Part of working with tech is knowing what to search for und using the right keywords. If you could give an example what guide / information you are unable to find, someone could give you an example how to search for it. I personally know a few junior devs and junior devops which use llama2 / chatgpt since they dont know how to search for it or read the docs.

You’re kidding?

I’m an avid home-labber trying to get into software engineering and I’m finding the software dev work to be super daunting lol

I can setup a reverse proxy in my sleep - but write a JS function without double checking google? forget it!

Check YouTube. I haven’t found a lot of written out guides but someone has made a video on it, I bet. Christian Lempa, Techno Tim, Db Tech, NetworkChuck come to mind for self hosting tutorials on containers and new applications. I’m sure there’s a lot more as well.

Each one of these things is a separate topic/project. There is of course, overlap, but concentrate on just one at a time. If you try to do everything, you’ll go mad. Application Administrator, Network Engineer, Sysadmin, Network Security, are often separate full time jobs. Just concentrate on your own needs, one at a time, then focus on how simple you can pare down that one thing you need/want to self host. What is the minimum set of capabilities you must have for this one thing? Then, expect to have as much research for this one thing as taking a 1 credit college course, at least. Maybe more, maybe less, depending what it is. That’s a more realistic set of expectations.

I’ve done a lot of web design and web hosting, and messing with Linux server stacks for around 30 years, now. You absorb a LOT over time, but it is all always changing. When I set up my first server, Cloudflare didn’t exist yet. VPNs weren’t a thing yet. The only people I knew with a network in their home were Computer Science professors. Wifi was not a thing yet. It isn’t you being inadequate, it’s you trying to do multiple jobs that each require constant professional development.

You don’t have to do it, 99.9999%^([citation needed]) of the population don’t. If it doesn’t interest or excite you then it’s not worth it.

I can barely find info on HOW to set up this things. Most blogs and articles I find are about what they are which I already know. And the few that actually explain how to set it up are just throwing so many more concepts at me that I can’t keep up.

This is big part of why I, even as an IT professional, don’t do much homelabbing or selfhosting. So often I’ll be scrolling through these subs and see something someone has done and I’ll decide I want to do that too, but so often I end up abandoning whatever it is because doing x requires googling for y and z because without y and z, x doesn’t work. And after a day of dealing with issues and tinkering on work stuff, do I want to spend more time troubleshooting? Often not.

Maybe you should read the documentation for the programs you try to install

Because most people who develop these things are, frankly, terrible at good documentation, or understanding the end-user perspective.

There’s also a downward spiral effect when you start getting into these things, because lots of them require dependencies, or ask you to do things but don’t explain why, and you’re just left wondering why you added that line to a config file somewhere, but if you don’t put it there, nothing works.

A vertical slice of the amount of knowledge you need passes through so many different disciplines, operating systems, GUIs, and programming languages that it would look like a Milhojas cake.

I’ve been a technical writer in the software industry for 17 years. The number one challenge in my work is extracting all of the information I need to write good documentation from the experts elsewhere in my company.

Back in the day you had to find a book if you didn’t know people

Most blogs and articles I find are about what they are which I already know.

Exactly… I neither want “what z is” nor all the history behind z, the Z inventors life, his grand parents lives etc. I just want a solution to the problem im having…

I’d just install Docker and add a CloudFlare Tunnel to securely host your stuff. No need to open any ports, your WAN IP remains hidden and you have the benefits of DDoS protection. It’s free as well; all you need is a domain name for which you can configure the nameservers.

Do you have a purpose to host whatever you happened to name here?

There is a reason to host reverse-proxies, however, one can do without them in a self-hosted environments. First, one needs to understand the point behind reverse-proxies (this is an example btw, you might very well know the how and why behind them), and only then would the instructions to set it up start to make sense.

The things of “how to set up” are usually covered by software’s documentation, further supplemented by googling certain specific questions and going on various forums. Never just use the first page you find but crossreference information as well.

​

Having gone into self-hosting as a hobby with the intention of supplementing knowledge in my career path (currently still studying), it can become quite daunting quite fast, especially if you try looking at a big picture, rather than taking it step-by-step. The best thing to do would be to follow a tutorial within the documentation of something you’re trying to host, then expanding per step as far as dependencies goes, and then go from there. Compartmentalizing is a very effective way to not get lost.

On the one hand, I feel you.
On the other, man, I am so happy you are suffering! lol If you, a Software Engineering student are struggling, then it does not make me feel quite as clueless anymore as a normie :)

I think this is going to be a super unpopular opinion here, but you don’t necessarily have to go with what “everybody else” is using just because they’re the big shiny pennies that everybody else is using. Years ago, “everyone” was using Windows and Internet Explorer. Does that mean that nothing else was worth running?

As to VPN, Docker, Cloudflare, Traefik, Crowdsec, Authentik… this definitely isn’t a “don’t use those things” post. If you want to learn Cloudflare, and want to use Cloudflare, then by all means, do so. This isn’t an anti-Docker, anti-Cloudflare, or anti-anything post.

But that being said, I started self-hosting with a Dell laptop running XAMPP with a single, tiny, intranet web portal on port 80. It was a couple of clicks to get a functioning LAMP stack, despite me being a complete noob. It was incredibly insecure, and not the best, newest, fanciest choice to host my little website. But it allowed me to learn incrementally, and quickly, rather than feeling like I was thrown into the deep end way over my head with – as an example – Docker and virtualization.

• I got sick of XAMPP pretty fast, and learned how to host my own database and Apache instances on bare metal outside of XAMPP
• then I learned how to buy and set up an actual domain name
• then I learned how to install SSL certificates for my domain
• then I learned about hardening Apache, and things like honeypots and fail2ban
• later, I acquired several more small servers with their own services, all behind the same router, and I learned about subdomains, virtual hosts, and proxies
• even later, with all of that foundation, I learned about moving individual services into Docker. I only had to learn Docker. I already knew the rest of what I needed.
• etc.

I wanted to self-host my own intranet-only DNS server, but I found BIND to be overwhelmingly daunting at the time, with poor information for how to actually set it all up and get it working. So I found dnsmasq and installed that instead. The configuration was incredibly simple, and it worked almost out-of-the-box. Am I going to have pretentious snobs telling me how much dnsmasq sucks and how I’m stuck in the 90s if I still use it? Probably. Does it serve my own, personal, use case for the features I needed? Yes, it’s been great and simple to maintain for years.

My “dynamic DNS service” is a bash script running from a cron job, accessing a webhook that my domain registrar offers, to update my DNS records automatically. Surely, there is some great piece of software with a catchy name and a GUI dashboard that I could be using for this purpose. But I don’t feel that I have to. My laughably basic bash script has worked without touching it for about 10 years now.

For my VPN, for a long time, I had a PPTP daemon running on my server. That, too, had been very easy to set up for a young sysadmin who knew nothing about anything. It wasn’t the most secure, I used it long past when PPTP fell out of favor, but it worked at the time, and there was ample documentation that was easy to follow to set it up.

I’ve moved on from most of the things I started out with when I was learning the basics, but all of that to say: Sometimes, it’s okay for something to be “good enough”. It’s okay if your server isn’t PCI-DSS-level secure and protected from absolutely everything. It’s okay to do the best you can, get your setup working… and then don’t stop there. Continue learning how to make it better, more secure, more reliable, more scaleable. But just because the popular kids are going to insist that you should never consider anything less than Nginx in a Docker container, managed by Portainer, behind Wireguard, behind Cloudflare… that setup can be your end goal, but you don’t have to start there and learn everything all at once.