LINUX Unplugged - 457: Automated Chaos

Episode Date: May 9, 2022

Each of us brings a secret topic to the show, and we discover a common theme about using the wrong tool for the right job. Special Guest: Alex Kretzschmar.

Okay, so we have some pretty spooky audio. All right, I'm going to play this. This is a real thing observed by scientists. See if you can figure out what it is while you're listening. All right, Brent or Alex, do you have a guess as to what we're listening to? Well, I have a guess. I don't know that I know what it is, but it sounds to me like, let's say you were in Miami right now and you might be at the, you know, the racetrack where they have the boats sitting

in the false water. If you were underwater watching the race, that's what it would sound like. Right. Because not only is today mother's day but it's also formula one day so happy mother's day everybody but go ahead what do you think it is alex i think it sounds a bit like plate reverb where they have the big metal plate in like a basement of a recording studio and that's how they create reverb i like that idea i think that makes it a lot less spooky you when you find out what it is, it's extremely creepy.

This is the sonifications from a black hole that NASA has observed. Right, Wes? Yeah, that's right. The black hole at the center of the Perseus galaxy cluster. That's what it sounds like. And it sounds like, to me, thousands, maybe millions or billions of souls. This is so cool because, okay, there's a lot of different sonifications, right, where you take some data and you map it to sound,

but this is actually measuring pressure waves in hot gas in this galaxy cluster. Like, it's actually sound. It's just, you know, to your weekly Linux talk show. My name is Chris. My name is Wes. My name is Brent. And my name is Alex.

Hello, gentlemen. I love these episodes. It's one of these weeks where I can't tell you what we're about to talk about. I got these episodes. It's one of these weeks where I can't tell you what we're about to talk about. I got no idea. We're each bringing a topic to the show, and we're going to find out live with you what each of us has to say. But then, of course, we'll get into the traditional steaks and eggs and sides, I guess. Hey, I brought biscuits. Great. You know, I was hoping, but all I have over here is cider. Anyways, are they sous-vided biscuits?

They are. Can you sous-vide biscuits? Really? We'll find out. And then, of course, along with them biscuits, we'll have picks, some boasts, and a whole lot more. Well, let's just get right into it. Wes, what did you bring to class today? I thought I'd revisit a tool that I've been playing with more again, but I used to use a lot.

And that's Canonical's LexD. Oh, yeah. Obviously, we use containers for basically all the things here. But most of the time we're running something like Podman or Docker. I was thinking back on this and it was when 16.04 first came out, if you remember those heady days, and Canonical had just started shipping ZFS, just built in, no fuss, no DKMS, like you just got ZFS. I wanted to play with that because, you know, I dabble with ZFS, just built in, no fuss, no DKMS, like you just got ZFS. I wanted to play with that because, you know, I dabble with ZFS, but mostly like from Freenas on a FreeBSD system or similar. So I rented a dedicated server, you know, had a couple of TB of storage, had a couple of hard drives, nothing super fancy, but it was enough to start playing with like a, you know, physical

system I didn't have to run in my actual home lab. And I got it all set up, had ZFS, had LexD on top. It was routed up with IPv6. And it was really neat because it was almost like running my own little VPS setup because you could just spin up all these like system containers, which is one of the things that LexD does that's a little bit different than Docker. You know, it's almost like a virtual machine, but it uses all the cool container tech. And so I could spin up Bastion hosts.

I could spin up whatever OS I wanted to play with. It was a great way to just run like the latest version of Arch on a server. What do you mean when you say it's almost like a virtual machine? Like, does it have virtual devices? What do you mean by that? At least the way that you see containers run a lot, at least with the Docker, is these sort of application containers, you know, where it's focused on one particular piece, right? It's I I'm running this process, a single thing, it's all, you know, declarative, you set it out. LexD runs what they call system containers, which is pretty much just like what you might be used to when you run a full Linux operating system. So it starts up systemd, you can run multiple processes, but it isn't a virtual machine. It's, you know, it's using the container technology of cgroups

and namespaces and that sort of thing to isolate it. So you get all the nice bare metal performance. There's no overhead involved but it's designed to work as if you have a full system access. Now this isn't for anything that's not to compete necessarily with Docker. I find it has a particular sweet spot

in the home lab for times where you don't need to scale things arbitrarily horizontally. That's not your end goal with this. It's not how you're going to run the system. But you want to be able to run multiple things and have a nice REST API on top to orchestrate it. So how does the kernel model work with this, Wes? With containers, like typical Docker containers, you would have one kernel shared amongst multiple processes. With LexD. How does

that work? Yeah, same thing. So in the container model, you get the host kernel in that case. I haven't really used it for a while. And it's been in the back of my mind, but it just hasn't been the thing that I've been playing with. We covered last month LexD 5.0 in Linux Action News, and that's what kind of triggered me to think, I should play with this again, because one of the things they've added now is they also do virtual machines. So you can do system containers if you're okay with relying on the host kernel, you want that density, if that works for you. But you now have basically feature parity on the VM side, including stuff like migrations and snapshots, also clustering. And they've added this whole overlay virtual networking.

It's powered by OVN, which then sits on top of like Open vSwitch. Okay, so this is like the networking layer? Mm-hmm. So you've got this overlay network on top, but it can do some pretty wild stuff, including if you've got like a nice NIC, like a Mellanox NIC,

can do offload, hardware offload, so it gets processed there. Cool. It can also talk BGP. So you can spin up new containers or virtual machines, and it will do all the BGP, peer with your routers, and announce them. So you don't have to worry about dealing with any of those routes. Man, this seems really nice if you're an organization using this

and you just want to spin up some infrastructure real quick. That's great. I wonder how deep we could take that rabbit hole. Could we put Docker inside an LXD and then run LibVert inside the Docker that's inside the LXD to use that to run Docker inside the Libvert? You see where I'm going? Like how deep does the rabbit hole go here? I think we have to find out. I know it does at least support running itself inside of itself because they've got to try it online and that's what it does.

That would be fun. That's all inception right there. They've got to try it online, and that's what it does. Oh, that would be fun. That's a little inception right there. That's pretty great. And so my question to you is, are you looking at moving any workloads over to this? I think I'm going to dabble.

You know, I don't know. It doesn't fit every use case. And, you know, at work, I'm already quite comfortable both with Docker and dabbling with Podman. But, you know, it takes a different approach where sort of the Docker and Podman tooling Docker and PubMed tooling, a lot of it kind of breaks things out and you're kind of assembling it, or maybe you're going to go run it on Kubernetes or Swarm or some other, you know, system. It doesn't really care. You get the pieces and you can go put it together. LexD, it's almost like ZFS in a sense, where it presents more of a holistic, like I can set this up on a server and it's going to go figure out all these things, including

fancy PCI pass-through. It's got all kinds of acceleration if you want. So I think if you had some hardware and you just wanted to manage that and you maybe weren't a particular fan of, say, the Libvert API, which I'm really not, LexD might be an interesting thing to try.

I've wondered, Wes, do you remember why PassWes got less interested and let this technology fade for you? That was a time where I was kind of just running more things. Maybe I was a less busy version of myself, or it was just, you know, containerization was new,

virtual machines were new. I was also at a stage in my career where those were things that I was learning for the first time. Not quite the first, but, you know, kind of first really trying to both learn them so I could apply them at work, which hadn't quite migrated that far, so I was trying to, you know, get on the cutting edge.

I'm not running as many things these days. I've tried to minimize some of my personal infrastructure just to keep things simple, especially I've just moved a lot in the past few years, so I haven't wanted a big home lab that I've been dragging around. And there is the employability of having a really solid understanding of Docker because that is such a normal standard deployment now in the business world. So if you're looking to build out like an employable skill set, I mean, not that LexD isn't used widely, but Docker is probably more widely used. As we've shown for better or worse, you can get quite far with, you know, just a Docker compose file these days.

There's a lot of community momentum behind that. a community momentum behind that. But I think if you wanted to get more complicated or perhaps, you know, your VPS bill was getting high and you wanted to like get a couple of bigger boxes that you were going to handle on your own or something like that, you know, there's a lot of use cases where I hadn't, I just hadn't really thought about LexD.

You know, I wonder if we're not going to discover a theme accidentally. I almost can sense it because my story is about figuring out I maybe am not using the right tool for the i maybe i'm not using the right tool for the job your story is about the right tool for the for the job mine definitely is is it all right brent so all right okay here we go so tell us brent what's been going on for you this week well as listeners may remember i'm still sitting here with alex yeah he's still here people will he ever leave i? Send us your feedback.

Raleigh's awful nice. Alex has got a great place. They're both very nice people, you know, so I can understand, Brent. I can understand. I'd probably be in the same. I'd probably be there if I were you, too. You know, just before the show, he did make me an account on his Mac. Wow, it finally happened.

You got a local account. That is a big step. Probably means he expects you to come back too. Yeah, I think he probably wants me to do a bunch of projects, it seems. So being here has been a real opportunity because I've been able to sort of tap into Alex's expertise a little bit and learn some things that I've had in mind for the last few years. And this might be quite topical after our NixOS challenge. Turns out Alex knows a thing or two about Ansible. And one of the main things I wanted to learn while I was here was sinking my teeth into that a little bit

and to see the differences that I could at least grok so far between Ansible's advantages and those of NixOS, for instance. And I gotta say, it's been super fun. For the last few days, Alex and I have been sort of diving in. I think I'm pretty lucky because he's got all the shortcuts. He basically said in a day and a half, he showed me everything he's learned in five years. Right. It's like you got sent to one of those expert courses. You're at the Ansible bootcamp, basically, except one-on-one. Yeah. And it's been a real treat. I gotta say. Thank you, Alex.

I mean, that's the thing about Ansible that I always thought in the early days was the learning curve is actually quite short, and the complexity can ramp as quickly or as slowly as you need it to. Like, the basics of just installing a package or something like that, we did that in, what, an hour or two? Yeah, I think once you sort of describe

the main structure underneath, installing packages or super symbols, which was really similar to the NixOS process, like wrapping your head around the concepts is kind of a big part of the hard part. While you were doing that, did you have moments where you felt like, oh, this seems like a lot of work

to get something kind of fairly basic done? I think had I not had the NixOS challenge that we did last month, I would have certainly thought that. But knowing that the work you put in has many advantages and is a real investment, then it made looking at Ansible just make a lot of sense from that perspective. I think probably the biggest thing that we struggled with, certainly me as the air quotes instructor, struggled with was Ansible is a collection of a lot of little files all glued together to build this configuration management structure. And I kind of struggled explaining, well, you need this directory and then you need this one and then you need. I mean, I do open source all of my code on my github at ironic badger if anybody's interested the number of times i said well if you look at my code on the on the you

know on github i mean that that repo is is used to deploy all of my personal infrastructure and that's answerable as i understand it so i mean in my defense that's kind of like my living and breathing encyclopedia alex you touched on something that I found a little bit more difficult compared to NixOS, for instance. In NixOS, you can start with just a Nix config that has a default location. One of the things with Ansible that you sort of have to get over a little bit is there's a directory structure and expected files in certain places and an equally thick nomenclature to go with the naming of those files and those, I don't want to say roles, because that's actually one of the names of something and don't want to confuse people too much. But that nomenclature was a little tricky.

And I think you saw me struggle with that. Yeah. So, you know, there's a lot of domain specific language to learn. So you've got to learn what a playbook is, what a role is, what a task is. Those are the three basic fundamental building blocks. And once you've got your head around how those three different things kind of link together, I saw your progress kind of skyrocket. And you went from, how does this connect to that to being, oh, well, I just do this. That's exciting. Yeah, when it all clicks. It's been fun.

And Alex discovered a reason for me to also use some other pieces of software, which was really great. I think Proxmox was a thing that you threw me at, to get a VM going so that we can play on this. Second mention of the show this week. Yeah, because we could use snapshots on Proxmox because the VM storage is just backed by ZFS. So it's really easy just to create a snapshot and run some Ansible,

break some things, and then roll back the snapshot. This has been a really, really good sesh for you at Alex's Place because you got your first hands-on with the Raspberry Pi, which we talked about in Office Hours 3. You've now got your first taste of Ansible and Proxmox. Like, Brent, this is like major milestone stuff. Yeah, I got to say, it feels like a whole new world has opened up for me in a way. The reason I wanted to bring that as my topic today was I hope I can encourage some other people to even get their, you know, sink their teeth in a little bit and feel some of that excitement, especially those who have joined us in the NixOS challenge. Cause I feel like that was a part of the conversation that came up often was, Hey, what's the difference between the two? Which one's more suited for me? And I think it's worth diving in even just,

you know, for an afternoon or something and seeing what it might be able to do. So I want to take a sidebar conversation that you and I had off air and bring it into this conversation because I think it applies. And that was, you were talking about how you had just deployed Umbral for yourself and you felt like, you know, it didn't really work well with this new Ansible paradigm

where you're kind of prescribing a system because Umbral itself is essentially a small little orchestration system in a box. But I think what we realized there is like, there's a fundamental shift perhaps in the way you look at how to build things now. Can you expand on that a little bit? It's true.

Yeah, so to give a little context, the last time you and I, Chris, talked about Umbral, I believe it's here on Linux Unplugged, actually. It was maybe a month or two ago. Which I should say is like this, you can set it up yourself as a self-hosting platform to install lots of really awesome self-hosting apps. And it's a Bitcoin node, although I guess they're going to make that optional in the future.

It's a platform that sort of showcases some premier self-hosting apps, and it manages all of it as Docker containers with its UI. So it itself is a platform in a sense. Yeah, and it can also manage a few Docker containers for you in sort of their, I don't know, umbral store, if you will, which is actually really nice for people running this, I don't know, on a PI or something like that who don't necessarily have that much Docker experience. So that's kind of neat. So back then I sort of installed it. I think we did it in a VPS, Chris.

Do you remember? Probably. Yeah. Probably on Linode. Yeah. And I kind of followed their documentation, you know, note for note. This time around, I thought, well, Alex and I are playing with this Ansible stuff. Maybe this is a nice way I can just sort of integrate that into Ansible and have a little tiny project to, you know, learn a thing or two.

Outside of having Alex help me. So, you know, got up nice and early this morning to do that. And I did discover a difference in how I understand how software is delivered. Help me maybe explain it if I don't get it quite right, Chris, but it seemed to me that Umbral, the way it's been packaged, you know, it uses Docker under the hood, but it kind of tries to obfuscate that from you a little bit. The way they package it is for those who don't necessarily have that expertise or that experience with some of this containerization toolkits. And they're trying to make it as simple as possible.

You know, their default is to run this on a pie. They have a pie image, which they call, I think they call it an Umbral OS, something like that. That's really great because you can kind of get up and running with the defaults pretty quickly. What I discovered trying to integrate this into my Ansible setup, the way Ansible works is that you can use a few, Alex, help me with the wording here, but you can use a few modules that are predefined in sort of the underlying Python code to do tasks that are really common. So let's just say you want to install HTOP. Well, there's a package module that'll just sort of know how to handle things. So you just pass it a few little variables like the package you want, I want HTOP and what state

you want it in. I want it to be present. And then it just goes and figures out the rest. And then it just goes and figures out the rest. What I ran into trying to get Umbral into my Ansible sort of setup was that they want you to just kind of curl a script. And most of what you do with Umbral is based on scripting that they've included when you download their tarball. And that's great because it's super simple to get up and running just by running one script. And you don't have to necessarily understand the underlying technology in there. Exactly. Yeah. But I was trying to use it as a way to understand the underlying technology, which I think is going a little bit outside of what they were hoping someone would do. You kind of jumped to a hundred too, right? Like you're getting started with this thing.

You chose a system that's like fighting against you and not designed to play nice with your new payments. Exactly. So what you've really come across, but now from the other side of it, is the difference between something that's designed for consumer use and something that's designed for production grade use. And not that Umbral can't be used in production. I mean, I'm using it in production. And I like that because this is a one-off isolated system that isn't part of the overall JB infrastructure or anything like that. I like that because this is a one-off isolated system that isn't part of the overall JB infrastructure or anything like that. I like that Umbral manages and orchestrates all of this for me. I didn't use their OS, but I'm running some of their scripts.

I could also see a future where like this probably is not likely, but say we had a future where boosts were representing like a third of our revenue or something like that. Well, then I would probably want to migrate that to an infrastructure that was a lot more managed and a lot more reproducible because it would be a real serious endeavor. Right. Like right now, I'm in the very early, very early stages where beta testing, all of that kind of stuff. It's a couple of years down the road before I really expect this to really gain traction. And so it's not a, quote unquote, production grade setup yet. And, you know. And some people would probably say that's not a great idea, but it has the advantage of letting you get started fast. Yeah, I would say, Chris, when you and I in January wanted to play with this when I was at the studio, we were able to relatively quickly get up and started with it. I think most of our time was spent understanding all the concepts around it, but getting the software up and running was super straightforward, wasn't it?

I find it interesting how the Umbral project provide all these scripts and they're essentially an opinionated set of scripts that say this is how we think the world should be and chris i i know that you with home assistant don't like it when they do that so what's different with umbral versus home assistant well the key difference is they did let me deploy my own OS. So I am managing the OS the way I traditionally manage an Ubuntu base, right? So I still am applying all of that at that layer. And then the stuff that runs in the container with the orchestration script, I just let them manage. But the underlying OS and storage and all of that is under my control. Linode.com slash unplugged.

Linode makes it simple, affordable, and accessible to deploy and manage a system for yourself or maybe for your customers in the cloud. And they do it at a better price and better performance than those large hyperscalers that have just endless options and want to basically lock you into their platform. Linode's how we run everything that we've built in the cloud for the last few years.

And you can tell when you use it. If you're a longtime Linux user, you can smell this kind of stuff. You can tell they love Linux. That's where the whole inspiration for the product came from. They started 19 years ago when this stuff was just getting baked into the kernel. They learned along the way that customer support is critical. So they've invested in having the best support in the business. And this is probably the number one signal I get back from the audience. Number two

is definitely performance. But I often hear like that because somebody just like had some sort of disaster happen. They blew something up or I don't know. And they contact me like you wouldn't believe how they saved me. And they stayed on the phone. It was the first person that answered. But on top of that, they really just have some of the best options. And they have 11 data centers for you to choose from. Each one is screaming fast because they are their own ISP. They've invested in MVME storage. They've invested in Epic AMD processors. So they've really, really made sure these things perform. And then they have a bunch of great backend features like firewalls and object storage and a powerful DNS manager, Kubernetes and Terraform support, of course, Ansible support. So they really, really make it a compelling option if you're doing infrastructure

as code. And along those lines, and I'll put a link in the show note, they have released a white paper, I guess you could call it as such. It's really an ebook, I guess. It's about infrastructure as code. And it goes through things like Terraform, Ansible, Puppet, Chef, and Salt. And it kind of explains all of it through the entire book. And it gives you a sense of what each technology is about and how it might fit into what you do. It's just a really good resource. And I've mentioned it on one other show. And I got really great feedback from the audience on it. So go check the show notes for that.

Or you can just Google Linode Infrastructure as Code eBook. And you'll see what I'm talking about. But go get $100. Go kick the tires go build something go learn something it's a great opportunity and they have a bunch of one-click app deployments too you can get things like a minecraft server next cloud gitlab lots of stuff one click go check it out go to linode.com slash One question I did run into coming back to Ansible was how to install non-repo software. Because the way Ansible, at least I understand, is the straightforward method for installing software uses your auto-detected distributions package manager to install software. And one of the questions I ran into pretty quickly was, well, how do I install software that's not in the package manager's repos?

And Alex, you had a pretty good answer that I think is worth sharing. There's a few ways, and you're quite right. The default package module will auto-detect whether it's running on YUM or DNF or APT or YAST or or pacman whatever it might be right and that's one of the beauties of ansible versus say the nix approach right is it transposes across linux as a whole the entire ecosystem pretty much where nixos is you know they're they have their own packaging system that you're pulling software from and you learn the nixos thing and you now know nixos but if you you know get a new job at somewhere that's running uh suzy you've got to learn suzy now so that's one of the key things but coming back to

your question there's a few different ways ansible has the concept of ansible galaxy which is like a bit like Docker Hub right so you specify a remote repo a remote source that you can pull in in the case of your question originally last night it was Docker how do I install Docker you can pull Docker from your repo in your package manager but quite often the version of Docker that's in there is out of date, or it's not packaged very well, or it has a weird name. And so what I tend to do is I tend to rely on Ansible Galaxy roles for that. And Jeff Geerling, who was on Self Hosted a little while ago, he writes a metric S ton of different roles that he uploads to ansible galaxy one of them is docker and so you can reuse jeff's expertise to install those sorts of packages by specifying a couple of lines

in your ansible and i do mean only a couple which is gearing guide.docker install that role from ansible galaxy and just magically Docker will be installed. Yeah, it is sort of magic. And you can do that for a bunch of other services. Tailscale was another one that we did. So we went away on Ansible Galaxy and found a role that installs Tailscale. And it actually did some really neat stuff. We ended up having to go and grab an authentication key from the Tailscale admin console.

And two minutes later, not only was Tail tailscale installed but also this role had authenticated that server with my tailscale account without me even really realizing that's what it was doing and i just found that really cool i'm reusing this this tailscale experts knowledge that i didn't even know i could do that yeah they're like these little as far as i understand it so far in my little mental model, they're like templates to use in your Ansible roles. And you asked me a question last night about how do I know that these Galaxy roles are safe to run? The beauty of all of them is that they're all open source. So the Jeff Geerling stuff, for example, if you go to Ansible Galaxy and search for jeff geerling there's a github repo link in every single galaxy role you can go and look at the code

and one of the things i truly love about ansible over any other configuration management software is that it executes tasks in a linear fashion so if you write task one do this task two do that it won't try and be clever and cute and do these things in the order that it thinks is the most sensible. It will execute them like a bash script would, line one, then line two, then line three. And so it's very easy to understand that this task happens here and then it does this.

And for me, as a layman at the beginning when i was learning ansible i found that so helpful because there's a there's a whole extra layer of complexity that i don't have to wade through and in terms of reading open source code we often say this in the community of i'll just go and audit the code it's open source go audit it and you'll be fine but with ansible you actually can you can see that this five line task does this this five line task does that yeah it's quite human readable i would say yeah for sure and for me that's one of the huge huge benefits of it so you've got code reuse out of the box that's a massive devops checkbox facilitated there you've also then inadvertently created a source of truth

which is your source code repository which you then should be putting into git or something like that to manage the version controlling there the only real pitfall really is if you start using ansible and then decide oh i just want to quickly install a package on the command line natively like you know want to do sudo apt-get install htop right and you don't put that then into your ansible repo and if you do that over a week you've got a very small amount of configuration drift if you do that over a month it gets a bit bigger but very quickly after a year or so you end up in where the fact that your desired state in ansible is nowhere near what reality is on that server and

so once you commit to doing it with configuration management, you've really got to go all in and try and do it completely forever that way. I've been really enjoying this journey and I hope to continue it. Oh, right. I go now. I go now and I just realized that my topic is encrypted in the show notes again. I like to do that. Create a little problem for yourself, eh? Do you remember the password? No, but you know what I have is I have this handy little app right here.

It's a Rust app, and it will basically encrypt text as AES, whatever you choose. I chose 8-bit AES. Oh, I love security. I don't really care. But I thought I'd start with a story that was sent in to me by a whole bunch of you out there. And it's about something I haven't talked about very much. And it's one of my absolute favorite Linux boxes,

my little Starlink dishy. And there is this week a new announcement for Starlink owners about a portability option coming. So you can now take your Starlink owners about a portability option coming. So you can now take your Starlink with you. You're not locked to the region that you signed up for. And I had noticed they were beta testing this because I got bit by this region availability thing. I moved out of my region. Oh no. Set up at a new place. Just no internet. And there was no capacity for me. But then a couple of days later, it just started working.

And the thing about the Starlink setup is if you use their app with their router, you can open up the app and go into debug mode. You'd like this, Wes. You go in there and it just tells you all the raw log information that the little dishy has. And I could be wrong, but I think there's either one or two different Linux boxes inside that little dish that talks to the satellites. And you can see what it's doing. And in there, there was this little flag about portability or roaming or something like that that had flipped to true.

And it turns out they were beta testing this and I was benefiting for free. And now they're rolling out for a $20 add-on. I had that. I mean, so many people sent that in to me. And it is really cool. But unfortunately, I got a bit of a tale to tell you. You see, I was packing up my Starlink the other day. Yeah, that's right, Wes. I don't like how this starts. No, it's not good. You see,

I was getting ready for our trip out into the woods. I put little dishy into stow mode, as they call it, and I packed it up. And a few days ago, i.e. Friday, I decided it was time to hook it up, get some additional tubes into the home because we needed more internets than kids, you know. I get it all set up. I put it out in the grass field. I power it up because it's on a smart plug. This is one of the ways I control it during the night as I sometimes turn it off and um it just sort of sat there limp dead never looking up at the stars never asking what else is out there can I be more than just what I am it just sits there limp and stomoed and I'm watching the app and all of a sudden this little red air comes up and it says Starlink motors are stuck and And I walk up to it and I can hear like a weird buzzing, but nothing else is happening. I reboot it.

I mean, I went through. Oh, no. I probably went through six hours of trying to get this thing to come back to life, just like refusing to accept my harsh reality. I mean, I was bummed, bummed, Wes, crushed. Because, you know, you go from cellular to Starlink and it's a pretty big upgrade and i was so crushed and so i just refused to accept i thought maybe i could move it just right and get the motor what if i smash it a bit or jiggle it or toss it up in the air plus this is the round dishy which is reportedly overbuilt and more robust and has more antennae in it so you broke

the strong one yeah and the new square one which smaller, is supposedly not quite as robust. It's more like- It's quite spelt. Yeah, it is smaller. That is nice. So I opened up a support ticket and now I understand why Starlink business is $500 a month. And the number one feature is prioritize support because I have heard nothing back. One automated reply and that's it, right? And I'm trying to get all my ducks in a row i'm trying to like pre-upload the debug info but i look like an idiot because when you use the app it has like this copy debug info option really handy i thought i'll get a jump on this because in the debug info it clearly shows that the motors aren't working right and it and it shows that it's not obstructed otherwise it's incredible what it knows about itself actually it's remarkable

when you hit copy in the app it just seems to be copying the debug info for the app and the router, but not the Dishy. So I realized that I'm like, oh, my bad. So I re-uploaded again thinking, well, this time I must have captured it. Nope. So now I have two tickets where I have like bad debug info in there and I look like a total noob. But so I've done like three responses to this ticket and I've heard nothing back yet i can kind of take a guess because i've been in this this spot before i can kind of guess where the satellites are like during the day so i i kind of like you know like weekend at bernie style kind of like positioned it in the direction of the satellites or like you know if you're old enough you might remember like having to tune in rabbit ears on your television yeah i wouldn't do

i'm doing that trying to point this thing up into space right and it gets a bit of a connection and so i can download stuff for a bit and i'm getting like 100 megabit connection and everything's good but then as the satellites move it doesn't adjust and i lose connection you might need like a home assistant integration that can automate right retuning i need like a little motor yeah i do i looked into like used dishes but they go for like a thousand bucks on ebay and it's not even clear how you even associate that with your account because these dishes you know they connect to your account you need like a fifth support ticket for that yeah i'm gonna definitely need a support ticket so i can get like temporary connection

but then i lose it but i've heard like when support does engage with you it tends to be really good but i was looking at the option. So they have this ridiculous $500 a month Starlink service that they're going to roll out. Have you seen this? It's their business version. And supposedly it's a more robust antenna. And you get more like 300 megabits instead of like 150 megabits. It's nice.

About the same latency. Okay. But the number one feature they list on there is support and i i it dawned on me right now like i could almost see the the utility of it for the business because without i realized without starlink i'm sol on remote broadcasts really or i go back to cellular which was always so horrible it could be pretty great you know like if you drug it with you into the studio yeah studio has internet out, that'd be a solid backup sometimes. I mean, it's just, I can't justify the price now, but if I canceled all my cellular data plans,

which I have three different cellular data plans, I'd probably almost pay for it. Connectivity is a big deal, right? I mean, this is what my whole job's online. So when I'm broadcasting remotely, which I'm going to be in a couple of weeks, because I'm taking Jupes in to get her ready for a summer road trip, I'm going to need to do, I weeks because I'm taking jupes in to get her ready for a summer road trip. I'm going to need to do, I think at least coder from Southern Oregon. And,

um, I was planning to rely on Starlink to do that. And I, I don't really know what happened. I'm thinking either it fell or something in transport. Like I tried to set it, but it's in a storage base. I couldn't see what happened.

Right. And I just took it out and it wasn't working. My guess would be the snails got in there again might have been there were there were a surprising amount of snails living in the uh in the pole when i took it down i looked in the pole and there's like a half a dozen snails all over the thing the good news is that it seems like spacex is serious about supporting hashtag van lifers and rvs in fact el, Elon Musk tweeted that Starlink is awesome for RVs, camping, or any activity away from cities. Yeah, I mean, it definitely is, right?

I would describe it as life-changing. I really would. I mean, imagine going somewhere where there's nobody around for a dozen miles, you don't have any other connectivity, and you put this little dish up in a field and you're getting 150 megabits down. I can't even with that. And it's sometimes 20 milliseconds or so,

which is fine for VoIP and stuff. It works. I'd love lower, but it gets the job done. So it really is a life-changing kind of thing. And I just am always, always happy to see use cases like this for Linux too. Like that's always in the back of my mind that I'm using Linux to talk to space. Now that I've gone through this experience, I can kind of see, like, say I was doing the shows out of the RV on the regular and not out of the studio.

I think I'd see I could see the logic in getting a business grade tool. Right. Using the right tool for the job kind of a thing, you know. In the meantime, I'm hoping to get a good support experience if anybody at starlink is listening uh look up my ticket for me help a brother out he's been waiting a couple days he'd really like to get connected back to the internet again bitwarden.com slash linux get started with a free trial for teams or enterprise or even a personal plan at bitwarden.com.

Bitwarden is the easiest way for an individual or a business to store, share, and sync all kinds of sensitive data. We use it here for our passwords, of course, and we love that Bitwarden is open source. It's trusted by millions in their community. And of course, businesses are using it worldwide. Our business, Jupyter Broadcasting, uses it. It's a really handy way to work with teams of people when you have common resources you need to sign into. And that's really where the password hygiene can fall down.

And now recently, Bitwarden has also rolled out username generators, which, man, oh man, I can't believe I hadn't thought of this earlier. So now not only will each one of my sign-ins for these different services I use, especially when it's not like a public profile, it's just something I'm using to manage an account or something like that. I'm not sure why I was ever using the same username for all of them. And if your email provider supports it, like Gmail, where you can do the plus in the address,

Bitwarden figures that out too. And it will support that. The username generator is such a great idea. And then you save it all in Bitwarden. Using a good password manager to keep different passwords for different websites is probably one of the number one things you can do for your security online.

You know, it takes sometimes a hard lesson to figure it out, but I bet a lot of you already know this. So if you know somebody out there who's maybe doing things a little less than ideal, point them to bitwarden.com slash Linux. Get them started. And then for our community, one of the things that gives us peace of mind is that if you want, you can self-host.

I've opted to use Bitwarden Cloud. I feel like they're probably going to do a better job of managing that than I will. But I know some of you out there just prefer to run it all on your own infrastructure. I like that Bitwarden gives you that flexibility. And then I think the other thing that's key to it is that there's actually a really large active community on their forum and on Reddit to help with all that kind of stuff. So you're not just off on an island by yourself either. It's really nice and it gives me the confidence to use Bitwarden and it gives me the confidence to recommend it to our audience. So go get started for free at bitwarden.com slash Linux. It's a great way to get started, make your online security better, and you support the show. And recommend it to somebody who needs this advice. You know they're out there.

Bitwarden.com slash Linux. Majid wrote in and he pleaded with us. Could you read my feedback on your lovely show? That would make my year. So I thought we could do that. And there's also a question here. If you are in contact with the NixOS devs, could you ask how accessibility is going with the installation? I'm blind and would like to try out NixOS and the NixOS challenge if

possible. All right. We'll put feelers out there. And if we hear anything back, we'll relate here on the show. That is a, that's a great question. And I wonder too,

if, and maybe somebody could let me know for my own edification, is this something that a VM software could help solve, like a VM that supports screen reading? I wonder too there, I don't really understand how that works. So I'd love to know for my own education, if that is indeed in the works, and if there is another solution our listener could employ. And now it is time for the boost. Our first boost is from the golden dragon with 4,500 sats. How'd this Linux time machine get here?

Oh, well, on my journey, I would have stuck out the pain points I had in college with my Wi-Fi drivers not working and just figured it out. Back then, I didn't have the same patience I have now, and that would have made all the difference. I love the idea that the audience took this idea of the time machine and sent us a few of their own ideas. Marcel wrote in with a similar kind of theme for 2000 Sats. He said, I wish I could say my biggest Linux regret was

rising I3 gaps instead of paying attention to my undergrad classes. Probably my biggest regret was not learning NixOS and Docker back when those things were just coming out. Thanks for taking us along on the time machine. It was fun. Another great boost came in from The Muzo. 2000 sats, thank you very much. Regarding ARM hardware, there are the SBBR and EBBR specifications that are supported by recent single board computers and ARM server hardware, which allows generic ISO images to be booted on compliant hardware. Been listening since early 2020. Keep up the great work. Well, thank you. The amazing rain sent in a boost and caught us. Did you guys announce the results

of the Nix server poll? So in our next OS challenge results, we asked the audience if we should consider nuking and paving our local server setup here, going from OpenSUSE Tumbleweed as the base OS and going for more like a Proxmox Nix kind of setup. I'm feeling nervous.

So we did get the results and we had almost 300 respondents, which I've actually found even that low. It can still be representative. I think people didn't vote on it as much because it was buried in what was our most show notes we had ever had for links. Of course. But we asked, should we use NixOS as our cloud and local server OS going forward? And with 57.44% of the vote the audience said yes keep us posted with 42

percent of the vote going to stick with seuss it was a 57 vote for yes go with nix that feels like a really close result yeah i don't know should we do both what do you think wes how do you how do you feel about this, Wes? I think we have some work to do. I could see also kind of having our cake and eating it too. Like if we do VMs. Wait, you brought cake? I should have. I should have said that. But if we did VMs, we could do Nix base and then OpenSUSE Tumbleweed VMs or something, right? Like there's still maybe a role for tumbleweed here, right? All right. We install OpenSeuss and then we install Wexty.

Yeah. And then we run Nix. Right. This again, this election is being called into question. You see, we had some meddling with the last election. There was lizard meddling. And now it is called into doubt this election. I don't know what to do.

We'll have to keep you posted because clearly we have not figured out this controversy yet. We should never have left Arch. You know what? Maybe that's the lesson. Alright, our last boost. 1200 sats from The Computer Guy. I absolutely love the shows. They inspired me to revive my own podcast.

Oh, that's great. JB is my favorite thing to tune into online. Well, thank you, the Computer Guy. Also, thank you to some booze from the Spherical Cow. He sent in a couple of thank you booze. We also have a few of you out there that are streaming in sats while you listen. Of course, thank you to our members. We are actually working on boosts for the member feed because that's a frequently requested

feature, but we have to build out a lot of other infrastructure first. So it's not going to be anytime soon, but it is on our radar. We are detailing out like those new website plans, things that will be new member features, stuff that we're building out over the summer. We are detailing that out in officehours.hair, starting with episode three and then kind of doing it fortnightly from there. So join in. Listen to that if you're curious about where all that's going. We have a pick.

I think you found this one, Wes. Yeah, but you, you tried it out. I think it's called Liar Bird. It's a voice changing app, simple and powerful voice changer for Linux written in GTK3. And it takes your microphone input, whatever it might be, and then it creates a virtual output that you can record in any app that you use to record your audio. And the idea is that it's supposed to make you sound kind of funny. You can make yourself sound like you're on a sketchy radio connection from far away. And it has like a Darth Vader mode and it has a gender changer

mode and it has like a Darth Vader mode and it has a gender changer mode and it has like a megaphone mode, just random stuff like that. And it's a nice looking app. Simple. You didn't try the Darth Vader mode? Oh, I tried them all. I just could have, you know, I didn't want to be obnoxious with all of the different clips. But yeah, before the show, we were playing around quite a bit. It's fun. If you've got a great pick idea or something that you're like, why these guys never talked about x on the show send it in to us go to linuxunplugged.com contact we got the

contact form there and of course you can always boost it in by getting a new podcast app at newpodcastapps.com and the great thing about those apps is it's not just boosts it's the entire podcasting 2.0 spec and the more people that use those apps the more podcasts will adopt things like transcripts and chapters embedded image and all kinds of nice things that podcasting is needed for a long time so go get a new podcast app at newpodcastapps.com and of course you can always grab linuxactionnews.com in one of those apps well at least linuxactionnews podcast at linuxactionnews.com i think that's how it works anyways Anyways, it's our weekly news podcast. And if you're not getting that, then you're missing out on what's going on

in the world of Linux. And we welcome you to join us live. Come on in. Wes sets up some chips and dip for the live members. They're virtual chips and dip. Could you imagine sending out chips to everybody? We don't want to make a mess. Maybe when we make a big time, big time,

we'll send out chips to everybody. But in the meantime, you get some virtual chips and dip when you come over to JBLive.tv. We do it on noon, around noon. Right around nooner, we start getting started. Sometimes there's virtual tacos. Sometimes there are indeed. And you can hang out in our chat room or our mumble room and participate here in this podcast.

We like it when you do that. That's real great. We also just like it when you listen in the chair with a friend. Just like all that kind of stuff. We should also mention, Chris, I think we're recording in a new Office Hours this week, aren't we? It'll be live on Tuesday, Brent. All right. Nooner.

All right. Nooner again. That's right. Yeah. Yeah. Let's talk website stuff. Let's talk project stuff. Let's talk podcast stuff.

All of that on Tuesday, OfficeHours. The Mumble Room will be open as well. See you next week. Same bad time, officehours.here. The mumble room will be open as well. See you next week. Same bad time, same bad station. In the meantime, links to what we talked about today at linuxunplugged.com slash 457. Let's see, you can go find, where are you on Twitter, Wes? Give a plug.

At Wes Payne. Yeah, there you go. The network's at Jupiter Signal. Catch more Alex on the self-hosted podcast at selfhosted.show. And I think that's all the plugging we have. I'll just spend the last few seconds thanking you for listening. If you've written in or you're thinking about it, do it.

It's a big part of the show. It helps us think of what you want to hear, make new show content, all that stuff. It's really useful. We like your feedback. We're making the show for you. Yeah. LinuxUnplugged.com slash content. All right.

Thanks for being here. We'll see you right back here next week. Thank you. you