Proxmox 8.1 | Linux 6.5.11-6 | Strange NVMe error prevents boot.

[sondzack]

Hello everyone,
I've run into an issue on my Lenovo m720q tiny, where proxmox fails to reboot. I've had the machine running for over a month, when overnight it randomly decided to stop working. It was unresponsive to ssh and when I plugged in a monitor I was met with a completely frozen session and a wall of nvme errors.

I had assumed there was an issue with the notorious power saving creep that's been haunting linux kernels (and I still assume the issue somewhat persists) but I had already taken the neccessary precautions and set the

nvme_core.default_ps_max_latency_us=0

flag in the linux kernel long ago.

I rebooted the system, but the system would always hang with the following message:

lvm:242 blocked for more than 120 seconds.
Not tainted 6.5.11-6-pve #1
nvme nvme0: Device not ready: aborting reset, CSTS=0x1

Odd, I booted into recovery mode and got the following errors (see attached images)
I check the status of the nvme using the lenovo built in bios-native test and the nvme passed without issue. The bios recognizes the drive without issue.

Any help would be extremely appreciated.
Kind regards.

 

[GzD]

Hi,
I am currently encountering the same issue with Proxmox 7.4-17 (Linux 5.15.143-1).
It seems to have appeared with the last kernel/pve upgrade imo.

 

[Docop2]

read the linux kernel manual. the 0 value is not a setting, specific drive require delay. you will find the formula to apply the correct one for your specific system.

 

[ystald]

same problem
6.5.13-5 kernel
Also tried

nvme_core.default_ps_max_latency_us=0,

but didn't work for me. System hang with

nvme nvme0: Device not ready: aborting reset, CSTS=0x1