There is a Powerlaw underlying global temperature, in the interglacial-times:
280-ish parts-per-million CO2 as the baseline...
The 9th-root-of-2 times that, gives you what CO2 level produces 1C global warming.
280*(9th-root-of-2)^2 gives you what ppm CO2 produces 2C global warming.
280*(9th-root-of-2)^3 gives you what ppm CO2 produces 3C global warming.
All the way up to 9C.
All the simulaitons which contradict the measured polar heating, the mega-rivers IN Greenland's ice, etc, are red herrings, and relying on them is incompetent: the evidence has already falsified their predictions.
Here is the paper:
https://www.nature.com/articles/nature19798
Notice that even not counting all the other greenhouse-gasses, like methane, sulfur-hexafuoride, etc, we already have guaranteed that the planet must stabilize at more than 5C,
.. not the 1.5C or 2C that the simulations-which-contradict-evidence predict.
This powerlaw is measured, historical fact.
Delusionally ignoring it .. isn't worthy of respect, OR able to create viability for our species.
I don't know that there will be any life left, in the tropics, within 1 century.
400+ cubic-km/year melting, now, iirc, and still accelerating, darkening the albedo of the planet...
: \
I'm an old guy, whose been in Linux, off'n'on since 1997 or 1996, and not a professional.
Keep that in mind.
I now always recommend a pair of NVMe's,
with swap on both,
with root mdadm mirrored RAID1 on both ( I've read that BTRFS "RAID1", when 1 mirror is missing on boot, simply won't permit you to boot, unless you get jiggy with the damn thing, telling it arbitrary stuff, to get it to allow that )
with /home mdadm mirrored RAID1
and use the extra space for whatever.
Use SATA for your backups.
I recommend using the fastest NVMe's you can get, but biggest is more-important.
Samsung .. what are they, EVO drives? go up to 2TB, iirc, and are reasonably cheap ( for people who can afford such things )...
This gives Linux's mdadm RAID1 speed ( it does RAID0 for reads, RAID1 for writes ), AND it gives greater reliability.
I've been stung by incorrect partition space allocation sooo many times, that now I'd stick everything on as few partitions as is sane, but as OpenBSD recommends, some filesystems on separate partitions breaks some attack-methods ( partly by breaking hardlinks ).
The difference that access-speed & bandwidth do, for your OS, and especially swap, is stunning, so if you've got the funds, consider the Samsung PRO NVMe's, instead of their EVO's, but definitely get quality & quick NVMe's, RAID1 'em up, and enjoy.
PS: I always do a prototype-install, now:
whole-device ( except swap, EFI, boot ), 1 partition, install everything I'm likely to want, of that OS, take a look at the filesystem use, for different parts of the root fs-tree, and then begin deciding what partition-sizes to be considering, using a 1.5x or 2x factor for expansion-space... ( different distros with /usr and /opt, especially ).
Then I repartition into the intended structure, & install in...
And, of course, I now expect to have to re-partition 1/2y later, as the things I've later found, & added, alter the ratios...
Obviously, if this weren't just some random guy at home, LLVM would make much more sense, because then partitions could be resized/redistributed on-the-fly.
But for now, for a machine I only-sometimes use, it's good enough.
Maybe this seems useful information?
I hope so...
( :