I’m in Germany, and it works pretty fine. They’ve got several datacenters around here, never had an issue with speed or latency.

I don’t like that they got that evil megacorp vibe, but what big Internet firm doesn’t?

Well, I need to run two separate tunnels to not run into hairpinning issue, so, some weirdness, I guess. More down to my services, though.

Interesting. As I said, I never tried yunohost. I usually work with podman, and just assign local ports to pods, then route traffic to those ports internally, which seems to work fine.

Anyway, I feel like we won’t be solving OPs issue here. Still, interesting to see some of the problems people with different setups have to deal with.

Yeah, I feel like we’re missing some info here.

I have to admit that I have no experience with yuno. Always seemed interesting, but not like something that fits into my work flow.

If they’re self-hosting at home (which I’m also doing for some services), I’d presume they’re probably running their stuff on a single machine, so I’m not sure where their router would come Into it. The data the cloudflare tunnel process receives should look the same to the router no matter the port it is ultimately sent to, and when it is sent to an address internal to the machine, shouldn’t pass through the router again.

I presume they mean pointing their cloudflare tunnel to direct lemmy.example.com to http://localhost/:[port], and I don’t think there’s any special rules about that port from cloudflares site.

I use tunnels and ports in about that range for all my sites, and don’t have any problems.

You probably don’t need me to tell you, but keep good backups. Friend of mine recently had his account nuked without any reason given, and without the possibility of recourse.

Why are you using a recording app with ads/tracking when there’s free alternatives that don’t do that?

Like, if you really wanna have transcripts, Google’s recording app works even if you deny it network access. Found that out because I was too lazy to look for a proper FOSS solution.

I actually kinda did that. Sent a preconfigured thinkcentre to my mum that boots into the jellyfin media player, connects to my server via tailscale. Just had to plug it into power, lan, hdmi. Immutable, atomic system that looks for updates on boot, applies them on next reboot, and does a rollback and ping me if the update fails.

I have ssh access, and my brother lives nearby in case everything fails, that makes things easier.

I kinda get it with “soft” targets (e.g. Let’s see what we can do in a day/weekend). “Hard” targets (you gotta do x in x Minutes) pretty much guarantee I’ll get nothing good done.

Yes, I hated every coding exam I’ve been in.

I don’t think it’s necessarily worth it for anyone currently on Linux, but if they provide support and a warranty, it might be helpful for some folks who aren’t that computer savvy, but still sick of Windows.

I guess you could install cockpit (via Terminal, sorry, but it’s pretty straightforward and there are good guides). After that, you could use the cockpit web interface to deploy docker/podman containers. It’s a bit clunky sometimes, but it does the job purely in UI.

You can also manage updates, backups, etc via cockpit if you install the required modules.

As base, I’d use any stable Linux distro that’s reccomended for server use.

Edit: Comment was in wrong place, refiled as op level comment.

I use atomic distros on my server and a media centre, but don’t see any reason to do it on my main systems. Stability is fine, and atomic distros make said tinkering more difficult.

Recently needed to set up a Win11 VM. It worked after removing the network adaptor from the VM setup, and then using the bypassnro command.

Fucking Microsoft.

I don’t think the average user thinks much about the platform they’re on, and about who controls it. I think they go to wherever most of their family/friends are.

Also, those platforms are firmly in the mainstream, the alternatives aren’t really - you’d have to actively go search for them. People just aren’t likely to do that, I don’t think.

In theory, maybe. In practice, I’ve had a lot of errors in that vein that very much wouldn’t go away, and where made much harder to diagnose by their obtuseness.

Honestly, I even dislike the mindset. Just make a big header with the generic error message and a little one below that gives some details. Having users interested in how your software works is not a bad thing.

While what you said isn’t untrue, .ml does Bill itself as a general purpose instance. Also, not all the replies are from accounts on .ml.

In a more general sense, I always felt that only reading sources that alligned with one’s political alignment narrowed one’s perspective.

Zeit, Süddeutsche Zeitung, taz, dpa, Reuters, ap, BBC, Deutsche Welle, Deutschlandfunk. Some others, if they come up. Mostly via RSS.

Finally someone who actually uses a Vostro. Always found that name unreasonably funny.

I set up my current desktop while reading Gaiman’s The Sandman, so it’s called Morpheus. Because I felt I needed to keep with the theme, my laptop is hades, my phone persephone, my server apollo, my router helios, the media centre PC is orpheus, the pi that boots and updates it outside of usage hours is eurydice, and the pi that runs home assistant is zeus (because it’s responsible for light(n)ing.

Oh, and the work profile on my phone is sisyphos.

Any video surveillance not controlled by you and you only seems like a bad idea. iRobots products have had privacy problems for years, regardless of whether they’re owned by Amazon or not (which they aren’t, merger got cancelled). Alexa does cost them money, but I wouldn’t use any non-FOSS non-self-hosted smart home system for privacy and security reasons.