UnpledgedCatnapTipper

My Bolt EV weighs like 3600 pounds, about 400 pounds more than my old ICE Honda Accord. Maybe we should tax vehicles based on manufacturer spec curb weight. That might push demand for smaller, lighter, more efficient cars.

It would not charge at the full speed the whole time, but the specifics depend on the individual model of car.

So, my car is a Chevy Bolt EV, and it has a 60kWh battery. If it were capable of charging a 1MW, it would take 3.6 minutes to charge from 0 to 100. My car can unfortunately only charge at a max of about 50kW, but I’ve owned it for 9 months and have only charged on my standard 120v charger (~1.4 kW) at home.

“Optimal” in what way? On an average day, I use ~10-15% of my max charge, with a max range of ~200 miles, +/- a bit depending on outside temperatures. I actually have the “hilltop reserve” feature turned on so I have regen brakes available always, which stops charging at 87%.

Most people do not need their full charge capacity, it’s just nice to have for the occasional road trip. With improved charging networks, ~100 mile range would he plenty for the vast majority of people for the vast majority of days.

I’m using Zoho to host my emails for my custom domain (I pay like $15 per year or something). I do have DKIM, SPF, and DMARC configured, as they’re not hard to configure, as they’re literally just a couple of DNS entries.

Moving to another provider would just be a matter of changing these DNS entries along with your MX record to point at your new provider.

I run a variety of self hosted things via my domain on a dynamic IP. I just have dynamic dns set up to check my current public IP periodically, and update the dns entry if it changes.

You’re welcome!

I use https://nginxproxymanager.com/ for my reverse proxy. It doesn’t exactly integrate with ESS, but it’s not super challenging to set up proxy hosts for the domains (you can specify multiple domain names in one Proxy Host if they all point to the same host/port).

I find it nice because I can manage it via a web portal, but it’s on my home LAN. Depending on how you access your VPS this might be less convenient than using one of the other options, but I don’t have much experience with the others.

I set up Let’s Encrypt certs to automatically renew for the Matrix domains, pointed them all at the ESS host server and port, and then enabled Websockets Support. That last bit is critical, or things will simply not work correctly, especially calls.

Alright that gives me a good idea what your working with.

I am running it along side some other projects, so I already had a reverse proxy set up, so I didn’t look closely at the other parts of the guide in that section.

If you want to be able to use this server for other hosting in the future, you may want to set up a reverse proxy. I can give some advice about that if you’re interested!

In your case, if the only thing you ever want to host on this box is Matrix, you don’t need a reverse proxy. You should be able to do the steps here and it should result in a working deployment.

It looks like that should give you a tls.yaml file, which you need to include when issuing the helm command to deploy everything. This one:

helm upgrade --install --namespace "ess" ess oci://ghcr.io/element-hq/ess-helm/matrix-stack -f ~/ess-config-values/hostnames.yaml <optional additional values files to pass> --wait

You need to make sure that in the <optional> section, you include:

-f ~/ess-config-values/tls.yaml

So your command would be this, if you have no additional yaml files, and if you do, simply put “-f path/to/file” at the end, right before the “–wait”:

    helm upgrade --install --namespace "ess" ess oci://ghcr.io/element-hq/ess-helm/matrix-stack -f ~/ess-config-values/hostnames.yaml -f ~/ess-config-values/tls.yaml --wait

Ok, so that sounds like either a DNS issue or a reverse proxy issue. Did you configure your domain/subdomains to point to the public IP address of where you’re setting things up? Are you using the reverse proxy in the guide or do you already have a reverse proxy and you’re adding ESS domains to it? Did you configure port forwarding on your router?

I have had issues with accessing my locally hosted services via domain name while on the same network. My router doesn’t like to route internal traffic back to its own WAN port. Can you access it from something on a different network (cellular data)?

If you want to be able to manage it via a web page instead of config files, nginx proxy manager works well, and can be run via docker/docker compose.

https://nginxproxymanager.com/guide/

It’s not quite as simple as a single docker compose, but the Element Server Suite for hosting a matrix home server (synapse) was fairly simple to get working.

There’s an additional component, and how you deploy it depends on the method you’re deploying things. If you do the Element Server Suite (https://github.com/element-hq/ess-helm) via their instructions, it includes the voice/video component and web client. If you deploy your homeserver in a different way, you should be able to find instructions for Element Call.

Synapse is the Matrix homeserver, it handles everything except for the voice/video call features.

Edit: it doesn’t handle the web client either, but technically you don’t need that for a working setup.

You will need a domain, but they can be had for a couple $ per year.

As for hosting on synology, I don’t know how difficult it is. I used the Element Server Suite on a Debian VM, and aside from some minor issues getting email and automatic room joining working it was pretty simple following their guides.

Does accessing your router page via caddy work when you’re actually on your home network and not accessing it via wireguard? Have you tried a different web browser to rule out your LLM suggested cookie issues entirely?

I got 3 Seagate Exos X16 14 TB drives for only $140 each (refurbished) at the end of 2022. I’ve got them in TrueNAS as a zfs array and they work great.

Mine were the SATA version, which isn’t currently in stock. The SAS version of my drives go for $299 now. The SATA X14 version is $350.

So prices for the same refurbished drives are more than double what they were like 3.5 years ago, so they really are expensive! I paid like $10 per TB, but they’re all $15-25 per TB now! I was looking for drives for a friend who wants to get started self hosting, and I was shocked by how much refurb drives had gone up.

This is all from https://serverpartdeals.com/ by the way, I’m assuming that’s the site you mean too.

I’m struggling a bit. I got the server up and text chat is working great, but the documentation for getting voice calls working is pretty hard to follow.

My searching around has failed to find a more step by step guide for modifying the gigantic sample continuwuity.toml file. It’s so unwieldy, and it feels impossible to know if there are some additional settings that need to be configured that I’m simply missing due to the length of the file.

Any tips, tricks, or guides you’re willing would be appreciated!

Well nVidia just sells the hardware to the AI companies, so even if the bubble pops, they won’t go bankrupt. They will stop making such obscene amounts of money, but they’re one (also the largest) of the 3 major GPU vendors. Personal computing still would buy from them, as would non-AI datacenters. He wants to keep the bubble going for as long as possible to boost their profits for as long as he can, but as long as people need graphical rendering and parallel compute power, I don’t think nVidia is going anywhere.

Think of them as the guy selling prospectors their tools. They hype everything up and jack up their prices for picks and shovels. When the prospectors don’t find any gold to make their investment back, the shovel guy just goes back to selling shovels at normal rates and prices. Sure, he’s not making as much profit, but he’s still solidly in business.

If you do have automation, that’s another thing you have to set up and manage.

Hosting a CA is a whole additional service to set up, as is enabling trust for said CA on every server you’re running.

Bare metal is commonly used to describe server OS installs that are not virtualized and are installed directly to the hardware. People are most likely conflating containerized and virtualized.