[SOLVED] Proxmox VE Crashes After Starting VM with PCI HBA Passthrough

[LightLuck]

EDIT: I've done all the firmware updates in iDRAC and turned off the rombar setting hostpci0: 0000:02:00,rombar=0. For some reason rombar isn't working with that PCI HBA330. I tried disabling rombar on both hostpci0 and hostpci1 but the issue would come back. This last part is kinda strange and I don't understand. It's been a long day, hope this helps someone out there!

Hi everyone,

I've been trying to get PCI passthrough working with a TrueNAS VM on my Proxmox server, but I'm running into an issue where Proxmox server crashes as soon as I add my HBA card. Below is a summary of my setup and the troubleshooting steps I’ve tried so far:

Hardware & Setup:​

• Proxmox version: 8.2-2
• Server: Dell r730xd
• HBA: Dell HBA330 (non-RAID) SAS Controller
• IOMMU enabled in BIOS (VT-d)
• Two PCI devices being passed through to the TrueNAS VM

Issues:​

• Without the PCI passthrough, the TrueNAS VM boots and runs fine.
• As soon as I try adding one or both PCI device(s) for passthrough, the whole Proxmox crashes instantly and I must reboot the server. This is the message that appears immediately in the server shell:

ext4-fss error (device dm-2): __ext4_find_entry: 1682: inode #1051332: comm task UPID: proxm: reading directory lblock 0

ext4-fs (dm-2): remounting filesystetm read-only

• The dmesg dmesg | grep -e DMAR -e IOMMU logs shows IOMMU entries for my HBA controller.
• No issues booting an Ubuntu Live Server VM, so the issue seems specific to PCI passthrough.

Additional info:​

I tried editing the /etc/pve/qemu-server/<vmid>.conf file with the following instead of adding through gui and the same error is happening.

hostpci0: 0000:02:00.0
hostpci1: 0000:84:00.0

As soon as those 2 lines are deleted, the VM start fine and Proxmox doesn't crash.

I also tried editing /etc/default/grub and modify this line to

GRUB_CMDLINE_LINUX_DEFAULT="quiet intel_iommu=on"

Then edited /etc/modules and added

vfio
vfio_iommu_type1
vfio_pci
vfio_virqfd

Then ran update-grub and rebooted. Issues persist. Since I can see my HBA without those changes, I decided to revert them back since I believe they're not useful in my case.

Everyone wants Plex and Nextcloud back up and running at home, any help is very much appreciated!

 

[AlBundy]

Have you ever found a solution?