PVE with thunderbolt storage kernel 6.8.12-7

rando8374651

New Member
Jan 11, 2025
3
0
1
I have been trying to understand if the bug mentioned with Thunderbolt devices has been resolved by any of the latest kernel updates, I’m on 6.8.12-7.

I know people use Thunderbolt for many things, eGPU, NIC etc. in my case I am wanting to connect a 4 disc DAS over thunderbolt 3 which exposes the discs contained as block devices, to then create a zfs raidz pool and use something like cockpit to then create shares. The DAS I will use has PCIe x4.

I see various reports of people not being able to boot, or devices not recognized, and the solution often proposed is to modify the kernel with thunderbolt.host_reset=0 which would revert to legacy behaviour in this context, or else to pin an older kernel version.

I tried to find some sort of official bug report, all I could see was something from Ubuntu https://bugs.launchpad.net/ubuntu/+source/initramfs-tools/+bug/2078573

However, this seems to basically say that actually the problem people are reporting is not a bug, but that maybe initramfs-tools needs to be modified.

Admittedly my understanding is limited of this issue, but to be clear I don’t want to boot from an external drive; I will use a separate M2 drive as my boot drive and I will be storing all of my VM’s on the M2.

I only want to use this external array to share file content and I’d like to allow access to a couple of LXC’s.

Assuming my hardware is fully supported by the BIOS, can I, on the latest kernel, simply:

  1. Connect a Thunderbolt 3 DAS
  2. Install bolt on the host
  3. Enroll and authorize the device
  4. Expect it to function as intended without kernel modifications?