

It is kind of a hassle to get all your DNS set up so your mail doesn’t go to spam. If you don’t already know what you’re doing it will probably take over an hour.


It is kind of a hassle to get all your DNS set up so your mail doesn’t go to spam. If you don’t already know what you’re doing it will probably take over an hour.


I verified my domain with Proton and I can send mails from my domain through their servers and pass all sender verification and reputation tests. This works without needing to control a public IP in a range that is allowed to send mail to the common mail services.


What do you do if somebody adds your IP to the crowdsourced blocklist used by your server?


This kind of thing has been going on since long before “AI.” Expect anything connected to the internet will have failed login requests. That’s why there is a login system.


Can you just connect a regular drive using a USB SATA adapter or something?


There are much better resources about these things. The YouTube comments are talking about very basic mistakes like not knowing what a shared library is. This is a similar device: https://www.youtube.com/watch?v=lrV0pbx7U7E


Orbital data center plan:
Step 1: Establish a slave labor force and advanced chip fabrication facility on the moon.


A bit flip might change a 0 to a 1 or a 1 to an infinity. Even if you could just do everything three times, that triples the hardware and energy costs compared to terrestrial computing.


Porkbun is a registrar that also has DNS hosting, and ACME DNS challenges just need DNS hosting. I also use Porkbun and I am happy with them as a registrar, but their DNS service wasn’t working out for me and I host my DNS records on Desec. It’s working well for me after I increased some timeouts in my ACME client configuration but I can’t recommend it if you’re having problems with Porkbun because the Desec rate limits are strict and their TTLs are high and if you don’t already know what you’re doing you’re likely to spend a lot of time waiting for propagation or rate limits.


If they cut the validity time for certificates, I’d expect them to also increase the rate limits by a corresponding amount. It’s not like they have anything to gain by making it so regular users can’t use the service anymore. They can’t upsell you to Lets Encrypt Premium with a higher rate limit.


You can use exit nodes, but at least in my case that would be really stupid of me to do. If you have a router connected to a VPN and you use that router as an exit node, all of your traffic is going from your mobile device to the VPN server to your router back to the VPN server to the destination host and any return traffic takes the same route in reverse, adding additional latency and limiting your bandwidth to the minimum of any link in either direction along the entire chain. You can potentially exclude the mobile to router Tailscale traffic from the VPN tunnel to skip a little bit of latency, but it probably doesn’t help much unless your VPN server is in a third location that is not along the path between your current location and your router. My slow upload speed would become a slow download speed, and when I travel long distances my latency to services at my destination would become half a second.
What I do is I have a travel router that I deploy where I’m staying, and that router has a site-to-site VPN with my home network. That way traffic doesn’t need to travel across continents to reach a server only 20ms away.
I also have a set of services that are exposed directly to the internet and I can reach those servers without Tailscale. I can live with being connected to a different VPN and not having the Tailscale-only services.
It may be possible to just use Wireguard. The main benefit of Tailscale instead of Wireguard is that two Tailscale nodes that are next to each other can connect directly without going through another server, and this is accomplished by continuously reconfiguring Wireguard. If you just want a private network VPN where you have a fixed route to your private network and a fixed route to a public internet VPN, you can do that without Tailscale. If you are traveling with a phone and a laptop, connections from the phone’s VPN IP to the laptop’s VPN IP will be slower as they route through your VPN server, but they will work.


This is sadly the only way to do it on Android. It’s not supported by Headscale and it only supports Mullvad.


You can load models for Frigate yourself, and the documentation tells you how to do it, but the recommended Frigate+ models are easier to use. For example, downloading and configuring YOLO-NAS becomes just copying and pasting a plus:// URL when you’re signed in to Frigate+.
As another example, I would consider GitLab not to be free because GitLab is a for-profit company, the open source version of GitLab intentionally lacks features that would be particularly useful to business users, and you can pay GitLab to get those features in a special GitLab distribution distributed under difference licensing terms. If GitLab had a plugin model, and unaffiliated developers created paid plugins for those features, then I think GitLab itself could be considered free. But if paid plugins were developed by the same developers, would that make it not free again?
More strange examples:
I guess to me it seems like there’s this gray area where you start having to think about intention and whether the software is really intended to be usable for the purposes that people in this community will want to use it for without having to pay the person doing the promoting.


Are Home Assistant and Frigate exempted? Home Assistant is free and open source and you can self host it, but there is a built-in feature where you can pay a subscription to use Nabu Casa’s ingress server and cloud GPUs, and many of the integrations are only useful if you have paid money for some piece of hardware or have a subscription to a cloud service. Frigate is free and open source, but it has built-in support for specially packaged computer vision models that are offered for a fee that supports the project. I wouldn’t consider either application crippleware, but you can pay money to people who are affiliated with the project for a direct benefit that is related to the software.


You’re already doing more than companies that charge people to host their photos.


KDE Connect can do all three of these.


A 172.16.0.0/12 address would be a very unusual fallback behavior. Normally you’ll have only a 169.254.0.0/16.


lv* commands are for LVM logical volumes. They’re not related to LUKS. Unless your LVM “physical” volume is on a LUKS encrypted device, you aren’t using LUKS. You’ll need to make the LUKS encrypted PV larger before you can grow your LV.
Unless your PV is already taking up the whole disk, in which case whether it is LUKS encrypted or not is irrelevant and the question you want to be asking is “how do I resize my LVM LV to fill the rest of my LVM PV?”, which is probably what the commands you have do.


You can use both at the same time and it is useful to have ULA if your ISP changes your assigned prefix.
If you have local SMTP and IMAP you can send mail on your own systems, or if you’re not using containers and you run everything on the same system you may even be able to skip SMTP and just deliver directly. But if you want to receive mail from outside your LAN you’ll need a way to access your IMAP server.