Proxmox sometimes doesnt boot

[walix]

Proxmox sometimes will just show a black screen on some boots, I have to restart and it will just boot up normally (it is the primary boot drive). The monitor may not always be on but it is always plugged in to a GPU (rx480 4gb stock). Its a little frustrating that something that I need to boot up reliably doesnt, what could be issue? Could it be that I have to physically press enter on the Debian boot menu to ensure 100% boot success?

 

[leesteken]

Just to make sure: it does not fully boot and you cannot connect to it using ssh or browser from another computer? Or there is only no output when you turn on the monitor?

No output can be caused by console turning off (use kernel parameter consoleblank=0 to keep it always on). It's also possible that the amdgpu driver disables its outputs when the monitor is off and you might need to unplug and replug the display cable or reboot.

I might have experienced something similar but I passthrough both my AMD GPUs and those would not display anything when the monitor(s) were turned off when those VMs started (so I might not be cause by Proxmox but more likely the amdgpu driver). I'm no sure thought if it's the same as your issue.

 

[walix]

Appreciate the feedback! I will have to verify whether or not im able to SSH but i set a VM to automatically start and remote into it but that did not happen so I would assume that SSH would not be working either. I am not sure if this is related but during an install i would sometimes get this issue? Do you think it has something to do with this?

 

[leesteken]

Appreciate the feedback! I will have to verify whether or not im able to SSH but i set a VM to automatically start and remote into it but that did not happen so I would assume that SSH would not be working either.

Likely, but maybe the VM also does not start (maybe for the same reason). Is your Syslog persistent? Then you can look at the logging of a previous failed boot using journalctl.

I've seen Proxmox not start the VMs at boot when it thought it was missing virtual disks (for just deleted VMs) and the time-outs can be quite long. If there is no display then I wouldn't know what was happening because it delayed the start of lots of other services.

I am not sure if this is related but during an install i would sometimes get this issue? Do you think it has something to do with this?

Very unlikely. It's just that the firmware for the latest (or upcoming) AMD GPUs and APUs is not (yet) included. Your RX480 should be fine.

 

[walix]

Regarding the logs I will have to verify it later today as im not home currently, however i can confirm that I have 2 containers that had their storage deleted and i tried to remove them but it doesnt let me coz the HDD that they were on was wiped (and oops moment on my part) but those are containers not VMs. Could this be the reason for that?

 

[leesteken]

Regarding the logs I will have to verify it later today as im not home currently, however i can confirm that I have 2 containers that had their storage deleted and i tried to remove them but it doesnt let me coz the HDD that they were on was wiped (and oops moment on my part) but those are containers not VMs. Could this be the reason for that?

Would not surprise me. There appears to be a connection between systemd and Proxmox containers. If a container fails (to start), systemd claims the system status is degraded. It might also insist on waiting for the drives. I'm not sure how to remove those missing containers completely, sorry.

 

[walix]

No worries, Ill try to figure something out and get back to you. Who know this might help someone one day