There’s plenty more reasons than filesystem corruption to use clean shutdown and even just plain UPS without any status information to the host. I haven’t seen a full drive corruption due to power loss ever. Various levels of corruption, yes, but not ‘all the data is gone’ scenario due to filesystem being broken. But I have seen multiple hard drives which have died on host shutdown without any warning (no smart errors or such), after power loss they just refused to start again.

Just today we had multiple short (~30 second) power outages and without UPS that would’ve put considerable strain on my hardware. Gladly those are pretty rare and those were well short enough that my UPS kept the system running, without one that would’ve meant that my hypervisor would’ve crashed and that would’ve resulted on several unclean reboots for the VMs.

Also, back to filesystems, with selfhosting, some kind of hypervisor is often in the picture, so the drive activity is generally much higher than with single workstation and corruption on a single file or some part of filesystem is way more likely. Total data loss is pretty unlikely, but a corrupted VM image can be a pretty big deal, regardless if it’s filesystem or single file which is the main issue.

there’s no standardized mechanism for forcing applications to save

On Linux (and I guess BSD and other *nix variants) there is. SIGTERM is exactly for that situation. Sure, programs can choose to ignore it, but I don’t think I’ve ever seen at least any FOSS software which wouldn’t respect that. Of course the program might crash and cause loss of data, but the mechanism is there.

You’re of course generally correct about modern filesystems being more resilient to power loss, but they’re not immune to problems and also you’ve omitted a large chunk of other reasons why people like to have UPS on their system, with or without monitoring them.

For me personally the hardware is least interesting part of running a homelab. I’d happily run everything on VPS if there was no constraints on money or bandwidth (among other things). But as the limitations are there running things on my hardware is a necessary evil and for that an UPS or two are practically must to have to prolong life span of hardware and prevent data losses. Of course that’s just one box to check on the stack, but when planning for a home server setup for anything I’d say it’s a pretty vital part for the whole setup.

But in the real world it doesn’t really matter if it’s the whole filesystem or an idividual file which is corrupted. Either way the data is compromised and nitpicking that the filesystem is actually just fine doesn’t make any difference.

Modern filesystems can corrupt as well with unclean shutdown. They are just less likely to do so. My server stores photos of our life over couple of decades with other nearly irreplaceable data so having the UPS and clean shutdown on power loss gives at least some confidence that the files stay where they are.

Of course I also run backups of the data, but it’s another layer on data safety setup.

Heat exchangers and/or heat pumps to move energy to water and then pump that to district heating network which then takes care of heating of all kinds of buildings.

Just to add to your points, all of those are already a thing in colder climates. Google DC in Finland is one example. When it’s heavily below zero half of the year it’s just not possible to run open loop systems and there’s plenty of use for the excess heat. Some DCs just donate the waste heat to municipal heating. PR obviously plays role in that, but that’s easy way to get rid of the heat and datacenters don’t need to manage the heating network and fluctuations on demand, so they also benefit from someone who’ll just take the heat away from their ‘hands’. Also sound is automatically better managed since you need decent insulation on the buildings to avoid condensation and other problems which come when temperature differences are pretty huge between inside and outside of the building.

I’m using Hetzner too, only via sync from proxmox backup server instead of rclone.

My brain reads those as ‘thee’, or however that old English should be written (thee, thou, thy…), so the image is that someone pretty pompous Englishman from 1700s with a powdered wig found fediverse.

Plus I tested those on duck.ai just for fun and it has absolutely no trouble of reading the thorns, so it only makes reading harder for other humans, specially for us whose native language isn’t English.

didn’t he break into a network closet and patch directly into their servers?

No. He placed a computer in a unlocked network closet and used that to download a bunch of data from JSTOR (which he had acces to, and a lot of the data is public domain anyways) and kinda-sorta caused DDoS attack against the service. Prosecution then slapped him with a shitload of federal charges which (in my opinion) were largely at least massively exaggerated if not straight made up. Threat of 50 year jail sentence and million(s) in fines then pushed him to take his own life.

Wikipedia has pretty detailed info about him and the whole case.

Old HPEs are pretty cheap and still quite capable devices. My main switch is HP 2530-48G, which is a complete overkill for my needs, but I got lucky at office e-waste pile.

Absolutely. And 1980s was about 20 years ago.

I came to say excactly that.

Immich itself is running on docker, so it’s the same than every other setup. For me it just makes more sense to run that on a separate VM than keeping all the docker containers in a same operating system due to backups, manageability, access control and so on, even if it costs a bit of memory to run a full operating system. Biggest consumers on the host process list are immich-api, postgres-processes and immich. Actual used memory is around 2,5GB for the whole system with ~5GB on buffers/cache.

But I often import hundreds of raw photos at a time from DSLR and/or 4K video footage from my drone and processing those takes quite a lot of memory. With less memory and slower drives the whole thing crashed when importing bigger batches to out of memory erros, but I suppose it could run just fine with even less than 4GB of memory, if the usage pattern is just to upload photos from your cellphone in a relatively slow pace.

It’s a full VM, so OS and docker creates some overhead, but before I upgraded the server it ran on far less resources and specially big imports were painfully slow. Also that’s just the number from proxmox console, so there’s likely some cache included, but stats show that the VM pulls over 10G every now and then with shorter spikes over 20GB of consumed RAM.

As I mentioned, it can run on way less, but it feels a lot more responsive when it has enough resources.

16GB of ram is a bit low for the load you’re planning. That can work, but Nextcloud and Immich are pretty hungry for memory. My immich instance right now is using 7GB of RAM and nextcloud uses a bit over 4. I have the hardware to give them pretty much all the memory they want, so those are not absolute minimum amounts, but I wouldn’t try to run either with less than 6GB allocated memory on my load (few users, about a terabyte of data on both).

With 16GB you can pretty much forget using ZFS, but that’s not a big loss. LVM works just fine and with thin provisioning you can use snapshots too just fine. Bigger SSD/NVME for drive cache would be good. And of course all-SSD setup would be nice, but that gets expensive pretty fast. Second ethernet port might be nice to have, depends on how you’re planning to build your setup. Otherwise that hardware looks pretty good, altough I’d search for something second hand. CPU is most likely a bit overkill for what you need, I have Xeon E5-2620 v3 @ 2.40GHz and it’s running on around 10% load most of the time.

You’re of course correct on that, but Mullvad the company gave (somewhat) clear news release that their company is against far right (or left or any ‘far’ direction) politics. I’m not a customer and I couldn’t care less about Mullvad, but I think it’s worthwhile to keep in mind that company isn’t the same than one of it’s founders.

Of course if that is a breaking point to you, feel free to vote with your wallet and I can very much understand the decision, but there’s very few fully reputable companies around so I can also understand if that’s not a hill to die on.

Don’t send spam, behave nicely, have proper configuration on postfix and dovecot, set up SPF and preferably DKIM, keep your domains reputation clean and so on. You’ll likely still end up on some lists, so you need to manually dig them trough and ask nicely for removal.

Self hosting email has a pretty bad reputation and many will strongly advice against it, but in reality it’s really not that bad. But email as a whole is more complex beast to manage than simple website or running Jellyfin for your family and friends. You need to understand how the mail routing side of things work (SMTP itself, authentication for it encryption included and so on) and also you need to understand how DNS works and how various entries are related to mail transport. Then you’ll likely want at least some kind of spam filtering and so on. And with email you’ll need to have proper backup scenario too so that you don’t lose the likely pretty important data. All of that is perfectly doable and it’s not as difficult as many claim it to be, but there’s a bunch of stuff you need to understand and implement all of them correctly.

There’s various permission errors on the log file, I’d start from those. Also your dovecot configs seem a bit funny to me, usually mails are not stored to user home directories with dovecot, but instead somewhere under /var, likely you’ll want /var/vmail. Logfiles are pretty good with these tools, so pay close attention to those.

And take your time to understand what the config options actually do. Too permissive configurations will get your sever to every spamlist on the world pretty quickly.

Fines are just the cost of doing business.

In theory fines in EU don’t have a ceiling. 10% fine of companys turnover with OpenAI (or any other company running on borrowed money) would be a pretty steep cost of doing business and I don’t think all (if any) of AI-hype entities could actually swallow that.

Obviously that’s just in theory, so far the fines have been a slight slap on the wrist, but mechanisms are in place to give actually meaningful fines for unethical practices. No one will go to jail tho, as EU doesn’t really have the power to jail US citizens.

That style stoves are pretty common around here on older buildings. We have one and it really helps during winter and cuts down the electricity bill, at least as long as we have enough firewood to keep that thing going.

The main material could be smashed/pulverized glass.

As in sand. Glass is recyclable material, it makes no sense to use that for this kind of application since sand is already everywhere. Also at least some flavours of glass can melt on temperatures used on these which I’d imagine could cause problems.

I hear aircrete is an insulation thats already prevalent in europe is this true?

At least here in Finland it is not. It could be, but it’s not really widely available right now. Maybe in the future, but that would need at least some changes on how things are built.