NaibofTabr

• a few git repos (pushed and backup in the important stuff) with all docker compose, keys and such (the 5%)

Um, maybe I’m misunderstanding, but you’re storing keys in git repositories which are where…?

And remember, if you haven’t tested your backups then you don’t have backups!

Yeah but the current build of libvegs has some conflicts with libfruit, so if you need to use both you have to build libvegs in a different directory and then simlink it in /lib.

however, i know nothing about self hosting. My knowledge is absolutely zero […] I dont understand nothing

This is going to be a problem, unfortunately. You’ll need to define your use case first:

• How much music do you want to have access to? Hundreds, thousands, millions of files? How large is your collection?
• Do you have downloaded copies of all the music you want to listen to? Are they all in one place, well organized and tagged? If you just have downloads in the Spotify app, you won’t be able to use those files, you don’t actually own that music. You’ll need DRM-free audio files.
• Where and how do you want to be able to access them? Just from one device like your phone? Many devices? Is having access at home good enough, or do you want to be able to access your collection while you’re away from home?
• Will you be the only user?
• What kind of budget do you have to work with?

An old PC might be enough to act as a server, but there’s more involved and the answer to what you need depends on what exactly you want to do. You will not be able to build a personal version of Spotify with just an old PC, for instance.

Never use a spreadsheet to do a database job.

Maybe it’s time to start funding software infrastructure the way we fund physical infrastructure - with tax dollars.

“Here, maintain this for us, we don’t want to pay anyone to do it”

This assumes that

1. There’s a process to read.

2. The steps in the process are complete and thorough.

Those are bad assumptions.

Metronet will be supplying an Optical Network Terminal, probably like this one:

This is basically the equivalent of a modem for cable networks. It does not provide routing functions. You’re probably stuck with the ONT they supply, but it shouldn’t matter much, definitely not for anything internal.

It looks like Metronet normally supplies Eero WiFi mesh devices for home networking - are the ones you currently have supplied by Metronet? They might just replace the modem with the ONT and leave the existing Eero gear, or they might upgrade the Eero gear to support the higher speed available on the fiber network.

In any case, if you are using ISP-supplied network devices then you don’t control the router, which means you can’t set up things like port forwarding to access your home network from outside, or configure VLANs to segregate devices on your network, or control things like DHCP.

Technically there’s no reason you have to use the Eero devices from Metronet, you should be able to plug any router into the ONT WAN port and have internet service. If you don’t want to get too deep into network config, then any modern consumer WiFi router will work (but not a modem/router AIO device). If you want to have a bit more control, look for one that supports OpenWRT.

You’re welcome. I tried to do basically the same thing a few years back (run a WordPress site from a RPi on my home network - and also a Minecraft server) and so I tried to write up the problems I ran into - I probably forgot some, but those are the major hurdles. I learned a lot along the way.

I should also point out that if you rent hosting space (from linode or inmotion or digitalocean or many other options) then problems 1 & 4 become much easier, and 2 & 3 go away entirely (most people don’t host public websites on their home networks because of these and other issues).

Happy to help. If you’re interested in learning networking more thoroughly, I want to point you to Professor Messer especially the Network+ content. He has far more complete explanations than I could write (and it’s free!). Even if you’re not interested in getting any certs, the explanations will be helpful.

So your goal is to host a publicly accessible static website from a computer in your home. There are a few problems you need to overcome before even worrying about configuring any software. You need some more basic networking knowledge first.

0. Basic Networking Theory - you should read a brief explanation of the OSI 7-layer network model. You don’t have to try to memorize this and you won’t really understand it until you start actually doing stuff, but you should read it for some basic terminology and to understand that there are distinct steps through which communication between computers happens.

   When you start running into problems (“why can’t I access the server? I did all the tutorial steps”), figuring out which layer the problem is in will help guide you to the solution:

   • is there a bad cable? -> 1. Physical
   • do I have the right IP address? -> 3. Network
   • is the firewall port closed? -> 4. Transport

1. Privacy/Security/Safety - don’t host a publicly accessible website from your personal computer. Just don’t. To make this happen you will have to open a hole in your network security that makes your computer accessible from the public internet. Don’t do this on your daily driver computer. Don’t do this with any device that has any files on it that you care about or any access to any personal information. Don’t.

   Set up your web server/learning environment on a clean, dedicated system. This could be an old laptop or a Raspberry Pi (an older 3B model will work just fine for this) or whatever cheap computer hardware you have, as long as it can run Linux and has a physical network port (using WiFi will give you extra headaches for getting this working). If you think you might want to expand your projects in the future, you can get a used Dell server for very little money, and add more hard drives as needed. Wipe the hard drive and install Debian or Ubuntu server as a base, there’s lots of resources out there for setting up web services on either.

2. Restricted Ports - you are most likely on a residential internet connection. Most residential ISPs close ports for security reasons, especially 80. For example, here is Cox’s list of restricted ports. You will need to find your ISP’s equivalent list and understand what you can and can’t do with your connection.

   There are workarounds, primarily through port forwarding. You will need admin access to your router to set this up. I recommend that you read that entire article because it probably applies directly to your situation.

3. Dynamic IP Address - most people still find it easier to work with IPv4 addresses - I won’t go into IPv6 right now, but you should read a little about it just for awareness. Your residential internet most likely has a dynamic IPv4 address, which means you can’t rely on that address staying the same forever (or even until next week), which means that you can’t configure your Cloudflare domain name to point to a single IPv4 address.

   Dynamic DNS is the solution for this, and again you’ll need admin access to your router to set it up.

4. HTTPS/TLS/SSL - if you get through all those issues then you probably have a working website, but now you’re seeing something like this when you try to view it in your browser:

   This doesn’t mean that you can’t get to your website - it just means that you can only do it via HTTP and not HTTPS, which the browser is warning you (and anyone else trying to view your website) is not secure. You can either just accept that this alert will always come up, and that you have to click through it, or you can learn about TLS and getting an SSL/TLS certificate. This is a later topic - it doesn’t matter and probably won’t make sense until after you’ve got your web server online.

I plan to RAID1 them and use them as boot drives

This will not work unless the mainboard is handling the RAID control in firmware. If you are doing software RAID then the OS must boot before the array can be accessed.

If you just want to set up a NAS, you can get a used PowerEdge tower for very little money, and it will work a lot better than what you have planned.

Like it or not, commercial computing is primarily Microsoft environments. Businesses are moving to Azure/O365, but there’s still a lot of on-prem AD out there, and a lot of businesses that are stuck in between with some form of hybrid hodge-podge. It’s definitely more difficult to do admin tasks for individual Windows endpoints vs. Linux, but on the other hand there is no FOSS equivalent for AD forest management. In a corporate environment, the ability to manage large numbers of endpoints at scale is more important.

You probably shouldn’t be using iPerf3 on Windows, but instead use the native nttcp.

tracert is included with Windows by default, no need to install a separate utility. robocopy is also included with Windows and can be used to do incremental backups if that’s your use case.

If you have to manage Windows systems you should learn about Windows-native tools, rather than trying to drag the Linux-native tools you’re used to onto Windows just for the sake of familiarity.

That said, installing (and updating) software on Windows is absolutely a pain compared to the relative simplicity of a Linux package manager and I’m 100% with you on that. I highly recommend chocolatey, which attempts to work as a package manager for Windows. All of the software that you install with chocolatey can be updated with a single command, similar to running updates in a package manager on Linux. If you can implement this on the Windows systems that you have to manage, it will make things easier.

You will own nothing and like it have no recourse.

No secrets. Keep it simple, keep it safe.

• red means stop (“I need a break”)
• yellow means slow down (“Let’s try something else?”)
• green means all clear (“Keep going!”)

Qubes - an OS that compartmentalizes system functions (including userspace) into separate VMs, with the intent of keeping them secure from each other. Kind of an internal zero-trust approach. Complicated to use.

Alpine Linux - stripped down to create a reduced attack surface, with the intent to provide only packages which have been vetted for security. Fairly straightforward.

Redox OS - a Unix-like OS written in Rust (not actually Linux). Limited, still kind of a prototype.

Damn Small Linux has been revived with a new version recently, which is nice to see.

HoloISO - a community built reimplementation of the Steam Deck OS.

I think it’s worse when they say they found a solution and include a link which is now dead.

With Google dropping its archive I feel like dead links are going to be more and more common.

It’s really great to see a continuation of Openboard.

Technically the US measurement system is metric since the Mendenhall Order of 1893 reestablished all customary units as conversion factors of metric units. In 1933 the ASA redefined the inch to be exactly 25.4mm, following the lead of the British Standards Institution in 1930 (precision was increasingly important for manufacturing, and the previous value of 25.40005mm had become impractical). The international yard and pound were officially adopted by the US National Bereau of Standards (now NIST) in 1959, the Metric Conversion Act was passed in 1975, and finally EO 12770 (1991) required all agencies of the executive branch to transition to metric units.

So, from one point of view we’ve been transitioning to metric since 1893 and it’s still not done. From another, the inch is just a metric unit as its length is officially defined in millimeters (all customary units are now based on SI units), therefore the conversion is complete.

Oh yeah, I find that it’s easier to get fine control of the outcome in SuperSlicer because it’s less refined. User-friendly features are nice when you’re getting started but a hindrance when you have more experience. I tried to use Cura awhile back and it felt like the Fisher-Price version of a slicer. SuperSlicer is probably less accessible overall, but it doesn’t hide controls from me.