/dri dev/dri none bind,optional,create=dir# Required for iGPU access
lxc.apparmor.profile: unconfined
lxc.cap.drop:

For trusted internal containers (like Emby), this does not meaningfully weaken host security.

Restart the container after saving.

5. Install Intel Drivers Inside the Container

Enter the LXC and update:

apt update && apt upgrade -y

Install VAAPI support:

apt install -y \
  intel-media-va-driver \
  i965-va-driver \
  vainfo \
  ffmpeg

6. Verify GPU Access Inside the Container

Check:

ls -l /dev/dri

You should see:

card0
renderD128

Test VAAPI:

vainfo

If this works → GPU passthrough is successful.

7. Configure Emby for Hardware Transcoding

In Emby Admin:

Dashboard → Playback → Transcoding

Set:

Enable hardware acceleration
Acceleration API: VAAPI
Enable hardware decoding
Enable hardware encoding
Enable tone mapping (if using HDR → SDR)

Optional (recommended):

Set the transcode temp directory to:

/tmp/emby-transcode

Or use a fast NVMe-backed dataset.

8. Confirm Hardware Transcoding Is Active

Start a file that forces a transcode (not direct play).

Then on the Proxmox host run:

intel_gpu_top

You should see activity under:

Video
Render

CPU usage should remain low.

If GPU usage increases → hardware acceleration is working correctly.

Proxmox – How to Mount a Synology iSCSI LUN to Proxmox

February 25, 2026 / bytesmith17

This guide explains how to connect a Synology iSCSI LUN to a Proxmox host and mount it for use with containers (LXC).

⚠️ Important: If using a standard filesystem such as ext4, only one node should mount the LUN at a time. Mounting the same LUN on multiple nodes without a cluster filesystem can cause data corruption.

1) Discover the iSCSI Target

Run the discovery command from your Proxmox node:

iscsiadm -m discovery -t sendtargets -p NAS_IP

Example:

iscsiadm -m discovery -t sendtargets -p 192.168.x.x

You should see output similar to:

192.168.x.x:3260,1 iqn.YYYY-MM.com.synology:target-name

2) Get the Proxmox Initiator Name

You’ll need to add this to the Allow List in Synology DSM.

cat /etc/iscsi/initiatorname.iscsi

Output will look like:

InitiatorName=iqn.YYYY-MM.org.debian:unique-id

Copy everything from iqn onwards and add it to your Synology iSCSI target permissions.

3) Log In to the iSCSI Target

iscsiadm -m node --login

If successful, the session will establish without errors.

To log out from another node (if required):

iscsiadm -m node --logout

4) Confirm the Disk Appears

Check block devices:

lsblk

Or confirm the by-path entry:

ls -l /dev/disk/by-path | grep iscsi

You should see something similar to:

ip-192.168.x.x:3260-iscsi-iqn...-lun-1 -> ../../sdb

5) Format the LUN (If Required)

If this is a new LUN:

mkfs.ext4 /dev/sdX

Replace sdX with the correct device.

6) Create a Mount Point

mkdir -p /mnt/storage

7) Add to /etc/fstab

Edit:

nano /etc/fstab

Add:

/dev/disk/by-path/ip-NAS_IP:3260-iscsi-iqn.TARGET_NAME-lun-1  /mnt/storage  ext4  _netdev,noatime,discard  0  2

Reload systemd and mount:

systemctl daemon-reload
mount -a

8) Ensure iSCSI Auto-Starts on Boot

iscsiadm -m node --op update -n node.startup -v automatic

Verify:

iscsiadm -m node

9) Add Mount to an LXC Container

Edit the container config file:

nano /etc/pve/lxc/CTID.conf

Add at the bottom:

mp0: /mnt/storage,mp=/mnt/storage,backup=0

Restart the container after saving.

Optional Security Recommendation

Consider enabling CHAP authentication on your Synology iSCSI target to prevent unauthorised access, even on internal networks.

Proxmox – Emby Intel GPU passthrough

October 1, 2024 / bytesmith17

October 1, 2024

Important: This is for use with Debian 12 (BookWorm) / Proxmox (8.2.7).

This guide is for the setup of an Intel iGPU or dedicated GPU passthrough on Proxmox between an LXC / CT linux system.

Run these commands on both the proxmox host and Emby CT. I have this method working on successfully on both a N100 iGPU and Sparkle Intel Arc A310 ECO

Step 1 – Adding driver download sources
nano /etc/apt/sources.list

add the following

non-free firmwares deb http://deb.debian.org/debian bookworm non-free-firmware

non-free drivers and components deb http://deb.debian.org/debian bookworm non-free

Step 2 – Installing the GPU Driver

Install the following
apt update && apt install intel-media-va-driver-non-free intel-gpu-tools

Step 3 – Confirm GPU’s major number’s for fb0 and renderD128

Important: Note down your major numbers from your output my example numbers may not be the same as yours and if this are incorrect the gpu will not work.

cd /dev

Use the following command to list all files and folders in the current directory

ls -lah

Make a note of the numbers in bold for fb0 (If it exists)

crw-rw---- 1 root video 29, 0 Aug 30 21:19 fb0

Navigate to the dri folder using the following command

cd /dev/dri

ls -lah

Make a note of the the number in bold for renderD128

crwxrwxrwx 1 root render 226, 128 Aug 30 21:19 renderD128

Step 4 – Amend Proxmox CT/LXC config to allow access to the GPU

Now we will add the numbers that were noted down to our CT config on proxmox for Emby – Important amend the command below with your CT’s ID number shown the proxmox webgui (Example Below) Mine is 3006

Now open your CT’s config files on your Proxmox host.

nano /etc/pve/lxc/3006.conf

Add the following lines to the file and save using CTRL + O and then enter.
lxc.cgroup2.devices.allow: c 29:* rwm lxc.cgroup2.devices.allow: c 226:* rwm lxc.mount.entry: /dev/fb0 dev/fb0 none bind,optional,create=file lxc.mount.entry: /dev/dri dev/dri none bind,optional,create=dir lxc.apparmor.profile: unconfined

Step 5 – Adding Persistent GPU Permissions

Important: Giving 777 permissions to your GPU is not considered best practice from a security standpoint. While this approach works, there are more secure methods you can follow. However, I understand the risks and accept them on my system. do this at your own risk…

Adding Persistent GPU Permissions on proxmox host so the GPU is accessible to the CT after reboot.

nano /etc/udev/rules.d/99-renderD128-permissions.rules KERNEL=="renderD128", MODE="0777"

udevadm control --reload-rules

udevadm trigger

How to Test?

Run the following command on your proxmox shell (This is basically task manager for your GPU)

intel_gpu_top

Step 6 – LXC/CT GPU Driver installation

Repeat Steps 1 and 2 inside your LXC/CT/Emby Shell

How to Test GPU in Emby?

Login to Emby and navigate on the admin webpage to Server > Transcoding. If you see your GPU and a list of Hardware Decoders its likely working. Test forcing something to Trancode by changing the bitrate and monitor the CPU usage and using intel_gpu_top to see if there is any ffmpeg processes running on the GPU

If your not seeing anything on this page I suggest giving your server a little nap (aka, a reboot) see if that makes the GPU driver jumps into life.

Proxmox – Persistent GPU Permission

October 1, 2024 / bytesmith17

If you are having issues with GPU transcoding in emby and your GPU permission keep changing after reboot use the following commands to fix the permission.

Run the following command inside your proxmox shell.

nano /etc/udev/rules.d/99-renderD128-permissions.rules

Add the following to the file

KERNEL=="renderD128", MODE="0777"

CTRL + O then Enter to save the changes to the file.

Now run the following command

udevadm control --reload-rules

Run the following command

udevadm trigger