Recently been trying to get a server node set up, but I keep getting random crashes (no specific amount of time passes) while the machine is basically idle.
I've also tried adjusting some configs in the /etc/default/grub file, first setting a max cstate of 1, then later after a subsequent crash updating the amd_pstate to be disabled:
GRUB_CMDLINE_LINUX_DEFAULT="quiet processor.max_cstate=1 amd_pstate=disable"
That didn't seem to solve the problem, so I installed the kdump tool to try to get some information from the crashes because I wasn't getting much from the regular logs. Got something that looked like this. I didn't type out all of the strings since I can't easily copy/paste it from this machine (SSH into it through a different machine than what I'm typing this from)
pid_max: default: 32768 minimum: 301
iommu: Default domain type: Translated
NetLabel: unlabeled traffic allowed by default
PCI: CLS 64 bytes, default 64simple-framebuffer simple-framebuffer.0: [drm] Registered 1 planes with drm panic
RAS Correctable Errors collector initialized
nvme nvme0: 12/0/0 default/read/poll queues
systemd[1]: Queued start job for default target graphical.target
softdog: initialized. soft_noboot=0 soft_margin=60 sec soft_panic=0 (nowayout=0)
platform regulatory.0: direct firmware load for regulatory.db failed with error -2
traps: pveproxy worker[1457] general protection fault ip: <number> sp: <number> error: 0 in libcrypto.so.3[<string of characters>]
OOps: segment-related general protection fault: d3e4 [#1] PREEMPT SMP NOPTI
default_idle_call+0x29/0x100
I'm fairly new to reading this kind of information, so I'm not sure what (if anything) is useful here. This is also where libcrypto showed up.
It's actually not the first time I tried getting this running. I had a different motherboard for a while, but decided to get another when I exhausted pretty much every config change and test I could find by scouring the forums. It's difficult to diagnose the motherboard. It was purchased new so there shouldn't have been a problem, but I figured swapping it with another new one might be the best way to check that. Although, with the previous motherboard it was crashing maybe once per day. With this current motherboard it's crashing about once an hour, give or take.
Any help would be much appreciated.
- CPU: AMD Ryzen 5 4500
- Motherboard: ASRock B550M Pro4
- RAM: 16GB DDR4
- Also has an NVidia GPU in it. Needed it for OS installation, which just meant I had to include 'nomodeset' during installation. Figured I could maybe use it later if I need a container or VM that has a GUI; Gigabyte GV-N710D3-2GL REV2.0
- Fresh Proxmox install, no VMs or containers running yet
- Proxmox 9.1.9
- Kernel: 6.14.8-2-pve (I rolled this back from 7.0.0-3-pve to see if it was a problem with the newer kernel; no change)
- Debian 13 Trixie base
- OpenSSL 3.5.0 (included because I ran into it at one point through libcrypto)
I've also tried adjusting some configs in the /etc/default/grub file, first setting a max cstate of 1, then later after a subsequent crash updating the amd_pstate to be disabled:
GRUB_CMDLINE_LINUX_DEFAULT="quiet processor.max_cstate=1 amd_pstate=disable"
That didn't seem to solve the problem, so I installed the kdump tool to try to get some information from the crashes because I wasn't getting much from the regular logs. Got something that looked like this. I didn't type out all of the strings since I can't easily copy/paste it from this machine (SSH into it through a different machine than what I'm typing this from)
pid_max: default: 32768 minimum: 301
iommu: Default domain type: Translated
NetLabel: unlabeled traffic allowed by default
PCI: CLS 64 bytes, default 64simple-framebuffer simple-framebuffer.0: [drm] Registered 1 planes with drm panic
RAS Correctable Errors collector initialized
nvme nvme0: 12/0/0 default/read/poll queues
systemd[1]: Queued start job for default target graphical.target
softdog: initialized. soft_noboot=0 soft_margin=60 sec soft_panic=0 (nowayout=0)
platform regulatory.0: direct firmware load for regulatory.db failed with error -2
traps: pveproxy worker[1457] general protection fault ip: <number> sp: <number> error: 0 in libcrypto.so.3[<string of characters>]
OOps: segment-related general protection fault: d3e4 [#1] PREEMPT SMP NOPTI
default_idle_call+0x29/0x100
I'm fairly new to reading this kind of information, so I'm not sure what (if anything) is useful here. This is also where libcrypto showed up.
It's actually not the first time I tried getting this running. I had a different motherboard for a while, but decided to get another when I exhausted pretty much every config change and test I could find by scouring the forums. It's difficult to diagnose the motherboard. It was purchased new so there shouldn't have been a problem, but I figured swapping it with another new one might be the best way to check that. Although, with the previous motherboard it was crashing maybe once per day. With this current motherboard it's crashing about once an hour, give or take.
Any help would be much appreciated.
Last edited: