Hello,
I use Proxmox in my homelab and until recently had a very stable setup running Proxmox 5 on 3 nodes with laptop CPU/chipsets (NUC's or the equivalent but in a desktop case). I recently upgraded to Proxmox 6.2 and since then one of the nodes has been regularly experiencing kernel panics where it will freeze completely. Symptoms are as follows:
I have tried installing kdump-tools and using netconsole as well, but I have not been able to get any useful logs when the kernel crash happens.
I have disconnected all connected hard drives except for a Samsung 860 EVO SSD which is used for the OS install and also run memtest to check for other hardware issues but nothing showed up in these tests. The fact that netconsole does not work leads me to think it is somehow related to the networking stack.
In order to test my theory, I wiped this node completely and rather than installing the Proxmox 6.2 ISO directly, I first installed Debian Buster. I found that running just plain Debian Buster with the 4.19 kernel, the machine was completely stable. I then upgraded the node to Proxmox by adding the repo's and following the instructions as per the wiki. I immediately started to encounter the random kernel panic's again.
It seems like there is some issue with the 5.4 Kernel and Realtek 8169 network adapter but so far my search results have not turned up anything useful. Is anyone aware of any issues with Realtek adapters in Proxmox 6 and perhaps how I can overcome this issue?
I use Proxmox in my homelab and until recently had a very stable setup running Proxmox 5 on 3 nodes with laptop CPU/chipsets (NUC's or the equivalent but in a desktop case). I recently upgraded to Proxmox 6.2 and since then one of the nodes has been regularly experiencing kernel panics where it will freeze completely. Symptoms are as follows:
- No response to ping or SSH over the network
- No WebGUI
- No response to keyboard when connected locally.
- The logs do not show anything abnormal apart from at the time of crash, there will be repeated "@^@^@^" characters in the logs, which I understand is due a hard crash.
I have tried installing kdump-tools and using netconsole as well, but I have not been able to get any useful logs when the kernel crash happens.
I have disconnected all connected hard drives except for a Samsung 860 EVO SSD which is used for the OS install and also run memtest to check for other hardware issues but nothing showed up in these tests. The fact that netconsole does not work leads me to think it is somehow related to the networking stack.
In order to test my theory, I wiped this node completely and rather than installing the Proxmox 6.2 ISO directly, I first installed Debian Buster. I found that running just plain Debian Buster with the 4.19 kernel, the machine was completely stable. I then upgraded the node to Proxmox by adding the repo's and following the instructions as per the wiki. I immediately started to encounter the random kernel panic's again.
It seems like there is some issue with the 5.4 Kernel and Realtek 8169 network adapter but so far my search results have not turned up anything useful. Is anyone aware of any issues with Realtek adapters in Proxmox 6 and perhaps how I can overcome this issue?
Last edited: