It could totally be a hardware issue. I’ve seen something similar happen because the temperature cycles caused the RAM to wiggle out of the slot a bit, but it wasn’t far enough to cause issues until the stick heated up and got pushed a little further out.

I don’t think it’s likely to be a hardware issue unless it’s a faulty drive, but I can’t rule out other components until I can see some errors.

I feel like it’s unlikely to be a hardware issue unless the drive Proxmox is on is dying. SMART is fine, but it’s on an SSD that’s nearly 15 years old, so it very well might be the cause. I need to migrate it to one of my nvme disks eventually.

But, I can boot into the Windows physical disk just fine (instead of booting into Proxmox and running it as a VM). Same with a live ISO on USB.

The fact that it’s perfectly stable after I do get it to boot is why I’m chasing down a software issue first. I also don’t want to think about the cost of having to replace hardware yet. But I’m definitely not ruling it out completely.

I’m going to take the advice of another commenter and turn off quiet in grub so I can get a better look at what’s actually going on under the hood.

That’s exactly what I was looking for, thank you!

I made so many changes following different guides on GPU passthrough on an LXC with a virtual display (never worked) that I’m tempted to rebuild Proxmox from scratch.

I have my router powering my pi, so rebooting the router will reboot the DNS server.

That’s generally true for any end user software.

As an aside, be careful running your laptop with the lid closed. Some need to be open for proper cooling.

Self hosting doesn’t inherently mean your stuff is publicly accessible, though.

It’s not at all uncommon for fines to be built into an SLA

The organizations they were talking about and I was referring to have a global presence

Plus, it’s not significantly more expensive to have a cold standby in a different geographic location in AWS.

So does an outage, but I get that the C-suite can only think one quarter at a time

It’s also silly for the orgs to not have geographic redundancy.

Yeah, if you’re a major business and don’t have geographic redundancy for your service, you need to rework your BCDR plan.

Thank you for this. I’ve been struggling to get wireguard running in hotio containers on my Synology, and this looks like it might be what I need.

Edit: it worked! Had to fiddle around with it because it didn’t like multiple containers in one compose yaml

It still doesn’t matter. Unless you specifically set up a separate network, everything will be in the same network.

Especially if you don’t statically assign your IPs.

Your stated use case doesn’t really require any special routing.

Because best practices for connecting an unsupported operating system are to not do it.

Even if the OS is safe on the day support ends, a critical vulnerability might be found just a few days later. It’s also possible that an exploit has already been found that the bad actor is sitting on it until support ends.

Even if that doesn’t happen, software developers are going to drop support for the OS and vulnerabilities found in those applications could be used to gain ingress.

No amount of “being careful using the Internet” is going to prevent hacking if the system has exploits. If you context a fresh install of XP to the Internet, your system will be compromised in a matter of minutes.

All of mine are sequestered to their own vlan that can only talk to Home Assistant. But they do have ones that use matter now.

Not immediately, no, but saying you can safely continue using it if you follow Internet use best practices is flat out wrong.

I did have my roccat keyboard die on me after less than a year, but all I had to do was send them a picture of it with the cable cut and they sent me a be one.

It’s the smart part that goes obsolete.

Don’t give out infosec advice if you have no fucking clue what you’re talking about.