How does this compare with other self-hosted packages like Ollama? Does Odysseus have any unique features that others might not offer?

In my case, the server has four gigabit network ports. And I do in fact create bridges on them because nearly all of my services run on VMs (I run two identical servers for redundancy of the services, and run load balancing off the firewall). Honestly this isn’t really a matter of the exact interface names, but rather why the ability to give them the desired names was intentionally broken starting with Debian 11. And as a sidebar, also trying to understand why the newer system refuses to assign the predictable names, because if it did that then I would be free to simply rename them as ethX with the systemd link files the way udev used to. On the firewall with eight network ports, yeah having to give all new names to everything was a really big deal tied to a lot of different software packages. On these servers, though, there’s not much running except kvm and I only had to readdress the bridges to the new interface names and everything worked.

It’s just the principle of it that REALLY bugs me…

I’ve considered Devuan, but haven’t taken the plunge into it. The problem here is, at least to my understanding, that predictable naming is supposed to be a core part of the newer linux kernel, and since I can find no sign in dmesg that those names are even being attempted, it worries me that something very basic has been broken that I can’t go back to. I Just don’t know enough about it to risk taking down a live server to try rebuilding in a different distro without at least first understanding why exactly this particular thing is broken in Debian. “The devil you know” and all that…

Thanks for trying though, it could still easily be something I missed. And there’s the chance that somehow the BIOS is blocking systemd from setting the names but that really seems like an overreach of what the BIOS should be doing.

I mean it’s kind of cool how systemd multitasks starting up services, so like I can already ssh in before half the other things are running. And it’s certainly nice to have a lot more information now about running processes. But when it screws something up, it REALLY screws something up.

In this case, it’s just easier to remember eth0 for each of the servers, rather than having to look up the name each time I work on a different machine. There’s no software reason for it, it’s just for my own convenience. I can accept naming the interface to dmz0, it just really bugs me that debian no longer gives me an option to name things the way I want when I could easily make those changes on older versions.

BOOT_IMAGE=/boot/vmlinuz-5.10.0-42-amd64 root=UUID=1a99d775-a924-43e4-9ffa-d170827cc8df ro consoleblank=600 quiet

It definitely help for re-world modeling, getting things figured out in advance. In the Spring of 2024 I decided to finally build a utility trailer I’ve been thinking about for years, and I modeled the whole thing in OpenScad down to the wiring and generating material lists (and yes, we got it built and have been using it through this past Summer). The fact that you can’t incrementally update variables makes some tasks stupidly difficult requiring a lot of needless redundancy, but I do like being able to write stuff out as code.

Maybe some day I’ll try Build123 or something else, but most of my projects are things where I just need to whip out a quick model, so I don’t want to deal with learning a whole new system at the same time. Case in point – yesterday I picked up a keyboard fro the thrift store but realized it was missing one of the folding feet (this one has two different sized stacked feet that fold out from each other). Half an hour later I’m making my first print, and another 15 minutes I’m ready to test the second foot piece. And I could have done it faster if I didn’t spend so much time capturing the details of the original feet. Point is, it’s hard to change when most of your needs are quickly covered with the existing software.

I’ve been using openscad to create a dual-filament direct drive print head for my Ender 3. Got the whole thing printed but I have to run a couple more wires up the cable harness to control the new servo and read the filament runout sensor.

Honestly, are there any HDDs that can really reach 6gb/s speeds? I haven’t seen any that could reach 3gb/s. My current array is all running on SATA-2 backplanes, but with 8 drives in the raid it clocks out at 460MB/s sustained (bytes, not bits). Considering my previous NAS could barely reach 70MB/s on a 6-disk array, I was quite pleased with the new setup.

Your server needs to have a static IP address. Once you have that, and have your DNS service set up, then you should be able to update your domain(s) to point to the new DNS. This new server doesn’t have to be your primary DNS entry for the domain(s), but it should be one of the first two entries. And that’s pretty much all you need to get started.

One other consideration is setting up the master/slave status of your DNS servers so you only have to make updates in one spot (helpful to ensure everything stays in sync). This isn’t a requirement, it just makes your life easier.

Nope, programming is like a black hole – once you’ve entered the event horizon it’s too late and you’ll never escape.

What do you mean by “traditional cad designers” and why is that important? I’m not really sure why that is something that should matter to me?

Aww man, I knew it was only a matter of time before someone was going to “force” me to learn python. Thanks for the link, I’ll take a look at this after dinner.

Another consideration… If you are a programmer type then OpenSCAD is a language-based program. I’ve been using it heavily for the past week designing a dual-filament extruder for my Ender 3, and last year I designed and built a utility trailer. As with anything it has its quirks, but I’m much more comfortable writing code and I always found the other GUI-oriented programs to be unintuitive.

I think my suggestion would be to use the PC as a dedicated firewall, but you will need at least two ethernet ports for that (one to connect to the ISP router, and the second for internal network). This lets you learn network security and control the traffic that can actually get to your other computers. You could also set up KVM to start running virtual machines here. The idea with a VM is to keep services separate and isolated, so like one VM to manage security cameras, another VM to host a game server, another one to host sonarr/radarr/jellyfin… etc.

When you are able to expand, your second physical machine should be your NAS. Get your storage space started, share it over NFS or samba, and move your backup/security camera/sonarr VMs over to this machine for direct access to the larger storage space. Pay attention to system usage and move your VMs to balance the cpu/memory resources. Eventually you may want to get a third machine dedicated more for hosting the game servers, maybe a web server to view the security camera feeds, or whatever.

You mentioned backing up Wikipedia… Have a look at the Kiwix project, you can already get access to regular backups for a lot of information sites.

Why does everyone think the term came from Star Wars? I know it was used in steampunk before then, and google suggests it goes back to a 1958 article about robots. Sorry, not trying to be pedantic, just feels like a lot of people give Star Wars unjust credit for things they didn’t actually create.

Don’t forget that managers think the same thing – if it’s free then it is somehow an inferior product but if you pay for something then that automatically makes it better. This applies forward as well… the more they pay for something, the “better” it must be.

For example… Cybertruck.

From my perspective, open-source products are greatly superior because you have the entire community of users and engineers working on a known issue, rather than a few paid engineers who may not even use the product. Even more importantly, the community will solve problems that a corporation has decided aren’t worth the effort or are “obsolete”.

Yeah I think there’s some good ideas that have been built up with the ABL devices over the years, and that mesh leveling seems great. However I also feel like more people should take the time to get their beds physically as flat as possible first, because otherwise your printer spends a lot of time micro-stepping the Z, and I’ve seen some mesh maps that looked absolutely horrible.

Playing around with arduino devices, I have some laser rangefinders that seem like they could work well for an ABL. That might also be a good way to map out the idea I mentioned about creating a thin full-bed shim to get a truly flat bed, and after that perhaps you would only need to do a fast four-corner calibration with the ABL to make sure nothing has changed.

I’ve never had auto-leveling on my own printer, but the one at work has it and I wouldn’t want it. It was great for the first couple years, then the sensors started suffering from corrosion and now it just doesn’t work for crap.

So on my own printer (fully manual) I have always treated the paper test as nothing more than an initial step to make sure the nozzle doesn’t plow your bed during the leveling process. It will get you close, but not close enough. To really dial in your bed, you need at least a 5-point 1st-layer print test so you can really get that initial layer adhering nicely at all points, and this of course performs the calibration with the bed at full heat so it matches conditions during actual prints.

The biggest problem I had was my aluminum bed was badly warped when I got it, but fortunately just bowl-shaped. I cut CD-sized discs of aluminum foil to build up the center – 13 layers in total, then put a glass or G10 bed over that to really get a flat surface. I finished mine with a PEI sticker on top of the G10.

Another problem is that all beds will be completely unique from each other. I have one of the original Ender 3 Pro printers, and purchased a Creality glass bed with it. There was nothing that wouldn’t stick to that bed, and I used a 10mm calicat (which has 4mm feet pads) as my tests for all new filament. Clean the surface occasionally with 91% ISO, and life was good… but after a few years the surface wore out and I was forced to get a new bed. Ordered directly from Creality again, and man, nothing at all would stick to that piece of glass, even for large prints. I tried everything down to brake cleaner to solve the problem and finally gave up on it. That’s when I started working with the G10. The point of this is that you can tell someone a particular bed will solve all their problems, and you will be wrong. The reality is that sellers use different manufacturers to make their product and you never know what you’ll get.

Your observations about the bed warping from the tightness of the screws is interesting, I never thought about that part before. I have heavy springs under my bed, and I have always suggested people get those springs as tight as possibly without being completely closed, then adjust the Z-switch to that point before you start your leveling process. Tight springs means the screws should never wander. It’s been around 3 years since the last time I even touched my leveling screws, and I just fired up the printer and ran some new pieces last week without any issues. A good tool is one that you can ignore for a year, then use it without having to recalibrate. Anything else is just frustrating!

There are a lot of suggestions we can make to help newcomers get their bed leveling correct, and there are a lot of variables that we simply can’t account for including manufacturer defects. One idea I had years ago but never got around to trying is to print a thin sheet of filament the size of the entire bed, ironed to create a top surface perfectly flat to the nozzle but taking up any imperfections in the aluminum plate. Then put a thin bed over the top of that, and you should have a perfect surface that lasts nearly the life of the printer. Seems like a good idea, but how do you figure out where to fill in those first layers until you have a final layer that covers the full bed?