I've got an HP EliteDesk 800 G4 Mini that I'm using as my first homelab, running Proxmox VE with a single VM to host my services. Lately, I've been experiencing random hard resets every 1 to 2 days, which takes my services offline and forces me to restart the VM manually. There haven't been any kernel panics, out of memory issues, or I/O errors—just a log entry showing 'crash' when I check what happened after rebooting. Here are my specs:
1. HP EliteDesk 800 G4 Mini
2. Intel i7-8700T
3. 64GB RAM (2x32GB Samsung DDR4 2666 SODIMM, non-ECC)
4. NVMe 1: SK Hynix PC611 256GB (OS)
5. NVMe 2: Samsung 990 PRO 1TB (firmware 5B2QJXD7)
6. ZFS for the root filesystem
7. 90W OEM HP power brick
I'm currently running:
1. Proxmox VE (based on Debian trixie)
2. A Debian VM with:
* WireGuard
* Gitea (Docker + Postgres)
* Joplin Server
3. Light homelab services; nothing too taxing
So far, I've checked everything and confirmed no out-of-memory events, kernel panic logs, hardware error logs, NVMe SMART status is clear, temperatures are within normal range, and ZFS ARC is tiny (around 250MB). However, the unsafe shutdowns on NVMe increment indicate possible abrupt power losses.
Given that the power brick is a 90W OEM HP adapter, I'm beginning to suspect it might be the power supply, especially combined with running 64GB of RAM, two NVMe drives, ZFS, and VMs.
Before I replace the power adapter, is there anything else I should check? Also, has anyone else had similar instability with 64GB in this model?
5 Answers
The 90W should normally be enough, but if you’re maxing out the resources, it could be borderline. If possible, test with a higher wattage PSU to see if that resolves the issue. It’s also a good idea to get a UPS; unstable power can really mess things up.
This might sound random, but have you checked to see if there are any settings related to failover or HA fencing that you might have accidentally toggled? Those can sometimes cause these kinds of resets.
A quirky potential issue could be with the Intel NIC you didn't mention. Disabling checksum offloading often helps with stability problems related to network traffic.
It does seem like a hardware problem to me. I haven't seen anything too alarming in your setup, but those power issues can really mess things up. You should look at swapping out your power supply just to eliminate that as a potential source of the problem.
Yeah, I agree. It couldn't hurt to try a different power adapter first; power delivery can be a sneaky cause of resets.
Given that it's a bit older, you could run some diagnostics that HP has built-in. Maybe stress test it to see if you can replicate the resets. And definitely keep an eye on your event logs.
Good call on the UPS! It's always smart to protect your equipment.