Proxmox seemingly shutting itself down, then repeatedly powercycling until I manually switch the PSU off.

faustumOssis

New Member
Nov 8, 2024
3
0
1
I'm currently running an E5-2697A v4 single CPU, on a Machinist x99 motherboard (not sure which specific model, but does it really matter?) and 32gb of ddr4 non-ecc memory. It runs fine for like 15-30mins and then sends the shutdown signal on it's own. I have 3 theories on why this is happening:

1. The CPU is overheating. I find this option unlikely because I'm using this with a 2-tower deepcool cooler (which I bought right before the ban, phew.)
2. The PSU does not have enough juice for the CPU. I find this option likely because I'm just using a 500w thermaltake smart for initial testing purposes.
3. This is divine punishment for buying an aliexpress x99 board.

Are any of these possibly my issue? Maybe something else? The CPU is supposedly designed to take ~145W but we all know how greedy Intel CPUs are. I don't even really have anything else connected to the power, besides a 2gb GT 710 (literally 20w) for basic display output, a laptop sized hard disk (like i said before, just for testing), and 4 fans. I've tried the whole disabling power button actions in the config and whatnot, still happens.
 
Last edited:
Some other general pointers that may help you:

Have you checked the RAM? It is possibly faulty - and only clears itself correctly when all power is removed.

What do your logs show - just before shutdown.

To generally check - if your problem is at all Proxmox-related, try with another plain Linux distro & see how much uptime you get there.

Try changing the HD for something else. Maybe even try just running a live Linux distro from a USB flash drive & see what happens.
 
  • Like
Reactions: faustumOssis
Some other general pointers that may help you:

Have you checked the RAM? It is possibly faulty - and only clears itself correctly when all power is removed.

What do your logs show - just before shutdown.

To generally check - if your problem is at all Proxmox-related, try with another plain Linux distro & see how much uptime you get there.

Try changing the HD for something else. Maybe even try just running a live Linux distro from a USB flash drive & see what happens.
Thanks for the reply, I've previously run memtest86 with no errors but I'll run it again. Could you remind me how to find the logs? I'll test with another distro as well.
 
journalctl --since <DATETIME> --until <DATETIME>
Looks like it's termination on signal 15, which from searching around probably means it's running out of memory? I have 32gb installed at the moment, and am only using 16 for my one VM that i've been testing with. Does this imply that my memory is having errors? And while I'm typing this it went off again, while running absolutely nothing besides itself.