External 2.5" HDDs connected via USB for longer term bulk storage and using it as a NAS, a smaller internal NVME SSD for the OS and a larger one (but SATA, so slower) for the directory were torrents go to.
The different drive performances fit my usage pattern just fine whilst optimizing price per GB.
External 3.5" would be cheaper for bulk storage but the 2.5" are a leftover from when I was more constrained in terms of physical space.
Well, the N100 does have a lot more breathing space in terms of computing power, so it’s maybe a better bet for something you want to use for a decade or more, and that remote control I linked to above does work fine, except for the power button (which will power your Linux off but won’t power it back on).
I actually tried an Android TV Box (which is really just and SBC in the same range of processing power as the Pi) for this before going for the Mini PC and it was simply not as smooth operating.
That Mini-PC has enough computing power room (plus the right processing extensions) that I can be torrenting over OpenVPN on a 1Gb/s connection whilst watching a video from a local file and it’s not at all noticeable on the video playback.
Kodi install instructions are here
I don’t use docker, I use lubuntu with normal packages. So for example Kodi is just installed from the Team Kodi PPA repository (which, granted, is outdated, but it works fine and I don’t need the latest and greatest) and just set it up to be auto-started when X starts so that on the TV it’s as if Kodi is the interface of that machine.
Qbittorrent is just the server only package (qbittorrent-nox) which I control remotelly via its web interface and the rest is normal stuff like Samba.
After the inital set up, the actual linux management can be done remotelly via ssh.
That said, LibreELEC is a Linux distro which comes with Kodi built-in (it’s basically Kodi and just enough Linux to run it), so assuming it’s possible to install more stuff in it might be better - I only found out about it when I had my setup running so never got around to try it. LibreELEC can even work in weaker hardware such as a Raspberry Pi or some of its clones.
Also you can get Kodi as a Flatpak which works out of the box in various Linux distros so if you need the latest and greatest Kodi plus a full-blown Linux distro for other stuff you might do the choice of distro based on supporting flatpack and being reasonably lightweight (I actually originally went for Lubuntu exactly because it uses a lightweight Window Manager and I expected that N100 mini-pc to need it, though in practice the hardward can probably run a lot more heavy stuff than that, though lighter stuff means the CPU load seldom goes up significativelly hence the fan seldom turns on and so the thing is quiet most of the time and you only hear the fan spinning up and then down again once in a while even in the Summer).
As for docker, there are a lot of instructions out there on how to install Kodi with Dockers, but I never tried it.
Also you might want to get a remote like this, which is a wireless remote with a USB adapter, not because of the air-mouse thing (frankly, I never use it) but simply because the buttons are mapped to exactly the shortcuts that Kodi uses, so using it with Kodi in Linux is just like using a dedicated remote for a TV Media Box - in fact all those thinks are keyboard shortcuts (that remote just sends keypresses to the PC when you press a button) and they keyboard shortcuts for media players seem to be a standard.
It really depends on what you’re doing with it and on what old PCs you have available.
I have an N100 Mini-PC at home in my living room connected to my TV which is both a home server and a TV-Box using Kodi (I even have a remote for it).
Having modern image and video decoding in hardware is pretty useful when I’m using it as a TV Box (there is zero stutter with it), whilst the rest of the time the thing mostly sits doing some low CPU-intensive server tasks (mainly torrenting and SMB server stuff).
Also, it’s a small box that fits fine on my TV stand without standing out and runs silent pretty almost all of the time.
Further, I don’t have any low power consuming old PCs around - the best are some chunky old notebooks, the rest are old gaming PCs which eat more power idle than the mini PC does at full load - and even the notebooks aren’t that low power as all that.
Mind you, for many years I used an old Asus EEE PC (a very small notebook running Linux) as home file server (with external HDs) and had a separated dedicated hardware TV Media Server box playing files from it, but eventually that PC stopped working and I found out I could just use my Router as a file server.
Last but not least, judging for how long I kept using my TV Media Server boxes (which over almost 2 decades I had 2 different ones and which as dedicated hardware could not easilly be upgraded when new video compression standards came out) 10+ years is definitelly my time-frame for using that Mini-PC.
All this to say that you should consider using old hardware, especially if you have some around and it’s task appropriate (like I did before using an old Asus EEE PC as a home file server), but also take in account what you’re going to do it and consider if new hardware won’t be better over the timespan you will likely be using it and if the being able to get a more task appropriate form factor (like how having a little box-size Mini PC lets me have it in my living room on a TV stand next to my TV and my fiber router) is worth it.
In summary, before you get hardware you should ponder a bit about what you intend to do with it before you decide what to get, don’t be afraid of using stuff you already have and also don’t be afraid to get new stuff if it’s actually justified by hardnosed reasons rather than merely some variant of the “new stuff smell” psychological effect when buying new.
“I have deemed my own work to be ‘great’ and that’s the only professional feedback I take”
The more services you have depending on a 3rd party which can do whatever the fuck they want, either directly or by changing the rules when the feel like it (i.e. not bound by rules they cannot change, such as root DNS providers are) and then doing it, the less your system is actually self-hosted, IMHO.
For me the whole point of self-hosting is exactly being as independent as possible of 3rd parties that can just fuck you up, be it on purpose (generally for $$$) or because they go bankrupt and close their services.
This is why I’ve actually chosen to run Kodi on my home server that doubles down as TV Box even though I can’t easilly use it from anywhere else (it’s possible but it involves using a standalone database that is then shared, which can only be safelly done through customly setup ssh pipes) rather than something like Plex.
It’s kinda funny to see people into self-hosting still doing the kind of mistake I did almost 3 decades ago (fortunatelly in a professional environment) of trusting a 3rd party to the point of becoming dependent on them and later getting burned when they abused that trust, and which led me to avoid such situations like the plague ever since.
Mind you, I can understand if people for whom self-hosting is not driven by a desire to reduce vulnerability to the whims of 3rd parties (which includes reducing the risk of enshittification) and is instead driven by “waste not” (for example, bringing new life to old hardware rather than throwing it out) or by it being a fun challenge, don’t really care to be as independent as possible from such 3rd parties.
Look for a processor for the same socket that supports more RAM and make sure the Motherboard can handle it - maybe you’re lucky and it’s not a limit of that architecture.
If that won’t work, breakup your self-hosting needs into multiple machines and add another second hand or cheap machine to the pile.
I’ve worked in designing computer systems to handle tons of data and requests and often the only reasonable solution is to break up the load and throw more machines at it (for example, when serving millions of requests on a website, just put a load balancer in front of it that assigns user sessions and associated requests to multiple machines, so the load balancer pretty much just routes request by user session whilst the heavy processing stuff is done by multiple machines in such a way the you can just expand the whole thing by adding more machines).
In a self-hosting scenario I suspect you’ll have a lot of margin for expansion by splitting services into multiple hosts and using stuff like network shared drives in the background for shared data, before you have to fully upgrade a host machine because you hit that architecture’s maximum memory.
Granted, if a single service whose load can’t be broken down so that you can run it as a cluster, needs more memory than you can put in any of your machines, then you’re stuck having to get a new machine, but even then by splitting services you can get a machine with a newer architecture that can handle more memory but is still cheap (such as a cheap mini-PC) and just move that memory-heavy service to it whilst leaving CPU intensive services in the old but more powerful machine.
Dust is going to be a problem (well, maybe not that much electrically, but it maks it a pita to keep clean) after some months, especially for the Raspberry Pi.
Consider getting (or, even better, 3D printing) an enclosure for it at least (maybe the HDDs will be fine as they are since the fan keeps the air moving and dust probably can’t actually settle down on it).