Hello,
I’m encountering an issue with NVMe passthrough on my Ubuntu Desktop VM in Proxmox. The problem started recently, and the NVMe controller (Micron/Crucial T700 SSD) fails to initialize within the VM, dropping into the initramfs prompt. Interestingly, this same setup worked fine before. My Windows 11 VM, which also uses NVMe passthrough (Samsung 990 Pro SSD), continues to work without issues. Below are the details of my setup and the troubleshooting steps I’ve tried so far. As you will see, I searched a lot but could not solve it alone. I am new to Proxmox and appreciate any help. I suspect this is related to Proxmox passthrough or IOMMU configuration, but I’m not sure what’s causing the failure. Are there any additional configurations or diagnostics I can try to resolve the issue? Here are some info and what I have done so far:
Device mappings:
Ubuntu VM qemu configuration
I’m encountering an issue with NVMe passthrough on my Ubuntu Desktop VM in Proxmox. The problem started recently, and the NVMe controller (Micron/Crucial T700 SSD) fails to initialize within the VM, dropping into the initramfs prompt. Interestingly, this same setup worked fine before. My Windows 11 VM, which also uses NVMe passthrough (Samsung 990 Pro SSD), continues to work without issues. Below are the details of my setup and the troubleshooting steps I’ve tried so far. As you will see, I searched a lot but could not solve it alone. I am new to Proxmox and appreciate any help. I suspect this is related to Proxmox passthrough or IOMMU configuration, but I’m not sure what’s causing the failure. Are there any additional configurations or diagnostics I can try to resolve the issue? Here are some info and what I have done so far:
- The Ubuntu VM does not boot and drops into the initramfs shell.
- Attaching a live Ubuntu ISO to the VM also fails to recognize the NVMe controller.
- When booting the same Ubuntu installation bare-metal (bypassing Proxmox), it works perfectly without any issues, indicating the disk and controller are functional.
- Upgrading the Proxmox kernel from 6.8.12-4-pve to 6.11.0-1-pve did not resolve the issue.
- Adding the following to /etc/default/grub on the Proxmox host did not resolve the issue:
pci=nomsi nvme_core.default_ps_max_latency_us=0 nvme_core.io_timeout=30 - I removed GPU passthrough for Ubuntu VM but nothing changed.
Environment Information
Hardware
- Motherboard: MSI X670E
- CPU: AMD Ryzen 9 7950X3D 16-Core Processor
- NVMe Devices:
- Micron/Crucial T700 SSD (used for the Ubuntu VM)
- Samsung 990 Pro SSD (used for the Windows 11 VM)
Proxmox Host
- Proxmox Version: 8.3
- Kernel: 6.11.0-1-pve
- EMU emulator version: 9.0.2 (pve-qemu-kvm_9.0.2-4)
Passthrough Configuration
Devices specified in /etc/modprobe.d/vfio.conf:
Code:
options vfio-pci ids=10de:2684,10de:22ba,c0a9:5419,1022:43f7,1b21:3241Device mappings:
- GPU Passthrough: NVIDIA RTX 4090 (10de:2684) with Audio Controller (10de:22ba)
- NVMe Passthrough: Micron/Crucial T700 SSD (c0a9:5419)
- USB Controllers: AMD USB 3.2 Controllers (1022:43f7), ASMedia ASM3241 (1b21:3241)
Code:
GRUB_CMDLINE_LINUX_DEFAULT="quiet amd_iommu=on iommu=pt pci=nomsi nvme_core.default_ps_max_latency_us=0 nvme_core.io_timeout=30"Ubuntu VM qemu configuration
Code:
cat /etc/pve/qemu-server/100.conf
agent: 1
bios: ovmf
boot: order=ide2;hostpci0;net0
cores: 8
cpu: host
efidisk0: local-btrfs:100/vm-100-disk-0.qcow2,efitype=4m,pre-enrolled-keys=1,size=528K
hostpci0: 0000:02:00,pcie=1,rombar=0
hostpci1: 0000:10:00,pcie=1,rombar=0
hostpci2: 0000:12:00,pcie=1,rombar=0
ide2: none,media=cdrom
machine: q35,viommu=virtio
memory: 80000
meta: creation-qemu=9.0.2,ctime=1729457387
name: achibuntu
net0: virtio=BC:24:11:22:07:5F,bridge=vmbr0,firewall=1
numa: 1
ostype: l26
scsihw: virtio-scsi-single
serial0: socket
smbios1: uuid=245fadff-0de6-4564-9dc6-9b4d256f5007
sockets: 1
vga: virtio
vmgenid: 6988ea9b-147d-4f84-83a4-c6e2aeddf6f8Ubuntu VM Information
- Device Details:
Code:sudo lspci -nn 01:00.0 Non-Volatile memory controller [0108]: Micron/Crucial Technology T700 NVMe PCIe SSD [c0a9:5419] - dmesg error output:
Code:sudo dmesg | grep nvme [ 0.839065] nvme 0000:01:00.0: Adding to iommu group 11 [ 0.841911] nvme nvme0: pci function 0000:01:00.0 [ 62.596278] nvme nvme0: I/O tag 4 (0004) QID 0 timeout, disable controller [ 62.598393] nvme nvme0: Device not ready; aborting shutdown, CSTS=0x5 [ 62.606297] nvme nvme0: Identify Controller failed (-4) [ 62.612397] nvme: probe of 0000:01:00.0 failed with error -5 - Modules Loaded:
Code:sudo lsmod Module Size Used by nvme 61440 0 nvme_core 208896 1 nvme virtio_gpu 94208 1 ahci 49152 1 xhci_pci 24576 0 libahci 53248 1 ahci virtio_dma_buf 12288 1 virtio_gpu xhci_pci_renesas 20480 1 xhci_pci nvme_auth 28672 1 nvme_core - Ubuntu VM is stuck in initramfs prompt:
Code:BusyBox v1.36.1 (Ubuntu 1:1.36.1-6ubuntu3.1) built-in shell (ash) (initramfs) blkid (initramfs) exit Gave up waiting for root file system device. Common problems: - Boot args (cat /proc/cmdline) - Check rootdelay= (did the system wait enough?) - Missing modules (cat /proc/modules; ls /dev) ... ALERT! /dev/disk/by-uuid/longIDxxxxxx does not exist. Dropping to a shell! (initramfs)_
Last edited:
