LINUX Unplugged - 428: Pi for the People

Episode Date: October 20, 2021

We try out POP!_OS on the Raspberry Pi 4, and chat with its creator Jeremy Soller from System76. Plus our thoughts on the perfect Linux laptop right now, and the clever initiative Valve just launched for the Deck. Special Guests: Jack Aboutboul, Jeremy Soller, and Neal Gompa.

All right, well, you remember that old 3D pinball game on Windows? I mean, just about one of the only good things on Windows? Yeah. Well, someone managed to decompile it, get it running on Linux, and put it on GitHub. Oh, yeah, yeah. You know, I should see if I can get a build of that. I'm going to try real quick. See if we can get that.

I mean, we're not doing anything else right now. You might as well. Actually, it was updated recently, too to resizable window, scalable graphics. Fancy. Yeah. Actually, pretty easy. All right, Wes, brace yourself. Let's see if we can launch it.

Let's find out. We'll find out live together. Oh, this is great. Okay, let's give it a shot. Let's see if it works. Ah! This is amazing! Look at that. Uh-oh, mumble room.

You might just have to do the show without us today. Hello, friends, and welcome back to your weekly Linux talk show. My name is Chris. My name is Wes. And my name is Brent. Hello, gentlemen. Pinball aside, the show must go on. And this episode is brought to you by a cloud guru,

the leader in learning for the cloud, Linux, and other modern tech skills, hundreds of courses, and thousands of hands-on labs. Get certified. Get hired. Get learning at a cloudguru.com. Coming up on the show today, well, I just got done trying out the Pop!OS 2110 Beta on my Raspberry Pi 400, and Jeremy from System76 is going to stop by in a moment to chat with us about it. Plus, we'll also have our thoughts on the perfect Linux laptop right now, a little bit

of community news in the show, as always, some pics and more. So just get right to it. Before we do anything else, before I even attempt to introduce Jeremy, let's say time-appropriate greetings to our virtual lug. Hello, Mumble Room. Hello. Hello.

Namaskaram. Hello, hello. Nice to see some solid numbers in there today. And they all wear matching outfits. Wow. I don't know how they coordinate that, but it is impressive. You know, that matrix channel is useful for all kinds of things. That's pretty great.

Well, so this last week, Pop!OS 2110 entered the beta phase. And, you know, we've been keeping an eye on things because it had been teased that there might be a Raspberry Pi version. And sure enough, it appears an image for the Raspberry Pi for Pop!OS 2110 has been published. As soon as we saw that, well, we just had to reach out to Jeremy Saller, principal engineer at System76 and Pop!OS maintainer. And he joins us now. Jeremy, welcome back to the show. Thanks for having me. I love being here. Thank you for being here. And I know behind the scenes, you told us you're a little under the weather and you still made it. So we really appreciate it. And I had to here. Thank you for being here. And I know behind the scenes, you told us you're a little under the weather and you still made it.

So we really appreciate it. And I had to reach out to you last minute too, because as a Pi fan, I was pretty happy to see this. I think the Raspberry Pi with external SSD storage as the primary disk is starting to get to a pretty usable desktop state. Like it's in the zone now.

But I tried this out with an SD card, Jeremy, and it's definitely usable, but it's slower for sure. And I'm wondering, as you've been tinkering with this, do you see storage, the way storage works right now in the Pi as one of the limiting factors for like a butter smooth desktop Linux experience? Yeah, so I've been, I'm the primary person behind this. I've been working on it in kind of pieces of chunks of time that I can cut out of my normal Pop! OS duties. And I do everything on an SD card because I want to see how bad it is for the average person.

I'm still using the Pi 400. So the recommended configuration is the Pi 400 because it has a slight CPU overclock with four gigabytes of RAM or more and using an SD card. And yeah, definitely the weakest part of the whole solution is the storage speed. If you start using a, I have an external 10 gigabit per second USB 3.2 Gen 2 NVMe adapter. That's a lot of jargon, but it's fast. It's probably a hundred times faster than an SD card. So when I'm using that, it definitely, definitely improves the experience to the point where I feel it is similar to a laptop

from about five years ago, a kind of lower end laptop from about five years ago. Which is fairly impressive when you consider the entry price of the Raspberry Pi, right? Absolutely. For a hundred bucks, you can't really do much better. So, and I mean, I have to ask this question, right? This is the obvious question, but aren't you just making it easier for people to not buy System76 machines by doing this? Absolutely not. And, you know, I thought the same thing. Like, is this going to cannibalize our market?

And I feel that when people have a low-cost entry point, when they do finally have the budget for a more powerful computer, where are they going to go? They're going to come back to the manufacturers of the operating system that they love. And I know this happens. So I think we've only amplified the thirst for our products among our customer base because they already had Raspberry Pis. They already had our products. It was a very closed Venn diagram of the two, making them happy only amps up our products even more. Yeah, not to be cliche about it, but it's a halo effect. Right. Okay, so I was curious because I'm not a developer and you're pretty deep into this stuff.

Do you find that it makes the x86 version of Pop!OS better by doing these kinds of ports and working to get it up on the Raspberry Pi? It definitely does. Part of the porting process was to do a complete... 2110 has a lot of changes, and we haven't really explained those well yet. And there's a lot of low-level changes that have happened that you're going to see later on when we explain it. So we are going to be switching to the stable kernel releases. We're working on releasing the 5.14.13 kernel currently on 2110. That will hopefully have us on the mainline kernel indefinitely. So rather than utilizing the Ubuntu kernel and then applying patches to it, the intention of that is to have a closer link with the upstream kernel development. The major blocker to that is the NVIDIA driver. That's one of the changes. Another change is we're moving off Launchpad and we are moving our repositories to our own server, apt.pop-os.org.

Because Launchpad had all the build infrastructure for ARM64 already, we didn't have that. So we actually took a System76 Starlink server, which we did not, honestly, we did not sell very many of, but it was one of our previous ARM offerings, and set it up as a build server so we could do ARM 64 builds. And we re-architected the packaging of Pop!OS to allow for Pop!Server. So a server edition of Pop!OS. So all of this stuff happening basically because of the Raspberry Pi port. And we also redid our build system to write it in Rust.

Nice. And to ensure that it was able to spit out builds for x86, for ARM64, and hopefully in the future for RISC v64 GC. So that all adds up to a lot of low-level changes to ensure that we are in place for what I see the architectures of the future being, which are AR64, RISC v64, GC. That's quite a bit of knock-on effect. I didn't expect all of that. I'd like to jump back just to the Launchpad transition. So this is everything, right? This is the build tools, this is the bug tracking, and this is also the package hosting. Yeah, all of that was moved off for 2110. So

now it's all in-house Pop!OS stuff. That means we still do use the Ubuntu mirrors for Ubuntu packages, but all of the Pop! OS packages, which is about 100 packages, are coming from our own app.pop-os.org server. Okay. And so if I understand you, essentially the next release will be based on 2204, which will be an LTS Ubuntu LAN, but you're going to maybe be using a different kernel from upstream, a mainline kernel,

with an LTS Ubuntu land on top of it? Yeah, we hope to continue. We're actually hoping to release that 5.14 kernel even on 20.04 LTS. So we'll be backporting it to our prior LTS release as well. The intention is that Pap!OS will be the distribution that will run on any hardware possible with absolutely no compromises. And the biggest compromise with the Ubuntu kernel development cycle is the SRU process and the fact that new hardware can take

at least two weeks, but usually a month, to get support. And the biggest problem with the Fedora process, which does something similar to what we're trying to do, is they don't really care about the NVIDIA driver. And I understand from their position why that would be the case. Anybody shipping the 5.14 kernel right now does not care about the NVIDIA driver. That includes Arch, that includes Fedora. The reason is that the NVIDIA driver has no stable configuration, no bugless configuration that works on the 5.14 kernel. The last NVIDIA driver that was released, the 470.74 driver, had issues with

G-Sync. It was the first driver to support 5.14. So you're either breaking G-Sync configurations or you are releasing the 5.14 kernel. You can't do both. So our intention with switching to the mainline kernel is to ensure that these regressions don't happen as often, that we have as close to mainline as possible Linux kernel that also supports all of the available hardware in the entire desktop ecosystem. I mean, that obviously would make Pop!OS compelling to both gamers, but like you said, people that want other architectures. But the big but that I see is it seems like it really puts a lot more of the burden of testing on you guys.

So are you thinking like you'll have some people in-house running the kernels for a bit, testing them, dogfooding them? Or what are your thoughts on how to accomplish the testing aspect? I think it's very important for us to increase the size of our team. And we're doing that now. So we just put out a job posting for PopOS developers. The only requirements are that you know Rust and that you have a good internet connection. We wanted to limit the requirements so we would get as many applicants as possible.

And we're going to be hiring multiple people from that role. It's not just one person being hired from that role. So we're going to build out the engineering team. And I believe Neil has something to talk about. I might have misspoken on Fedora, so I'd love to hear that. Yeah, so we absolutely do care about the NVIDIA driver with the latest kernel stuff. Part of the complexity here is really getting NVIDIA to care. Like from a community perspective, we try to engage and make sure that things are patched and we try to do the best we can in terms

of testing. the major complexity uh issue we have right now is that um so for example you brought up g-sync i don't know anybody on our team on the team that actually like checks this stuff that actually has a g-sync monitor and knows how to correctly test for g-sync like it's not a question of whether or not fedora wants to care about it or not it's a a question of we just don't know how to. We don't know how to test and make sure that that works. I completely understand the amount of hardware out there to test. And the reality is NVIDIA doesn't really care that much.

So they're happy to leave. When 5.14 came along, it took them three months to make an NVIDIA driver that actually worked on 5.14. And then it had the G-Sync regression. So, yeah, the big problem is NVIDIA. But I feel that if there is somebody like System76 that does have the ability to test, and we do the same thing, we try to do the same thing Fedora is doing, we get to mainline kernels, then maybe it won't be such a problem for Arch and Fedora anymore. If we're all targeting mainline kernels, and we can start saying, you know, hey, as a team, as a, you know,

desktop distribution team, Fedora, Arch, Pop!OS, and all the others, this new kernel has regressions with the NVIDIA driver. We know that a significant portion of the desktop Linux market is using NVIDIA cards. Even if the upstream Linux developers are willing to break the NVIDIA driver, we still do have to have something that works for the people who have NVIDIA GPUs. Absolutely. That's always been one of the best and most important roles that I think distro maintainers can take. And I know you guys also have customers that you want to keep happy,

which is one of the reasons why I think it actually ends up working out for all end users. Wow. There's a lot to unpack there, Jeremy. It seems like there's a lot of things in motion, including things like a pop server and other things that may be down the pipe. I'm wondering if you have any other kind of anecdotes to share with the Raspberry Pi, like expectations you might want to set or anything like that, because we're going to link to it in the show notes. And I have a feeling some people will check it out after the show. Well, I'm pretty happy with it. Just based on trying out Raspberry Pi OS and Ubuntu on the Pi, both Ubuntu server and Ubuntu desktop on the Pi. I think it stacks up well compared to those. You definitely need the four gigabytes RAM or more. And I would definitely prefer if people were trying USB 3.0 or faster storage. Other than that, it's pretty much stock Pop Desktop. We haven't really made many changes at all to adapt it for the Raspberry Pi

other than making sure all the packages were building for ARM64. Right. Very good. Well, I enjoy it. I think it makes for a, my ideal little use for the Pi 400 right now is kind of like an emergency go-to machine that has my email accounts,

my calendar, my web browser logged into my profile that even if I forget my laptop or my desktop is crashed for some reason, it's like it's my go bag of computers that is standing by.

And I think the Pop Desktop is kind of perfect on there. So I just think the way it all works with the launcher and all of that has been ideal. Jeremy, thanks for coming on the show and explaining all this to us. Yeah, thanks for having me. If you were wondering how you would know which games might work best on the deck versus those that might not, well, it seems Valve has a solution. This week, they announced the Steam Deck Verified program.

This seems like a pretty good idea, right? Because how were we as consumers, especially if you're not maybe as savvy as a news follower, like how would you really know what games work best on the deck or not? And it seems like this is probably going to solve that, right? And it strikes me that we have inspired with the deck, it seems, a lot of Windows users to try Linux. We've had some listeners that have been Windows users for a long time who are currently trying out the deck. You see the Linus Tech Tips folks taking out a Linux challenge right now.

That's probably the most prominent one. That's probably the one I should ask. Of course. And but like the Linus Tech Tips challenge, a lot of people are missing the forest through the trees. And this is something I've observed in the notes that we've received

and something I've observed watching the Linus Tech Tips Linux switch challenge. What Valve has done here with the deck is they've created a carefully crafted set of hardware and software to make sure that the end user gets a great experience. And they've built for Linux.

But a lot of these Switch Challengers, and I don't mean anything disparaging if you're a listener, but you've got to understand, you're cobbling together hardware that you've had forever, that you bought specifically for Windows, and you're getting frustrated when together hardware that you've had forever that you bought specifically for windows and you're getting frustrated when it doesn't work,

which is a bit akin to like being a windows user forever. And then going out and getting one of these new max and then being frustrated that the hardware you bought to work with your windows box doesn't work with your Mac. Like that'd be silly. Everybody would just take that as obvious on its face. Well, yeah, you bought something that only works with windows. Of course it doesn't work with your fancy new M1X Mac or whatever. So it's sort of like that same philosophy isn't being applied to the Linux switch.

Like Linux is just expected to work with everything, I guess. And I mean, that would be awesome, but it's just not happening. For most things, it does, right? For an incredible array of hardware. But unfortunately, hardware and drivers matter. And sometimes it just doesn't work. Yeah. And I mean, I'd love it to. But it's like, you're missing the point here. I don't know. I see people struggling with this. And I think this is a great idea. So the deck really does teach us to get this working. You

need a curated set of hardware. You need a set of drivers you know that work. And then putting a certified for the deck program on top of this, it also kind of gives developers like a goalpost now. Developers kind of have an idea of, well, what do we need to do to be deck compatible? Is it just Proton? Is there something more to it? There's actually something to accomplish now for game developers. And for consumers, I would assume there'll hopefully be a list that gets compiled either by Valve or someone, and consumers can check that list before they buy a deck, I would think. Yeah, hopefully this, you know, assuades any fears right now and makes both sides of the marketplace a little more interested in investing.

I'm curious, Brent, if you're in the market for a new machine. One of our emails later in the show is about an ideal Linux laptop, and Apple just announced these M1 Pro and M1 Max-based Macs, a Max Mac. Are you curious? Are you tempted to get a Max Mac, Brent? I saw all the notes in our show notes ahead of time because I have that insight. I got really excited. My antennae are tingling. So I will say Alex and I built this sort of Franken X250 of mine here in front of me. But it's just disappointing.

So I'm really looking forward to getting something else. And with all of the offerings we've seen come out in the last few months, I'm actually happy to say it's going to be a hard choice. Yeah, I was kind of kidding about buying a MacBook, but we may be nearing a future where you could actually buy a current MacBook from Apple and get it home and put Linux on it. We don't know yet, but it seems Asahi Linux feels like they have a good shot. They feel like they've got a chance of getting this working.

You know, that actually feels a little bit unbelievable to me because it feels like only a month or two ago, we were talking about, you know, the very beginnings of having that work be done. And how are we where we are now? It seems fabulous is the only word. I know. We don't know for sure if these new machines will work. No, we do not. Although it sounds like Hector Martin has ordered a new MacBook with one of these shiny new SoCs in it and is going to do a live bring up stream sometime soon. So yeah, fingers crossed, right? I mean and it's going to do a live bring-up stream sometime soon. So fingers crossed, right?

I mean, this is going to be a test of some of their design, some of their projections that, you know, the unique setup of this Apple Silicon might make this process going forward easier. And now we kind of just have to hope. Yeah, there are some unique ways that Apple set it up in such that a lot of what runs the internals of the M1 SoC is done in hardware. And so the OS doesn't have to be directly in control, but it has to have interfaces to that. And they think because some of those interfaces haven't changed since very, very early in Apple's SoC days,

they think there's a good chance that Apple won't change them for the next iteration just to keep things simpler for macOS development. And knock-on effect to that is it means a lot of the work they've done to bring Linux up on the M1, which is not all upstreamed, but a very significant amount of it is now upstream, and the rest is pending upstream. They think a lot of that work that they've done may just automatically apply to the new chips that Apple just announced. And like Wes said, Hector is so confident that he's going to live stream it. I think it's like next week or sometime.

You'll have to check his Twitter feed, but that'll be interesting. He's just going to do it live and he's going to troubleshoot and work his way through it. And a reason why we mentioned this is I think we have to acknowledge that from a technological standpoint, these are pretty impressive chips. What Apple is accomplishing here with performance to power ratio is really good to see, actually. And I hope it drives the rest of the industry. And getting Linux onto that, I think, is going to be pretty significant because not only does it open up new hardware to more people and that on its face, more choice, I think is always a good thing. And if you don't think choice is a good thing,

you're probably probably not very happy as a Linux user. And then secondarily, I think just from a repairability and a lifespan and also a data forensic and security standpoint, just nerdery too, right? I mean, like you put Linux on here and look at how much already we've learned more than we would have learned from Apple directly by Hector and the team's efforts, right? Like we know so much more about the internals. It's not that it doesn't feel weird, right, Brent? But it's compelling hardware and it does,

especially for a photographer, it does sort of address a specific market use. Yeah, and I remember a time not even so long ago when we were all quite disappointed in Apple's hardware and then they've just kind of leapfrogged everybody into this stuff. It's really impressive to see and exciting. I haven't been excited about Mac stuff for a long time. I agree.

In fact, on Coder Radio, I was quite critical of their Mac hardware over the last few years, especially the early butterfly keyboard systems that only had 16 gigs of RAM. the early butterfly keyboard systems that only had 16 gigs of RAM. And now that we're seeing systems that can support 10 cores and 64 gigs of RAM, now you're starting to speak my language. I start to take notice too when you're talking about doing all of that around 35 watts. I find that unbelievable. So I very much would like to see my favorite operating system on there. And I think the Asahi Linux team has probably got a pretty good shot.

I think if you were in Vegas and you were betting, I think the odds are on them. But there's something else going on this year. Along with all of this stuff about new architectures and ARM in the background, and let's be honest, in the space that actually makes money for Linux, the server market's in a bit of an upheaval, and we've all kind of learned the lesson

of how important a community enterprise Linux distribution is. So we're keeping an eye on what's going on in that space as the December deadline for traditional CentOS approaches. And this week, it looks like the AmoLinux folks have done the undoable. Nobody would take this challenge on well and stick with it, I should say. Jack, it sounds like AmoLinux is introducing a new migration tool. And if I follow, it means I could migrate, say, from older versions of RHEL, potentially even like version 7 to AlmaLinux or what?

Tell me how this works. Yeah. Hey, so from versions of CentOS to really any enterprise Linux derivative distribution. really any enterprise Linux derivative distribution. So that's Alma Linux, Oracle, even CentOS itself, and also Rocky Linux. That's amazing. You're telling me that I could just use this to get my butt off of RHEL 7 onto just CentOS Stream? Yes, if you wanted to. We're not billing it as a conversion away from rel.

Right. Right. Of course. Why would you, that'd be suicidal. Well, I mean, it's not,

it's not what we're after. And I think that's not the audience for it, but I think it's more anyone that's still running like CentOS seven. And we hope that eventually we'll be able to support CentOS six and maybe even an earlier version to just, you know, come up to a newer version and, you know, bring their stuff into modernity. Oh, my gosh. I can't. I don't think I can express to you the level of gratitude I would have felt back and moving older RHEL systems to newer RHEL systems.

Or what I used to end up having to do the hard way was hard migrate an existing RHEL box to CentOS. And it was almost always a couple of versions of RHEL behind. So, I mean, it was to say it was an all-nighter is an understatement. So a tool like this, boy, I just I'd be buying you guys all the beers right now. So congratulations, because not a lot of people want to take this on because it seems like something that is fraught with landmines. Listen, you still need to be careful. You still need to make sure that, you know, you're following like sysadmin best practices and make backups. And, you know, you may want to do like

a dry run in a sandbox or something, but, uh, yeah, I mean, it's def it's definitely doable if you're smart about it. And I think that as time goes on, you know, listen, we're, we're doing this as a community wide, uh, project. So I think that as more people get involved, as more people test it, and as more people contribute back, it'll only get better and only get easier over time. That's nice. And yeah, I have no doubt that anyone in our audience, especially anyone who's been using CentOS or RHEL that's getting pretty old, they're going to do a dry run. I have faith in them. Well, thanks for coming on and explaining this. I know it's a busy day over there for you, Jack. So it's good to get the scoop directly so that way I didn't miss up the details.

And it's a great tool to see. So thanks for joining us. Yeah, absolutely. And if anyone wants more info, check out amalinux.org slash elevate. And we look forward to working with everyone across the EL derivative community. the EL derivative community. So hopefully this is something that, you know, us, Oracle, Rocky, even Santos and Rell themselves can all contribute to. And we definitely look forward to that. I think that's something that's, you know, a little bit missing in the community today. So hopefully this is something we can all come together on. Linode.com slash unplugged. Go there to get $100 in 60-day credit on a new account,

and you go there to support the show. Listener support, and this is one of the ways you can support the show, is what lets us stay independent and stay true to our listeners. It's by your support, and you can do that by going to Linode.com slash unplugged, and you get $100 in 60-day credit. Linode's my favorite way to host something. I've been using them for over a couple of years now, and I absolutely wholeheartedly recommend them. Even if they weren't a sponsor, I'd still

tell you this is the place to go. Don't tell Linode that, though. You know, it makes sense, really. They're independent. That always checks a box with me. They've been around forever, 18 years, which is forever in internet time, and they have remained focused on their mission to make computing simple, affordable, and accessible to all in the cloud. And that's led to a best-in-class experience. Nobody checks all of the boxes Linode checks. Oh, yeah. There's lots of ways to host stuff. And some companies check some of the boxes.

And some of those boxes are pretty important boxes. But nobody checks all of them. And when you get $100 in credit, you can go see what I'm talking about, linode.com. For example, go deploy a NextCloud server, maybe a Minecraft server, go build up a development environment, or do what I do all the time in A-B test stuff, different distros or different

projects. I've mentioned this before to you, but one of my favorite activities on Linode is trying out several different permutations of a project and picking the one I like the most. I did this a bit ago when we were looking at infrastructure options and I wanted to try out various different WordPress configurations. But to get WordPress really fast, like, you know, rank really well in the Google search kind of fast, you got to kind of have to tweak it a little bit. And Linode has a one click that I ended up going

with that was better than anything else I built myself, which is all based on top of the light speed stack. You'd be really impressed by how fast you can actually get WordPress going on Linode. And it doesn't take like a crazy expensive box either. In fact, their prices are 30 to 50% cheaper than the major hypervisors out there that want to lock you into their system anyways. And they have an easy to use cloud dashboard that's so much more straightforward than anything else those guys have cooked up. I don't know what they're doing over there. It must be designed by like a committee of a hundred over there. And then Linode layers on top, great services you're going to eventually find are just invaluable.

Like S3 compatible object storage at fantastic prices, DDoS protection, VLAN support, a powerful DNS manager, and so much more that, you know, us nerds really appreciate. But don't worry if you're brand new,

you've never even spun up a server. They've got stuff that'll work for you, too, like those one click deployments and a really nice interface. And the best part, and even as pros use this sometime, the best customer support in the business. Man, I can't tell you how nice it is when your production stuff is having an issue to have somebody you can call that's willing to help. Linode is great and it could be part of your multi-cloud strategy, or it can be your cloud like it is for us. And they've been supporting open source initiatives that we love from the very beginning,

like Kubuntu or the All Things Open conference, which Linode is at this week, or my beloved LinuxFest Northwest, which I hope will return to in-person soon. Either way, Linode is dedicated to offering the best in virtualized cloud computing. If it runs on Linux, it runs on Linode. So go to linode.com slash unplugged, get that $100, and support the show.

linode.com slash unplugged. We do have some housekeeping, some things to take care of around here, and I want to say right up front, thank you, members. We had a couple people sign up last week, long-timers, and it's awesome to see that. If you've been thinking about supporting UnpluggedCore.com, go there, join us. You get two feed options, a limited ad feed, nice and tight, same full Joe touch, just a shorter runtime, or the big sucker, the live show, which as of right now is already an hour long. So it's a lot more show. And it's going to keep on going, too.

So you get that. There's tons of good content in there. In fact, the businessman in me says I should be monetizing that content or at least cutting it up and sharing it more often. I may every now and then share a special one. But part of it is the loose atmosphere. It's the version we just shouldn't be releasing. Yeah. It's like the stuff you say because you're not releasing it, right?

And then you release it. Whoopsie. Well, to our members, because we know that they get us, right? Like, they get us. They're supporting us. It's actually so cool. Yeah, so like... It's a lot of trust in this membership.

That's what it is. It's like we trust that they get us and that we can say stuff in front of them. And so that's sort of one of the benefits of being a member. Unpluggedcore.com to sign up for that. We'd love your feedback. We want to integrate more of it. In fact, it's something that I'm trying to stay on top of more and more. So please do.

Linuxunplugged.com slash contact for that. Our Matrix community is continuing to grow. I'm really getting excited more and more about Matrix. There's a lot of platforms that are great. I'm on Discord. I'm on Telegram. But I feel like long term, Matrix is solving a lot of problems and opening a lot of platforms that are great. I'm on Discord. I'm on Telegram. But I feel like long term, Matrix is solving a lot of problems and opening a lot of doors. Why not get in on that?

Linuxunplugged.com slash Matrix. Go there. Join. There's a LUP plug group. There's all kinds of, there's new spaces that Kernel's worked on that are beautiful. So you can, if you're just a Linux unplugged listener, you can just focus in on that. It's great. So that's over at linuxunplugged.com slash matrix. And then I mentioned that LUP plug. That happens on Sundays at noon Pacific in our Mumble Room. Details at linuxunplugged.com slash mumble so you can sign up. And then the nice thing is you get to hang out with like-minded Linux users. And then you got a Mumble client that's already good to go.

So like perhaps you're like Jose, who joined us in Denver, and has made it to the mumble room today. Jose, tell everybody where you're from. I'm from Puerto Rico. I'm a self-taught web developer. I have a freelance shop, and I won a Raspberry Pi in Denver. That's right, you did win the Raspberry Pi. And what time is it where you're at right now? Right now, it's almost 3.30 in the afternoon. It's basically Puerto Rico. It's basically Eastern time but we don't have daylight savings so that's the same time

the whole year. Jose's making it in the middle of his work day. I'm just saying be like Jose. You know, just take a long lunch and come join us on a Tuesday sometime. You're not feeling great? Maybe something went wrong? Come hang out with your Linux friends. Go get our mumble set up. Linuxunplugged.com slash mumble.

Be like Jose. Be a longtime listener. And then eventually one of these days you can just pop, you get an afternoon, you can pop in. It's great. And you hang out and chat with us directly. And you also, kind of a nice benefit, get the lowest latency, highest quality audio live stream. Straight from the source. With all of our extra mistakes.

Hey, well, how about some feedback, huh? Yeah. Eric writes in and says Garuda saved his butt. I'm a big fan of Arch on my personal laptop, so I've been testing out some of your recommendations like Endeavor OS and Garuda. Well, I'm at my first

in-person conference this week in over two years and I updated my laptop before leaving. Didn't think much of it until I got to my hotel and tried to use my laptop. Uh-oh. While updating my system really meant I ran Pac-Man, that's why you and didn't check any outputs. Oh, man. When I got to my hotel, it wouldn't boot because there were no kernels to be found. Oh, no.

Luckily, I had my handy-dandy Ventoy USB stick and started trying to set up my Chirrut environment when I realized Garuda had been taking ButterFS snapshots. Heck, yeah. I proceeded to reboot my laptop, selected a previous snapshot, logged in, restored it, rebooted again, and I was up and running as if nothing had happened.

Man, isn't that what we've been talking about right here on the show? Why we needed to up our file system game a little bit just to make these things more reliable workstations? And yeah, he's rolling the dice with an update right before he leaves. It sounds like one of my moves, but it was pretty nice that he can roll back, huh? Yeah, no kidding.

Now, there is one thing Eric did note that he was very grateful that there's just this tight integration between Pac-Man and TimeShift and Grub and all of that set up for you in Garuda. Yeah. He did just have one complaint, though, and I thought maybe you could inform him here, Chris. This is also the first distro that I'm sticking with that uses KDE as the desktop environment. And boy, is it sluggish on my laptop with an Intel N5000. Do you have any plasma performance tips I can try?

Hmm. Hmm. Hmm. So that's one of those silver edition Intel processors that has like a 1 gigahertz or 1.1 gigahertz base frequency. It's pretty rough. It bursts up to 2.7, but depending on your cooling, it won't necessarily hold that for very long. It is low power, though. That is nice. That's rough. You can go into

the, you may have seen this already, but you can go into the compositor settings and you can try different versions of OpenGL. You can also, there's a, there's an option there to speed up the animations or slow down the animations. And I would recommend you play with that. And I'm trying to think, is anybody in the mumble room have any plasma performance tips, any,

any tips or ideas for making plasma run just a little bit quicker? You know, I, I have turned off, you know, you mentioned animations and stuff. I go through and turn all that stuff off. Like Wimpy was saying last week, sometimes it's nice to just have it be a little bit snappier. I think maybe my laptop's a bit more performant, but I still, it just makes a big difference, I find, on KDE. And so I'd give that a shot, go through all the settings and turn them low or off is even better if you can and if you don't care about that stuff. Yeah. And you would know you're struggling with low-end hardware at the moment. I know. I'm making it though. It's going okay.

Counterintuitively, I would make sure that you are not falling back to CPU rendering compositing. Yeah. rendering compositing. If you're rendering on OpenGL, you'll actually get better performance than you would if you're on, say, X render or no compositing at all. Because when you have no compositing or you're using X to do compositing, you will force the CPU to do all of the work instead of offloading it to the GPU. And I'll take it one step further, Neil. Depending on how good your GPU is, and in the N5000, this might not apply. But if you have a more modern Intel GPU or any of the larger commercial GPUs, having some animation actually kind of papers

over the loading time of the window in a way. Because if you just turn off all animations, you will get the absolute fastest. The machine can draw that on the screen, absolutely. But that's not always a satisfying speed to tell you the truth and with the with the animation effect like a little fade in or whatever it is kind of papers over your perception of that a little bit because the window is actually still rendering and drawing as it's fading in and that just makes it a smoother

process you don't always get a performance improvement and i actually find i become more sensitive of how my system's performing when I turn off all the animations because I can clearly see it's struggling. But it really depends on your combination of CPU and GPU. And it's just something you have to go through and kind of mess around with. But if you have any tips, linuxunplugged.com slash contact, maybe we can compile them together. Brian wrote in with a neat new tip about functionality in the kernel. He says, you mentioned, he's talking to me, that you were trying out a battery charge threshold on your laptop using TLP. And then I would forget to use TLP full charge almost all the time. This bit me

on the road trip and I'd go to use my laptop and it would have at best like a 65% charge. Right. It's no good. But Brian points out that this functionality was recently exposed by the kernel. So you don't actually need TLP to do this anymore. And then you can trigger the charge or charge threshold with systemd. And you can start and you could even include a unit file as an example. You can start and stop a service that sets the thresholds or it sets it to full charge. And he says, I even automate it with a hotkey.

That's a pretty neat idea. And I didn't realize that battery charge threshold settings had been exposed by the kernel and that systemd knew what to do with that. That's awesome. Looks like it's nothing too crazy in the systemd setup, just using bash to echo some values under slash sys. But hey, that's the power of Linux, right?

Oh, I see. So he's using systemd to orchestrate just like when to kick off a script or something like that. Yeah, when it starts, it'll kind of set that, and then when it stops, it'll set things back. Ah, back in my day, we used cron. Yeah, that's one option too.

Cron is bad. You want to take Danny's? Well, Danny might find that last tip handy because he's shopping for a new laptop. I'm a longtime Mac user who's dabbled with Linux since 2006. I need a new laptop and was wanting to run Linux Mint Cinnamon or Manjaro KDE, but would also need to run some older Mac apps like Adobe CS5.

Should I make sure I buy Intel or is AMD okay if I use something like Quick EMU? on some older Mac apps like Adobe CS5. Should I make sure I buy Intel? Or is AMD okay if I use something like Quick EMU? I'd love to buy a MacBook Air M1 if I can afford it, but does Quick EMU work on it as well? Thanks for any help, Danny. Hmm. At first when he said I want either Linux Mint Cinnamon or Manjaro KDE, I was like, what?

That is two opposite sides. But then I realized actually kind of what I like about Cinnamon is it's kind of like some of the best stuff about KDE and a GTK desktop. So then it started to click. I still think it's kind of strange, though. It's a pretty big dichotomy between Mint and Manjaro. So he may want to revisit exactly what he wants out of a distribution and consider Fedora.

Hey, sometimes you like what you like. You know, if you get a CPU that supports the right virtualization extensions, which is pretty much most all of them now for desktops and laptops, it should work either way, AMD or Intel. I wonder, though, if you have considered if you want Adobe CS5, you didn't say what other older Mac apps. So this may not apply, but Adobe CS5 is also available for Windows. And Wine is probably going to be a lot better time than a whole VM, right? Because Wine means not only are you running it natively on your system, but you're also going to have access to your local native file system. And so that's a lot simpler to work with, and it's going to work a lot more than you'd expected on the Mac. But you absolutely can go QuickMU on the M1 platform.

It's not completely done yet, but check out UTM. It's a virtual machine manager for Mac that sits on top of QuickMU and we have our QMU and we have a link in the show notes and they tout Linux support in there. It looks basic, but covering most of what you'd want from something like this, and it absolutely can run macOS. That's maybe one of the advantages, but you have to be on the M1 system. So perhaps it's a future scenario where you get Asahi Linux or some Asahi Linux-enabled distribution on there, and then you could install UTM and have a go. Or use UTM and macOS and run Linux in that and run Windows in that.

I would strongly also look at the System76 Lemur Pro. I'd also consider the Framework laptop. Oh, yeah, Framework would be good. Their marketplace just opened. It seemed like that's rolling out nicely. Yeah. XPS 13 perennial option, check those out. What was that laptop that you tried out,

that gaming laptop that I think what Dylan has now? Yeah, yeah, yeah. The those out. But then what was that laptop that you tried out, that gaming laptop, I think what Dylan has now? Yeah, yeah, yeah. The Asus. Yeah, the Asus, the G14. It's like one rev behind, but it's got a decent like RTX 2060 or 3060 in there and it has 120 hertz screen. And as long as you're using something that's got like kernel 5.12 or 5.13, everything works out of the box. Dylan's been really happy with it. So Manjaro KDE might work? Yeah, yeah, that totally would. I definitely want to double check Mint because I don't think 2004-based distros work out of the box on the G14. But anything a little newer, like a newer version of Pop or 2110 Ubuntu, that would totally work.

That's totally doable. We got several recommendations for different YouTube archival tools. That was pretty nice. Wufefe sent in his. It's called YTCC, and it's a command line tool to keep track of your favorite playlists on YouTube and then archive them to custom folders

based on your settings. We'll have a link to that in the show notes. He says, it's command line only, but you set up the playlist you want to track, and then you run it as a cron job, or maybe a unit in system D, and it'll go fetch new videos just periodically for you and store them locally.

Yeah, it can be YouTube channels, SoundCloud, RSS feeds. And I love that this is kind of just a standalone tool. You know, maybe you want a whole system, but if you want to roll something yourself, you've kind of gotten to automation. You want to do some of that yourself. YTCC might just be a nice piece of that. Very much.

And I was thinking about this is, I wish we had a way in Matrix to have like, and maybe there is, but not a wiki per se, but wouldn't it be amazing if we had a simple way to submit things like this

that the audience sends us and keep a list in Matrix somewhere and maybe the other picks that we do. Just like a picks room we make somewhere? Yeah, it needs, the key would be somebody would have to be able

to maintain it because these lists always, always end up maintained. Ever. Every single list that any podcast ever starts, it's not just us, eventually it gets unmaintained.

And I'd love to find a solution to that. Probably something with the blockchain. Brent, do you want to take our last email from Marcel? Marcel has created a neat tool as well. He writes, I wanted to share this system status message of the day generator I made with a friend. And of course, it's written in Rust. the day generator I made with a friend. And of course it's written in rust.

I don't know if you click through the link that he sent us, Brent, but it's beautiful. You log into your box and his message of the day generator gives you a beautifully formatted weather report. It shows you what your uptime is. If your SSL cert is active, your file system usage, the last couple of logins. And if you have failed a ban, it'll give you like a super brief summary of failed a ban. And the last time that was checked. And it, man, it's like one of these things that I just love because it's a simple Rust app with zero runtime dependencies, and it just looks great. So we

obviously have to put this on the new server, Wes. Oh, yeah, I think so. I mean, it looks so shiny. A nice reminder every time you log in. And Rust, I kind of love in the readme here, the author says, I got stuck in dependency hell one too many times trying to update interpreted alternatives, e.g. something like Python, probably. And that's how it ended up in Rust. That's great. You know, I kind of got back into having something in my message of the day, which shows up when you first log into the console or when you first open up a terminal.

And I was using Garuda Linux and they auto-start NeoFetch and give you the NeoFetch output, which is great, but takes a beat every now and then to do its analysis and run. And of course, you know, I'm a speed hound. So what I like about what Marcel's built here is it's so simple and it's so fast.

It really, by the time your terminal's done, it's loaded and it gives you that output. It's really choice. And it's in the AUR. I want to take just a bit here and mention that our friends over at CloudGuru have a certified entry-level Python programmer certification course. Python powers so much of what we often talk about on this show, and getting a solid entry-level Python understanding is invaluable. And why not get certified in the process if you want? A CloudGuru has the certified entry-level Python programmer certification, a great way to learn about Python data types, control flow structures, data collections, functions and generators, and a lot more.

By the time you complete the course, you should feel more than comfortable taking and passing the Certified Entry-Level Python Programmer Certification exam. And more importantly, you'll have a good understanding of the fundamentals of Python programming. We'll have a link in the sponsored section of the show notes, or you can go to cloudguru.com and search for Certified Entry-Level Python Programmer Certification at cloudguru.com. Well, Chris, I thought you might need just one more rust pick today. Somehow these have been sneaking in all episode. Not that I'm complaining, but this one also involves WireGuard. WireGuard. Enter OneTone, a cross-platform user space WireGuard port forwarder that requires

no system network configurations. What? How can this be a thing, Wes? How can you set up a WireGuard VPN with no network configuration? Is it just like a user space Go app that... I still don't get it. It's still creating a network connection. I don't understand.

The use case here is, let's say you have an existing wire guard endpoint something like a you know a router running wire guard and you have a peer on the wire guard network that's running a tcp server if you want to be able to access that tcp service from a second computer that's not part of your wire guard setup and for some reason you can't install wire guard on maybe you don't have access, maybe it's some server that you really don't want to install things on, or, you know, it's just not blessed in some way. Yeah, sure. That's where OneTon enters the picture. And it's using a bunch of neat libraries in Rust that basically implements its own

user-based TCP stack to make this magic happen. Wow. Wow, that must have been a hell of a project. That's impressive. This reminds me a lot of that tool we talked about years ago, S-Shuttle, that made sort of a quasi-VPN connection through SSH without needing any super fancy setup. But this sounds so much more elegant. Yeah, I haven't had a chance to give it a shot yet,

but I definitely will. So small TCP is the TCP stack they're using. And then it's also using Borington, which is that Cloudflare Rust implementation of WireGuard. So much Rust. It's ridiculous. I can't even anymore. I feel like this is the prodification. I'm trying to think of the right term. This is what we expected to start

happening around WireGuard once it got mainlined and we started to see WireGuard filter out into the general population. Something else that comes to mind that the audience is super excited about right now is TailScale and that's another product that truly is like a product with a service

and a good user experience that sits on top of WireGuard. Now we have this one ton, which lets you participate in a WireGuard network without even having administrative access to your box. I could see how that'd be super useful in a corporate environment. This is really neat.

And there's other products out there built on WireGuard. And it's sort of like, it's like the sort of Valve certified stamp for the deck. This is like the, you can trust the crypto stamp for a VPN product. Assuming they're not using a bastardized implementation of WireGuard. It means I can kind of trust the fundamentals of the software is running on top of. And I like the idea of of that specifically, because then I don't have to worry about if tail scales done some weird crypto that they can crack or they can read or is going

to get broken by some script kitty, because I know that they're using something that's tried and true and tested. And if it gets updated, that update goes out everywhere. And that makes these products, I think, to me, just a little more trustworthy. I love to see this. It's also kind of the, you know, one of the reasons we love open source is like we can have this shared base that we all get to use. And that just levels the playing field and raises things up so that, yeah, a whole bunch of other parties can build neat stuff on top without having to reinvent the wheel each time. Yeah, I mean, and for me, you know, on Linux, like, that's great. I prefer to go the route where it's built into the kernel.

I prefer to go the route where I could use it with my other tools and all that kind of stuff. But I totally appreciate the use case for that. use it with my other tools and all that kind of stuff. But I totally appreciate the use case for that. But I'd say with all of these, just double check that, you know, they're not using some crazy esoteric wire guard or some crazy esoteric port or something like that. You know, make sure it's good industry standard stuff. And that's a big part of the due diligence now, I think. So there you have it.

And we'd love your picks. Let us know linuxunplugged.com slash contact or join our Telegram group, which is going all the time. That's at jupiterbroadcasting.com slash Telegram. You'll see several of the hosts in there throughout the week and drop links and interact with folks. But if you absolutely want to get in the show, that contact page, that's really the way to do it.

And if you do the Twitter thing, you can also follow the show at Linux unplugged. The network is at Jupiter signal where you can get notifications about the entire networks. Yes, there are other podcasts on the network and they release throughout the week. This is a self-hosted week too, I think. So at the end of the week, towards the end of the week,

there's going to be a Coda Radio first and then there'll be a new self-hosted. And if you follow us on Twitter, you find out. So much to look forward to. Yeah. There's also one of those fancy combined show feeds, right? Where do people get that? That's a good idea.

Go get the all shows feed. In fact, if you go to jupiterbroadcasting.com, there's a shows menu there with all the feeds for all the shows listed, including the all shows feed. So that's really the pro tip. All a card or all in one, it's up to you. That's right. That's right.

And you're always welcome to join us live. We do the show on Tuesdays at noon Pacific, 3 p.m. Eastern at jblive.tv. We'd love to have you. See you next week. Same bad time, same bad station. Indeed. Go get JBLive.tv. We'd love to have you. See you next week. Same bad time, same bad station. Indeed. Go get yourself some Linux action news. Go find out what's going on every single week in the world of Linux and open source LinuxActionNews.com. Links to what we talked about today

at LinuxUnplugged.com slash 428. That's where you get all the infos. You want that Raspberry Pi image of Pop! OS? You get it over there. That Alma Linux transition tool? That's where you get all the infos. You want that Raspberry Pi image of Pop!OS? You get it over there. That Alma Linux transition tool? That's linked at linuxunplugged.com slash 428. While you're there, hit that contact page, get the subscribe button, all that stuff. Thanks so much for joining us on this week's episode of the Unplugged program,

and we'll see you right back here next Tuesday. Tuesday! Tuesday! Tuesday! right back here next Tuesday. Tuesdayitles.com. Go pick our show title. Brent, it's so nice to have you back. You've been traveling.

You've had incidents that have happened. I mean, it's been a bit. I am home for the first time in like four months. Like your real actual home? Yeah, yeah, yeah, yeah. My cats are here and I'm here. It's weird. It doesn't really feel like home anymore, I got to say.

I bet. It's the strangest feeling. But I got to say, it feels good to be back. And I also want to say one of the crazy incidents that I ran into was I got into quite a scary car crash a few weeks after our Denver meetup. And I was near Denver when that happened. And I just wanted to say huge thanks to the whole JB community. Uh, I got so wonderfully overwhelmed with, uh, messages from people in different ways, just saying, Hey man, hope you're doing all right. And so got tons of love, lots from the, uh, self-hosted discord, um, some from

the telegram group and people just messaging me just to say, Hey, hope you're okay. And I really appreciate that. It's exactly what I needed at that time. So thank you. Thank you everyone. Super, super appreciate it. I'm totally fine by the way, miraculously. So I'll be, I'll be here more regularly. That is really good. And yeah, when we saw the pictures, it was one of those, is one of those accidents you look at and you're like, how did he walk away from that? There's not much of a car left there. And for anyone who's been in a car crash, this is my first, which I would not recommend it by any means. And it just is a scary thing. And it's one of those

crashes for me where it just replayed in my head for about two weeks. So it's a pretty bad one. I bet. Do you, uh, even now a few weeks later? Yeah, I definitely do. I'm, I'm physically totally fine, but I gotta say like super traumatic. And Chris, you and I were talking about how exactly the accident I had was one of your worst nightmares, which is, you know, traveling down the interstate and something, just another car just appears in front of you. There's not a single thing you can do except crash directly into them going at highway speeds. And, uh, you know, I was driving a Subaru at the time and I'm pretty sure that thing kept

me alive. And so thank you to the technology and those things, you know, thank you to the engineers at Subaru. They kept Brent alive.