Keep it in his pants?

Interesting article, but man, the random capitalization thing going on there is distracting as hell.

Well, our discussion didn’t go too far off the rails, but it sure escalated with the other commenter. I think this guy just likes to argue.

Years ago I had a registrar go tits up without warning, taking about 70-80 active domains for an MSP’s customers with it. I managed their email servers and DNS, which was with the registar, of course. It was a bloody nightmare to recover that situation. Because we couldn’t supply them a DNS change to prove our control of the DNS, hence ownership of the domain, we had to individually affadavit each domain. Took weeks.

I get you don’t think it’s important, but there’s plenty of sysadmins that do, with experience backing that up.

Moved to Porkbun, saving about the same. The price creep pisses me off.

I don’t agree and it’s no extra work to do it the other way. And when one or the other goes fucky, you can recover immediately.

This is the way. Never use the NS of your registrar.

If you click the HA header, there’s 5 different scripts there, each install HAOS in a different way. As a container in an LXC, as a “baremetal” LXC, as a VM, as a Pimox VM, or as a Podman container.

Blue Iris and any ONVIF compatible camera. BI has a phone app for Android and iphone that you can integrate via ngrok if you have CGNAT internet and can’t open a port directly.

It’s a little bit of money (license is cheaper via Amcrest website) and you have to run it on Windows. I use Dockurr/Windows and run it in a KVM-based docker container on a couple sites.

Very easy to set up, has motion/shape detection built in, bulletproof, features until hell won’t have it, and works with a plethora of cameras.

So there’s 5 methods in there of installing HA, not to mention having to take apart the scripts to figure out what they do in order to help you troubleshoot this.

Not sure what you expect of asking a question like this with no useful info about what you actually have in front of you.

Try the OCCWeb app in nextcloud apps.

I wouldn’t be surprised if they were spit out of Claude.

I’ve set up OpenWebUI with the docker containers, which includes Ollama in API mode, and optionally Playwright if you want to add webscraping to your RAG queries. This gives you a ChatJippity format webpage that you can manage your models for Ollama, and add OpenAI usage as well if you want. You can manage all the users as well.

On top, then you have API support to your own Ollama instance, and you can also configure GPU usage for your local AI if available.

Honestly, it’s the easiest way to get local AI.

it’s come a long way in the last few majors. I’ve been using it since v7 and it’s been a struggle at times with users complaining about how glacial it could get. These days it’s pretty blazingly fast comparatively, I get no complaints.

I’ve quite come to look forward to this newsletter.

Yah, I don’t think a Pi3 is the place to make many determinations on the efficacy of VMs vs bare metal.

What’s the drive config? If it’s a SAS raid, you’ll probably want to look up how to flash it to IT mode so it’s not doing hardware RAID. I’d install Proxmox in a RAID 10 ZFS configuration and go to town.

I do this but in a docker VM. Then I can snapshot and back it up. I haven’t noticed any performance disadvantage since it’s running as a KVM guest, so it’s pretty much the same are running on bare metal.

https://doc.dovecot.org/main/core/man/doveadm-sync.1.html

most backup providers encrypt data on their servers so you dont have to manage that I dont think.

That’s something you should manage yourself, so the provider isn’t the one with the keys, by encrypting the backup locally before sending it. Most solutions you mention let you do that.