What’s yours look like?

This is what I’ve been trying to figure out, so I’m glad you did it. I have a Synology too and I’ve outgrown it, but building my own server scratch is daunting so I just wanted to hook a DAS up to a mini PC and call it good. But there were a lot of people online saying it was a bad idea.

Same! I was one of the early migrants just before the mass migration, having seen the writing on the wall and having already joined Mastodon. It was a great time initially, a bit of “you’re new? okay, I’ve figured a few things out, try this”. We caught Ernest by surprise because he was only making a proof of concept to offer an alternative to Lemmy. And then we showed up in droves. But he did his best, and while it was bumpy, it was still an overall great experience, much like the early days of Reddit, pre-Digg.

I still like the kbin-style instances over Lemmy, but yeah, it’s not quite the same any more. That said, still preferable to Reddit these days.

Thank you, Ernest. Hope you’re recovering wherever you are.

I did the notification thing. If the outdoor temperature is lower than my indoor temperature and above 17° and the humidity is under 60% and the windows are closed, it sends me a notification to open them.

Same if the outdoor temps get above 24° and the windows are still open, it reminds me to close them.

I love the integrations with immich and owntracks, as well as having an OIDC login. What kind of resources does this need for a server?

Thanks!

You have a CNAME record that points service.my.domain to machine-that-hosts-caddy.ts-domain.ts.net, and with tailscale enable it hits the caddy server and then reverse proxies it to the machine:port of “service”? Which may or may not be the tailscale IP address and port?

We all got to learn somewhere!

Lot of good advice here, but sometimes people forget what it’s like to be a beginner. Since you don’t know what you’re doing, I would recommend not trying to host things on your home server and access it from the outside world. That usually involves port forwarding on your router, and that comes with a lot of risks, especially if you don’t know what you’re doing. Others have mentioned it, but a better option when you’re starting off is to rent a vps and host your software there.

Squarespace might work, but my guess is it’ll be easier to transfer your domain elsewhere. You can follow guides for that online and it’s pretty straightforward.

Having a vps, a domain name, you’re most of the way there. On your vps, you’ll want to install a reverse proxy, which is what routes incoming urls to the right place (nextcloud.domain.tld goes here, calendar.domain.tld goes there).

Docker is another thing I’d recommend learning as a lot of what you’ll self host will likely be in a Docker container. I’d watch a few YouTube videos to see how it’s done. This channel has some great videos, and there are others out there.

It seems like a lot, but learn a little here and there and don’t expect to have this all working overnight. You’ll get there!

Same with the Misskey clones. There have been some very dramatic opinions on both sides.

Nextcloud got too bulky for me, and in my search I tried a number of different options before installing OpenCloud without realizing it isn’t fully finished yet. That said, it still works well enough for my use case.

Tailscale is great for not opening your ports to the internet. Having it playable on a friend’s appletv adds some extra complexity. Reverse proxy on a subdomain with something like fail2ban would work, but it does leave you more vulnerable.

I have a 220+. It works well for what it’s supposed to be. If you want a set it and forget it nas, this is a good one. However, after a year and a half, I’m ready to move on for the same reason I don’t like Apple: too walled garden. It was a great starter nas, but it’s too limiting now. But again, of you don’t want to think about it and just have it work, it’s a good choice.

The LLM for home assistant, or just HA in general doesn’t move the needle? My HA is also pretty low key, but I was considering the idea of running my own small llm to use with HA to get off of OpenAI. My current AI usage is very small, so I wouldn’t need too much on the GPU side I’d imagine, but I don’t know what’s sufficient.

I think I mentioned somewhere, but if not, over the last couple of years I learned a lot about the software side of running my homelab via synology and the vps’s, but I still know almost nothing about hardware, so this is all really useful information. Thanks!

So I want to reduce the cooling, CPU (get one with integrated graphics), and motherboard, and not necessary but look into adding ssd and more memory.

I’m okay with spending a bit up front if it’ll last a long time, but I also don’t want to buy too much and be useless.

Yeah, I posted this assuming I would get a lot of comments about it being more than I needed, and this was already me paring it down. But then I got a lot of comments saying to add more memory, reduce the cooling, and add a ssd but otherwise not much about reducing it.

Yeah, I’m not sure if I really want to deal with an llm. It would mostly be for home assistant, so nothing too crazy.

Can I make use of two existing 3.5" and 4 2.5" drives with that?

Everyone else seems to be recommending going with Intel because it’s apparently the gold standard in transcoding. But you don’t have any problems with yours and being an amd?

Interesting, why intel?

Someone else mentioned the cooler too, so that’s out for sure. To be honest, I never really thought about graphics in the traditional sense, but I do need something for at least Jellyfin transcoding. And maybe a small llm. Would it be better to get a dedicated GPU or CPU with integrated?