One thing I ran into, though it was a while ago, was that disk caching being on would trash performance for writes on removable media for me.

The issue ended up being that the kernel would keep flushing the cache to disk, and while it was doing that none of your transfers are happening. So, it’d end up doubling or more the copy time because the write cache wasn’t actually helping removable drives.

It might be worth remounting without any caching, if it’s on, and seeing if that fixes the mess.

But, as I said, this has been a few years, so that may no longer be actively the case.

I’m a fan of the Bambu printers because they just simply work.

You want to print something, they print something, done.

If you want to fiddle, then they’re the wrong printers, but if you want to model shit and make things then they’re really hard to beat right now.

And, yes, I have reservations about the closed sourced nature, but honestly ask yourself: are you going to contribute to the code? Are you going to build your own firmware to run on your printer? If the answer is no, then that’s probably not really a concern that should be driving your decisions.

Sounds good to me?

Straight up piracy at this point.

I have vanilla-ass white boy musical tastes, so I’ve had little issue finding what I want on Soulseek.

That said, there is one thing about Soulseek that’s not advertised: there’s a freaking enormous list of “blacklisted” terms that won’t return search results even if the data is there.

Lots of banned artist and album names that will return zero results, unless you do something like search for a song or two that’s on the album you want and finding the data that way.

Might be worth seeing if changing what you’re specifically searching for improves your results, since I was dealing with like 70% completion until someone told me about that ah, feature.

Edit: and you can have my iPod from my cold dead hands.

Comedy NNTP option here.

It’s an established, stable, understood and very very thoroughly debugged and tested protocol/server solution that’ll run on a potato and has clients for every OS you’ve ever heard of, and a bunch you haven’t.

Setting up your own little mini-network and sharing groups is fairly trivial and it’ll happily shove copies of everyone’s data to every server that’s on the feed.

Just encrypt your shit, post it, and let the software do the rest.

(I mean, if it’s good enough to move 200TB of perfectly legitimate Linux ISOs a day, it’ll handle however much data you could possibly be backing up.)

Disclaimer: it’s not quite that simple, but I mean, it’s pretty close to. Also I’m very much a UNIX boomer and am a big fan of the simplest solution that’s got the longest tested history over shiny new shit, so just making that bias clear.

A big point of confusion that keeps happening in relation to OCI is that there’s actually two “tiers” of free, and one of the two is subject to resources vanishing.

If you convert to a pay-as-you-go account, all that shit stops, and you’re treated as an actual customer while keeping all the free tier stuff.

I suppose you could get hit with a surprise bill if you’re not careful and use things that have a free tier and then convert to billing (example: you exceed your object storage free amount), but if you don’t use anything outside of the compute resources, it’s just as good without the resource reclamation stuff.