How is it that if hardware is old, that means it's unmaintained, or that if it's old, it can't have SSDs? Neither of those things are typically inferred from age.

I still maintain old servers, and even my Amiga server has an SSD.

▲

wtallis 3 days ago | parent [-]

If they're running hardware that old, and it's causing them software compatibility problems, then we can infer that their infrastructure is unmaintained, because the cost of moving to newer hardware is so low that the cost of newer hardware could not plausibly be the reason they haven't moved to new hardware. There's dirt cheap used server hardware that would be substantially faster, cheaper to operate, and not have software compatibility issues like this. Money can't be preventing them from using newer hardware.

We don't know for sure the servers don't have SSDs, but we do know that back in the days of server hardware that didn't support SSE4.1, SSDs had not yet displaced hard drives for mainstream storage, so it's likely that servers that old didn't originally ship with SSDs. It's not impossible to retrofit such a server with SSDs, but doing that without upgrading to a more recent platform would be a weird choice.

A server at that age is also going to be harder to repair when something dies, and it's due for something to die. If they lose a PSU it might be cheaper to replace the whole system with something a bit less old. Other components they'd have to rely on replacing with something used, from a different manufacturer than the original, or use a newer generation component and hope it's backwards compatible. Hence why I said using hardware that old would imply their infrastructure is fragile.

But all of this is still just speculation because nobody involved with F-Droid has actually explained what specific hardware they're using, or why. So I'm still not convinced that the possibility of a misconfigured hypervisor has been ruled out.

▲

johnklos 2 days ago | parent [-]

> If they're running hardware that old [...] then we can infer that their infrastructure is unmaintained

You lost me there. One thing has nothing to do with the other.

People have reasons for running the hardware they run. Do you know their reasons? If you do, please share. If not, there's no connection whatsoever between old hardware and unmaintained infrastructure.

Is my AlphaServer DS25 unmaintained? It's very old server hardware.

Is my 1981 Chevette unmantained? It's REALLY old. Can you infer that the fact that I have a car from 1981 means it's unmaintained? I'd say that reasonable people can infer that it's definitely maintained, since it would most likely not still be running if it weren't.

> It's not impossible to retrofit such a server with SSDs, but doing that without upgrading to a more recent platform would be a weird choice.

I don't know where you learned about servers, but no, it's not a weird choice to use newer storage in older servers. Not at all. Not even a little bit. Maybe you've worked somewhere that bought Dell servers with storage and trashed the servers when storage needing upgrading, but that's definitely not normal.

	▲	wtallis 2 days ago \| parent [-]
		> If not, there's no connection whatsoever between old hardware and unmaintained infrastructure. See, this is just you being unreasonable. Yes, we can all imagine why people might keep old hardware around. But your AlphaServer is at best your hobby, not production infrastructure that lots of people and other projects rely on. Nobody's noticing whether or not it crashes. Likewise for your Chevette: nobody cares until it stalls out in traffic, then everyone around you will make the reasonable assumption that it's behind on maintenance. If F-Droid is indeed using ancient hardware, and repeatedly experiencing software failures as a result, then the most likely explanation is that their infrastructure is inadequately maintained. Sure, it's not a guarantee, it's not the only possibility, but it's a reasonable assumption to work with until such time as someone from F-Droid explains what the hell is going on over there. And if there's nobody available to explain what their infrastructure is and why it is showing symptoms of being old and unmaintained, that's more evidence for this hypothesis.