Forward proxy. Reverse proxy is meant for inbound connections, not outbound.

You’re probably overthinking this though. Unless you absolutely need a SOCKS proxy, just setup a wireguard tunnel. It’ll be faster since it’s in-kernel and uses UDP.

Here’s an easy script that’ll get you setup fast:

https://github.com/angristan/wireguard-install

https://github.com/rapiz1/rathole if you’re wanting to forward raw ports

Wireguard if you’re just wanting to run a boring reverse proxy. I use this to get me up and running fast on Wireguard:

https://github.com/angristan/wireguard-install

This guide should help you be able to visualize what you’re trying to do better: https://blog.fuzzymistborn.com/vps-reverse-proxy-tunnel/

Good luck!

You appear to be already using Cloudflare so it might be doing a “good” enough job to combat spambots and other adversaries.

Since I don’t much like cloudflare, my instance uses Crowdsec which uses a crowd sourced block list aggregated from other servers to block out nasty IPs that have a bad reputation for attempting to run automated exploits, spam, overzealous crawlers, and other assorted nasty stuff on said servers that the list is crowdsourced from. A lot of Tor exit nodes, to no surprise, are on that list too.

Idk this is something that might pique your interest if you want to further harden your setup.

Using a third-party app or mobile web browser?

Really? I could’ve swore I saw the likes of beehaw and other “safe” instances interact with the vanilla porn instances.

Indeed.

Done.

Sure thing. Gimme a sec.

Hopefully I fixed the logic once and for all. The second part, you’d have to take that up with Disa, he’s the one who wrote the UI portion, assuming he’s interested.

Dunno what happened but it works now. I probably mistakenly ran it on my laptop’s dev environment instead of the server.

Looking into it… I’m a bad coder lol.

I ended up adding some improvements to this current script. Dead/nonfunctional instances are excluded as well as instances with < 50 users. (Sorry single instance users who comment here, please don’t take this personally.)

To add, I improved upon it and have it automatically remove the nonworking sites from the list. Since there are a lot of dead/nonfunctional instances we’re still linked to.

Should be back up now.

Debian works, I just wanted something with newer packages. Using kubuntu 23.04. Hell, OpenSUSE tumbleweed also works, but Plasma crashes on me randomly, murdering my session.

I’ve concluded that Arch has to be toggling something in their kernel that my hardware doesn’t agree with at build time, because it’s the only distro that kernel panics happen on.

It is mlmym, but it’s set to be plugged into our instance.

It’s not just you. I need to work on it again since I forgot to copy it over when we moved servers and lost all of my work for making the logic work since my /home partition got reformatted.

Yeah, if Arch wouldn’t fucking crash my laptop, I’d be using it. But I have to settle with Ubungo.

I’m a unique case because my first foray into Linux was using the CLI via SSH on a server, and I thought the CLI was insanely cool and was immediately hooked. So I don’t really have any good advice other than to just force yourself to use it I guess?

Ironically, I have a more difficult time using desktop Linux just because it’s not a headless environment like I’m used to. I still use it ofc.