LINUX Unplugged - 303: Stateless and Dateless

Episode Date: May 29, 2019

We visit Intel to figure out what Clear Linux is all about and explain a few tricks that make it unique. Plus Wes and Ell are back from KubeCon in Barcelona and return with some great news for open source. Special Guests: Alex Kretzschmar and Brent Gervais.

Although maybe before we start, we should probably start with a very special happy birthday. Happy birthday, Brent. Happy birthday, dude. Oh, you guys. Thank you. Still out there traveling the world in part. It's the new normal. I assume he wouldn't want it any other way. Well, I'm glad you're here, Brent, even though it's your birthday,

because we went down to Intel's Portland campus and learned all about Clear Linux and got some really good takeaways. So we'll be talking about that in today's show. But that's for later. Just right now, I want to say happy birthday, sir. You know what? I kind of didn't even realize it was coming up because I don't even know which day of the week it is anymore. Time compression.

It's a real problem. I don't know how Elle deals with it, to be honest, because how do you deal with it? Do you not suffer from time compression like I do? I think I just I never know what time it is. Yeah, I think we're all there's some of us out there

that when you travel a lot you're just, you're dateless. I, the only thing Floating through life. The only thing that keeps me pinned is the show recordings. Like I, if I don't do a show I don't really know

what day of the week it is. That's why when you change which day you do a show it really, I mean your whole week changes. It really messes with you. It's a weird phenomenon. You're running stateless is what you're saying.

Stateless and dateless. There it is. Ouch. This is Linux Unplugged, episode 303 from May 28th, 2019. Well, hello there, and welcome into your weekly Linux talk show. My name is Chris. My name is Wes.

Hello, Mr. Payne. How are you feeling? Jet-lagged at all today? I think finally I'm back to normal. You actually look good. Oh, thank you. You do look good. I'm ready to do Linux Unplugged again. It was a pain to be away last week, but you guys did a great job. Special thanks to my pal, Brent. Somebody has to go to Spain.

Somebody. I mean, yeah, it's... Barcelona is somebody's job. We didn't want to have to go alone, right? I mean, just it's not as much fun that way. Yeah, so Wes and Elle are back from Barcelona this week. They went to KubeCon, and they grabbed a few really good insights

from the entire event. We'll share those with you. And Brent and I are back from our trip to Clear Linux in Portland, Intel's campus. We'll tell you all about that. There's a ton of stuff to get into. But before we can do any of the community news and our own legit form of journalism, I want to say time-appropriate greetings, virtual lug. Hello, mumbles.

Hello. Hello. Hello. I want to say time-appropriate greetings, Virtual Lug. Hello, Mumbles. Hello. Hello. Hello.

And I do believe, rumor has it, we also have Alex Allen. She's here, too. Hey, guys. Hey. Hey, what's up? What up? Oh, you know, doing a podcast. Doing a podcast. How you doing, Alex?

How are you doing? Oh, it's 98 here today. I am melting. It's too hot. It's too hot. Well, geez, I hope your equipment doesn't fall apart. Yikes. I'm like on top of the world right now.

I pulled off a thing recently, and it feels good. You know, something I've been toying with doing for a really long time. It's kind of been a dream. There's a nice thing happening right now. Now that all of us are working full-time for Linux Academy, we're kind of like infecting each other with Linux enthusiasm. You wouldn't believe what we did to the studio today, but that's a story for another time. It's like being a Windows switcher all over again.

It's so exciting. And so we keep, like, everybody is experimenting with different stuff right now. And I decided to tackle hardware pass-through in a VM on my ThinkPad. Because those that listen to the show for a while now may recall that a bit back, I bought one of those GPU docks. It's a Thunderbolt 3 Lenovo dock with an NVIDIA 1050 built into it. So you get a discrete graphics card and you get all the ports and all that kind of stuff.

And I got it to work. It's actually, it's nice. But it has its drawbacks. Mostly that when you don't have it connected, the NVIDIA driver goes crazy looking for the hardware. It's pretty comical. You've got this separate graphics card that mysteriously connects and disconnects,

and not many other systems have that limitation. No. So I thought, well, wouldn't it be cool if I just said, ah, bag it, and let's just pass the entire dock through to a VM. Make it simpler on the host system, and now you have two computers in one. See, at first I thought I'd pass my GPU through to KVM Make it simpler on the host system, and now you have two computers in one. See, at first I thought I'd pass my GPU through to KVM

and I'd call it good, but now going through this, I've learned that all of the devices that are in that PCI group have to be allocated to the VM. So if you've got, say, in this case, an NVIDIA card that also has an HDMI port with a sound chip on it, those are both in the same PCI group, and both have to be removed from the host system and allocated to the VM. All or none.

I didn't realize that. And it gets even more complicated with a Thunderbolt 3 dock because you've got the Thunderbolt 3 bus, you've got the USB adapters, plus you've got low-energy mode bus. There's multiple things. So I really did this deep dive on my Fedora 30 install of my ThinkPad of going through

and blacklisting all of these PCI devices and making sure that the Nauvoo driver wouldn't load and all of this. But at the end of it, I am happy to say I successfully managed to allocate the entire Thunderbolt 3 dock to a Windows 10 VM running on my ThinkPad under Vert Manager.

And just works? I mean, Windows sees the GPU? Yes. Now, there's one big drawback, and I haven't taken the final plunge yet. In order to get it working, you have to uninstall the virtual machine graphics card. You can't have both the VM graphics and the host graphics.

So you've got a little graphics card dance to perform. And I know how much you love administering Windows. Yeah, so I was getting everything set up. It's funny, though, because now the Windows machine is connecting over the Ethernet port on the dock. And if I don't have the Ethernet plugged in, my Windows VM doesn't have networking. It really is too machine.

It really is. So I'm going to hook up a mouse and keyboard and a monitor to the dock, and I will have a separate machine there. Not really any plans to use it for anything other than I just wanted to see if I could pull it off. But so far, it's been really cool. So now I just got to take the plunge, remove that virtual graphics card, hook up a monitor, and see if it works.

If you're on Fedora and you want to mess around with something like this, Red Hat themselves actually have decent-ish documentation. The problem with all this stuff is it gets out of date really quick, especially if it's documentation for RHEL. So just make sure you double check everything. Sometimes like the path for where the grub config is changed and stuff like that. PCI pass-through was what got me interested in Linux in the first place. I wanted to put one graphics card into my server so I didn't have to buy another computer

because I was a poor student at the time and now it's so easy using ovmf which support uef uefi graphics cards that in the old days you used to have to eject the gpu before shutting down windows because the bios on the graphics card didn't know what was happening whereas now they support resetting and it's just dead easy now um but if you're interested in this kind of thing i highly recommend there's a arch linux wiki article called pci pass-through via ovmf and that takes you through everything you need to know about this topic um and more and if you want to reach out to me as well i i still run this every day i have a windows vm that i use for Lightroom and

other Adobe stuff. But there's a really cool project called Looking Glass, which I think you might want to investigate, Chris. I actually have it bookmarked right now. It looks really promising, although it looks like it's also early days and it's a little finicky to install, possibly. So can you explain what it is and why I might want to check it out? Yeah, I'll do my best because it's not simple so um when you are outputting graphics through a graphics card you are sending frames to a frame buffer which then go out through the hdmi port so what this guy i think his name is jinnif or gillif or something is his username he worked with wendell over on level one text to do some um really funky stuff actually you install a driver which hooks into that frame buffer then on the linux guest you run

a client which can view that frame buffer which is a kind of shared piece of memory between the vm and the host and then you can basically like a vC window, just open up the client on your host machine and view it, view windows in a window, if that's not too much of Inception for you, with maybe two or three milliseconds of latency. It is incredible. And recently there was a new plugin announced called Scream or IV Scream or something which does audio through that same method as well. Oh, no kidding. So in the Windows VM side, it's using DirectX to do all that frame buffer stuff.

So whatever DirectX can render, as fast as it can render it, it can toss it in that frame buffer. So it is very high performance. It's just also very early days and a little finicky. But it seemed like a great way to get very high performance Windows without having to have a dedicated monitor, which is what they were going for.

It's a really interesting kind of journey I ended up on because I think long-term it might be how I try to solve this issue for my son if he ever gets in a position where he really wants Windows 10. And I want to say thank you because I got a couple of audience members who submitted me part systems from PC Part Picker. That's great. Just a nice little build you can put together. A couple of really nice ones submitted me part systems from PC Part Picker. That's great. Just a nice little build you can put together. A couple of really nice ones, including a nice little small one that uses a case that Linus Tech Tips just featured.

And I'm really looking forward to doing this Ryzen build. But now they just got the new CPUs. We'll see. We'll see. I might have to wait. So on that note, there is something to be aware of with Intel versus AMD in this space. So Intel requires for this to work the extension VT-D and VT-X. AMD have their own and the name escapes me right now, but the important thing is that you can pass through IOMMU groups, right? And that's the

important thing you want to look for in any motherboard. And it can vary down to the exact BIOS version on that motherboard. So sometimes can vary down to the exact BIOS version on that motherboard. So sometimes the only way to be sure it's actually going to work is to buy the product and try it out. So buy it from Amazon because you can send it back if it doesn't work. Yeah. So those I M M O U groups, right? If I'm getting that right is like what I was talking about where you have a couple of PCI devices that are grouped together, like an NVIDIA graphics card, GPU, as well as the HDMI audio output. And anytime you have devices that are essentially like that in the same group, they all have to be allocated at once to the virtual machine. It's all or nothing, which is tricky, but doable. There is a kernel patch you

can apply called ACS, which can sometimes break down those IOMMU groups into smaller chunks. So occasionally, I mean, into my Arch install right now in front of me, I have that kernel patch installed, which enables me to only pass through a couple of devices instead of 10 or 15. So I can have two GPUs in the same system, pass one through and leave the other one with the host, for example. So there are ways around these things, but it does get quite involved quite fast. Yeah, but nothing more than maybe an evening or two. That's about how I would sum it up.

Nowadays, you're absolutely right, yeah. That's nice. I mean, not super simple, but something someone looking for a project could easily handle. And then once you have it done, you essentially have two computers out of one. And what I'm really trying to get out of this, I was telling you this yesterday, is

I just, I kind of feel like I'm done repartitioning my disk for a while. I've got a really solid base install on that Fedora 30 on that ThinkPad. I might, I mean, I may have to repave just to be compatible with the studio machines, but right now, Fedora 30 is working so good. And, you know, if I can

avoid having to reload that system by using things like virtualization with hardware pass-through and kegsec, hey-o, I might. I might try to make it like the entire release of Fedora 30 without repartitioning. You're in love with this installation. I can tell, right? You don't want to make any changes.

You're trying to figure out anything you can do to make stuff work on Fedora, which has actually proven pretty possible. Yeah, it's been really good. All right. Oh, there's our warning that we've waded off into the weeds. Let's get into some community news. What do you say? Start with, speaking of Ubuntu,

Ubuntu 19.10 is going to have a new trick. It's going to bundle NVIDIA's proprietary driver right into the ISO. The NVIDIA binary drivers won't be activated by default, but they will be present on the install media to make it easier to turn on afterwards. Yeah, that's right.

NoVu will still be the default if you've got an NVIDIA graphics card on new Ubuntu installations. Really, the change is just positioning the mainline and legacy proprietary drivers onto the actual ISO so it can be easily obtained locally post-install, right? So maybe you have a system set up,

you want to use the proprietary drivers, you might not have internet after or during installation. You don't need it now for that. I guess the not having internet thing is a business customer request. I guess I got the word that that trickled down from requests on large installations in enterprises. I mean, it is nice when you can sort of do all these things

and not have to futz with gateways. Do you care that it made the ISO bigger? Not really. I usually use something like Debootstrap anyway to build my systems from a more minimal ISO anyway. A lot of people seem to be super excited about the new ZFS on Linux 0.8 release that came out in the last week. I guess there's some pretty nice features that people have been waiting for for quite a long time.

Oh, yeah. Things like native encryption support, as well as raw encrypted ZFS send and receive support, which, man, send and receive stuff is just great. which, man, send and receive stuff is just great. Another prominent feature is support for device removal and pool checkpoints as well as trim or discard for solid state drives. All of those things people have been waiting for to be shipped as part of ZFS. Now, in the past, you could layer things together, right?

You could use Lux and ZFS or some other combination. This just means ZFS has more features and since it already does, I don't know, everything, you know, violates all the layers, eats everything related to the file system and disks. It does, doesn't it? Now it does more of those things. If you trust ZFS,

you're probably going to like these changes. Yeah, I mean, I think the encryption support's huge, right? And so is the trim support. That's for solid-state drives. That's massive. And hey, okay, Python 3 compatibility with its tools, the conversion marches on.

Okay, yeah, you're right, you're right. Have you seen what, there's a website now that tracks how many days, hours, and months left until Python 2.7, or whatever. Good for them. Yeah, right? I mean, steady improvements. Are you running this anywhere?

I'm not updated to this release, but I do use EFS on Linux. Yeah, of course. Do you? Who do you think I am? Where? I mean, on my arrays at home. Now, I do still have one that needs to get converted that's on ButterFS, but that's just kind of been something to watch and to play with.

How many arrays you got? You got ButterFS arrays, you got ZFS arrays. Yeah, right? Now, there aren't any Bcache ones right now. Or Stratus, so I'm not that hipster. I got to work on that. Yeah, man,. Or Stratus, so I'm not that hipster. I got to work on that. Yeah, man, get your Stratus set up.

What the heck? Alex, you got any Stratus rolling at home in your home lab? With the new fancy rack I saw. Congratulations. Thank you. Yeah, I had some offcuts of wood left around, so I thought we'd get it off the floor, away from the spiders that jump.

Built himself a server rack. Nice. Very nice. So you got any ZFS or ButterFS or Stratus rolling any of that stuff? So after listening to a couple of episodes of TechSnap where Jim and Wes evangelized it and going to the talks at LinuxFest Northwest from Jim and the one and only Alan Jude,

they convinced me that I was missing out by not having some ZFS in my life. And yeah, I've given it a go. And I'm presenting a ZFS pair now. And then on top of it, I'm using merger FS to kind of blend it in with my normal. There you go. Very neat.

Normal storage stuff. So anything I really care about is now on ZFS. Anything I don't care about is just on the JBOD merger FS kind of the XT4 drives. So yeah, I'm using a ZFS. Oh no, I said Z. I'm turning into an American. Yeah. So I'm using it for a bit and I must be honest, uh, the learning curve I felt like was quite steep because, um, I feel like if you make an array, if you make a pair, like a pool wrong on day one, that can have some serious ramifications

in a year or two's time when you're trying to recover data. So I wanted to make sure I was right. So I ran everything past Alan and Jim, and they said it was okay. So, so far, so good. Welcome to BSD Unplugged, your weekly BSD talk show that's too busy getting actual work done to care about your silly display server.

My name is Alan. Alan Jude. You know, Chris, you don't want to touch your install. I'm thinking it's time to pave my ThinkPad, and you know what? It's probably going to have ZFS on route on this next generation. I think so. With Distro? Sticking with Ubuntu?

Probably, but that's most likely. I haven't convinced you to try Fedora 30, huh? Well, I think I'll have set it up. There may be some room for a Fedora installer too on there. Or at least a K exec. Well, maybe the biggest hoopla news this week in the desktop Linux community has been this big post by a group of GNOME application developers. Stop theming my.app.

Please don't theme our apps. An open letter from independent app developers to the wider Gnome community. This is all about users that are changing the themes on their Gnome desktop and, in the opinion of these developers, breaking their apps because they're now inconsistent with how the application developer had designed them. Isn't this fascinating? Right on the heels of our conversation with Cassidy last week about this. By the way, congratulations to Cassidy, who is now a new dad. That's wonderful.

Like he was just on the show last week right before that big news. That's some commitment. And was talking about trying to create a free desktop standard around dark theming. Because that's a big part of why people are switching their theme is to go dark. They want the dark theme, yeah, of course. Not everybody, but a big part of it. I'll raise my hand. That's right, me too.

They say, all of our efforts designing, developing, and testing our apps are made futile by theming in many cases. And they do have a couple of examples. Yeah, they do. Things like GTK style sheets, well, they can make applications look broken and sometimes, in corner cases, unusable. Icon themes can change icon metaphors, leading to interfaces with icons that don't express what the developer actually intended. And app icons, well, they're the identity of an application. So changing an app's icon denies the developer

the possibility to control their brand. I really actually kind of agree with that last point the most. I think that is true. And I think at the user individual level the offense is mild, but it's a much bigger deal when you have

a large distribution that does it, because by default out of the box, it's changed for the users. You, but it's a much bigger deal when you have a large distribution that does it, because by default, out of the box, it's changed for the users. You know, it's interesting. I'm not sure at some point, like, there was a section of open source, maybe just free, I don't know, open source software that, I don't know, in the past, we would say had a brand, right? I mean,

if it's some code in a repository, but now things are thought of as apps, right? We're using those terms, and they have brands. Everything has a brand. And of course, I mean, right, there's trademarked art and logos, and there is a brand, but we're taking that more seriously, and maybe there's older schools of thoughts about how open-source software is shipped.

It does seem to be a confliction of those two things, doesn't it? You're right. We now live in a world where people want Thunderbird and GIMP to look like Thunderbird and GIMP. And developers want to be able to ship an application expecting a style sheet to look a certain way.

And then you have distributions that sort of make it a point of pride to customize the distribution to represent their brand. This is not a new thing either. We kind of touched on this about six months ago. We said, hey, this thing

is brewing out there in the community, especially in the GNOME community. And now we're really seeing it being taken to the next level. And there's also a response to this letter that tries to point out in its own letter that this crap's open source. So no matter what you do, we can still get to it and theme it. And it is, quite frankly, much, much, much cheaper for us to retheme these things than it is for us to create new ones from whole cloth. And as long as it's free and legal for us to do so, and the code's right there, it's always going to be in our best interest to retheme this stuff. So why don't we just try to come to some common ground?

Yeah, that does seem like a good approach, right? Acknowledge it's going to happen. Maybe we can build things in a way that at least mitigates some of the worst cases of their examples, which, I mean, those do happen. They make a good point here, Cheese, and I want to get your thoughts on this. They say, the problem we're facing is the expectation that apps can be arbitrary restyled without manual work, which is and has always been an illusion. Like, if you restyle it, there's things that have to be fixed up

and tweaked and changed. That's just not how design works. Well, I think that first off, if you want Linux on the desktop, these are growing pains for that, right? Everybody, like you said, they want GIMP to look like GIMP. They want Thunderbird to look like Thunderbird,

so on and so forth. And it is free and open source. So regardless of how they feel, people are still going to iterate on this and they're going to continue to do what they do. I mean, I think that we do need to find some common ground, a kind of unified standard that everything can be designed under to some extent. But that's also part of the reason that I got into Linux was because you could tinker with it and change it and make it into what you wanted. So to say that all of a sudden, this isn't cool, guys, stop doing this, I think is just you're pushing

people away from these different distros and Linux in general. As for the iconography and the branding and redoing that, you know, that comes on the heels of the purism stuff too, and them lifting some icons that maybe they weren't okay to use. to use. So that, to me, that's the bigger problem, is that if you take something and brand it as your own, and then put it into something that's going to be made commercially available, there's a problem with that. And you should not do that. From a moral standpoint, really. Absolutely. Absolutely. Part of me thinks with this whole situation

and Cassidy talking last week and Gnome and just everything that's going down, I see why elementary kind of went its own way to develop its own thing. Right. They might have the most sane approach when you play this whole thing out, long term, especially if they work in a dark mode. Absolutely, man. And they will, right? Because

they're listening to the feedback and they're going to respond to that feedback. So they absolutely will. And I expect that when they do it, they're going to nail it. They're going to do it right. I want to pick up on something you said earlier. You said one of the things that drew you to desktop Linux was the ability to theme it. I'm going to echo that. I thought that was like, oh, this is next level power user. Because you remember one of the stupid things you could do in early Windows? You could get that

stupid plus pack? Window blinds, I think it was called. Yes, or you could go third party and do window blinds. Yeah, I thought that was the bee's knees, man. It was kind of the first step in. Oh, look at all the tweaks I can make. So the I thought that was the bee's knees, man. It's kind of the first step in. Oh, look at all the tweaks I can make. So the fact that that was

built into my desktop was like, oh, I'm amongst my people. You know, that's what I mean. Exactly. But I also, I completely connect with where these developers

are coming from. It is probably one of the most frustrating experiences in the world to build something and care about it and then release it and an end user has a bad experience not because of a bug you've introduced but because of something that happened simply by a style sheet that that developer never expected

hiding something making something hard to recognize making the ui less intuitive or even break in some cases and they do have some examples where that. I would imagine that's got to be supremely frustrating because it's just something that's totally out of your control. Like you did everything right. You made something good. It wasn't a bug you've created, and yet the user is still hitting a wall. And you've got to sit there and go, this is impossible. I could never be successful.

And I could see how they'd want to get together. They'd probably get together at events and kind of think, we've got to do something about this. This has been in the works for a long time. When you look at the developers that are on this list, a bunch of great apps on here that I like, the maintainer of podcasts, the maintainer of Uber Rider, and Fractal, and Gnome Games, and Lollipop, and Passbook,

and Authenticator, amongst others, are on this list. So I think even if you as an end user react to this and say, don't take away my theme, which I can also completely connect with, I think it's worth hearing them out because we've got to find some common ground here. And this is a direction that I think is going to become more and more important to these developers on the GNOME side of things. Well, I think, like I said, this is a growing pain for desktop linux right at some point we're going to have to come to this crossroads where we say hey guys we need everybody

that's that's developing a distro around gnome we need to set some common goals for gnome and how we how we style things what's expected to be here what's not expected to be here, and then cut you loose and let you do what you do best and create an awesome distro for us. You know, and I think too, like in a way, if we don't figure this out, if you ever did have a Photoshop or a Premiere or, you know, one of those applications like AutoCAD that people live in to do their jobs and they could like, it's one of the top of the list of things we really need on Linux for people to switch. You think they're going to use GTK or Qt and then have the user themes or distro themes

break the experience? No, they're going to build their own custom toolkits. They're going to build their own custom UI elements, and they're going to ship that just because they can't. I mean, look, that's what Reaper does. Yeah, right. We're using Reaper right here to record this show, and that's exactly what they do. It's kind of a shame, though, because then you don't get to take advantage of all those new mean, look, that's what Reaper does. We're using Reaper right here to record this show, and that's exactly what they do.

It's kind of a shame, though, because then you don't get to take advantage of all those new technologies. Yeah, that's true. It is like, she says, a growing pain, I guess. Yeah, we are having, instead of them shipping this thing that's meant as a blank canvas that you sort of drop on and theme and customize, they have a higher level goal, right? They're trying to

make this whole end-to-end package and experience for you. And that does mean pruning some branches. You're just not really allowed to go down. Couple of housekeeping items today. So I think for this one, I say we kick off the Joe Rez original. Let's do it.

Now, first up and foremost this weekend, ladies and gentlemen, it is finally here. Texas Linux Fest. 2019. It's happening. It's happening. May 31st through June 1st. Link in the show notes. And there'll be a family dinner too.

Friday night at meetup.com slash jupiterbroadcasting for details on that. If you're going to Texas Linux Fest and you want to hang out with L, I, and I and others. I almost got it out. Cheese and Alex aren't going to be there till later. They're missing the dinner? They're going to be there, though, but just not till later.

I actually might get there a little earlier now. Really, girl? Yeah, you know that. I did not know that, actually. That's great. Well, you do now. That's great. Because otherwise, L and I were just going to totally talk trash about you the entire

time. We might still happen, but. That's never stopped me. Shocker. All right, okay. So that's coming up really, really soon. And I know a lot of you can't be there, but if you're going to be there, I'd love to see you at the meetup.

I want to also mention, now that we're doing the housekeeping, Friday stream, fridaystream.com slash five. I've got some big news. You know, I've been thinking about this. Like, a lot of our shows, like this show, it's about Linux and the open source community. Oh, that's what it's about. Okay, okay. Linux action news, you know.

Same thing. But this show, the Friday stream, it's really about us. It's just totally a selfish show. It's really about the people that make these shows possible. They're warning. Their stories. And some of the stories we're interested in that don't fit in other shows. Ways to express the stuff we are interested in.

We did talk about some SpaceX stuff in that one. Yeah, why do you talk about all the space stuff when I'm not on? I know. Yeah, you got to come on the next one, actually. Because there's a problem that came up in the show that only Wes Payne can solve. Uh-oh. Yeah, that's how Brendan and I ended up going to Portland. In the Friday stream, we were chatting about it. And in this last one, I'm sharing some life.

The wife joins joins and we share some big news. Some big news. Fridaystream.com slash five. Yeah, things happen on Fridays. You got to watch it. Also, speaking of things happening, we should mention Linux Academy just did a bunch of new content updates and just recently pushed out 26 new hands on labs with a new interface and a lot more. No better way to learn than actually doing. And a great example of this is if you just finally want to completely wrap your head around SELinux, a hands-on lab is such a great way to... There's probably no other way to get it, right?

Yeah. Also, the community edition will begin its June rotation soon. So if you're not a community member at Linux Academy, which is the free account, where Elle maintains a fresh rotation, go sign up over there. And we'll have a link to the Linux Academy weekly update, which has some information about the new UI stuff and the new labs. Go check out the YouTube channel at LinuxAcademy.com on YouTube.

Some good stuff there. Also, speaking of the community edition, the mobile app has that now, too. So you can get in on that stuff. There you go. That right there, ladies and gentlemen. That is the housekeeping, as I have it at least. Did I miss anything? I think you got it all.

Perfect. Thank you. Ten out of ten. Thank you. Big shout out to Alex over there for working on that UI, UX stuff, man. Yes. It looks good. It does.

I know. They got a good team over there. So we have been doing some legit straight-up journalism. I feel so proud of us as like a team right now because we are covering multiple events across the world at the same time. And I don't know, man. It's just like big time. It's something we could never have done on our own. And now as part of a large organization, I just feel like we can crush some of these things.

And so Elle had an opportunity to go to KubeCon, and she decided she would bring Wes Payne. Drug me alone. So it was a great event because I got to hang out with, what, 7,000 of my closest friends. And the keynote was really focused on the concept of diversity. Well, the keynote on the first day. There were lots of keynotes. on the concept of diversity. Well, the keynote on the first day,

there were lots of keynotes, and they really did a lot to promote the diversity that was going on for not only the projects, but the contributors behind the work. That's great to hear. Well, let's fast forward to, I believe, it's Friday, Westpain wakes up and takes us on a stroll to the keynote. This is obviously a huge year for Kubernetes

and the Cloud Native Computing Foundation. The estimates I've heard is that there's 7,700 attendees this year. And keep in mind, Kubernetes didn't get a 1.0 until 2014. And I know, we keep calling this KubeCon, but really, it's KubeCon and the Cloud Native Computing Foundation convention, because that's a big part of it. Sure, Kubernetes is the platform, but the cloud native family of tools and open source utilities, they tie this all together. And that's a lot of what people are here to talk about today. Yes, you're probably running Kubernetes, but maybe you're just running FluentD.

I've also been very pleased to note that while there are plenty of business, marketing, sales, and the usual assortment of team members, there are a lot of engineers here. You can tell. You walk up to any of the booths in the sponsor showcase, or you just look around

at the clothes people are wearing and the name tags, and you can tell. The people that are here. Sure, there's also some, there's operators and sysadmins, but primarily it's application developers. These are people who are using Kubernetes and users excited about the possibilities of this platform.

And that's a big difference when you compare it to something like the Open Infrastructure Summit and the OpenStack community, which is a lot more focused on the people running the software and not the people targeting the software. All right, now I'm going to play a couple of moments from that keynote that Mr. Westpain was hustling to from his hotel. So yes, Kubernetes in 2019, don't stop believing. So let me set this talk. leaving. So let me set this talk. So there's been 31,000 contributors to Kubernetes. Oh, my God. That's a lot of contributors. That is a lot of ideas. That is a lot of thoughts. You know,

that's a lot of time. And if you sit in the SIG meetings and you see how the sausage is made, you understand, wow, that's crazy. If you look at Slack, that's crazy. There's been 164,000 commits. This is a large, large project. Brian is a VMware senior software engineer, and he makes one other point that I want to play before we circle back to this that I thought was very interesting that I want to get Wes's take on. Kubernetes is a cloud-native platform for building platforms. Sounds weird. A lot of people don't think about it that way.

Kubernetes is not the end. It's not a product. You know, we don't, you know, unless you're a cloud or a vendor, you don't get anything from selling Kubernetes. The real power of Kubernetes is what it allows you to build. And that's how I want us to think about it. Kubernetes is a cloud native platform for building platforms. And then this was Wes's thoughts right after that keynote, which we'll pick up from. That wraps up the keynote for the day, the last and final day of KubeCon EU 2019.

I think the theme of don't stop believing is very much on point because there's something of a tacit acknowledgement here. It's an undercurrent, buried a bit below the hype. There's a lot of companies here and a lot of them are using and working with Kubernetes. But there's an open question of how much of this is production? I think by and large, most of what people are doing is still in the exploratory stage. And we saw that, as Brian mentioned in the keynote. Kubernetes didn't graduate as a CNCF project

until about a year ago, March 2018. And that was the date when they started to say to the world, look, Kubernetes might be ready for you. It's one thing if you're a developer or an advocate or someone at the forefront who really does have the largest problems. Maybe you're someone like Amazon or Microsoft or Google. Okay, maybe you need to develop that. It's very different if you're trying to just use this platform, or platform for a platform, as they said, to develop things. So Don't Stop Believing is a call, a call to action, a call to continue to action. Many companies have begun the journey to Kubernetes.

And while there's still many problems left, right? I mean, stateful applications, from Brian's point of view, are still something of a maybe. But you can see the momentum. I mean, 7,000 attendees here, engineers, business leaders, everyone. And there is an embrace of this common tooling. Kubernetes, all of the cloud-native computing foundations projects,

the theme is that instead of hiding away in data centers, because we're all using the cloud, because we all have very similar goals, there's this democratization, standardization of the base layer of computing. Whether you're on a cloud provider or doing a little more work and rolling your own hybrid or on-prem solution, these tools are all designed not as full products, but as components of a solution so that you, a vendor, a consultant, your engineers can work together to take these components and build your own custom solution, but at the same time, leaving a robust upstream path, a way that regardless of what company you work on,

there's a community. There's a community of engineers all working on the same stuff because these tools are a common business goal. It's a common business platform too. So Elle, I'm curious, since you were at DockerCon and you said it almost felt like Docker was in the shadow of Kubernetes, now that you've gone to KubeCon, is this the common platform? Did you

get like the other part of that picture? You know, I think that it's definitely going to be when the worlds collide kind of concept because I walk in and I start seeing people from the Docker community from day one. Like there is so many of the contributors of actual Docker, you know, Inc. staff there. They had a booth there. I think they really are wanting to just be that underlying, you know, container runtime for Kubernetes as a standard. I'm thinking that's where they're going to hedge their bet.

Like, that's just my opinion. You know, I didn't have anyone confirm it, but that's really what it feels like. That seems like if I was going to put money, like if we had betting, if we could bet on this kind of stuff. Next time, next time. Like we should come up with something. Overall impressions, Elle, of the sort of, not just the size of the conference, because it sounds like it was huge, but

just comparison to other industry events and the things that they tried to cover. What was your takeaway in that regard? You know, I come in from more of the community architect mindset of how are we going to build the community? How are we going to get the contributors? And one thing I just kept talking to Wes about is how distant this conference feels compared to others. You know, when I would go to open infrastructure at the time OpenStack Summit, there were a lot of user groups that meet, you know, there are established dinners, established times to go in and talk about contributing to the project. And it really is about building up and fostering that contributor groundwork. So people want to spend the time helping this project grow.

I didn't really feel that at the Kubernetes conference, I just really felt that it was more vendor targeted, more trying to sell more trying to get the name out there. The keynotes were great with a message, but I didn't feel there was any follow through. I think if you were to attend and you're one person in 7000, it would be easy just to get lost in the bunch and not really get your money's worth or find out how to become a contributor or part of the community. Hmm. It did seem like they had some they made some efforts like there was a mental health in tech talk. It sounded like that's that's something but not quite connecting with what you're talking about. It was more like, you know, here's some information, take it in, but then what? You know,

it's like handing somebody, I don't know, a deck of cards, but not really showing them what to do with it. Where I felt that, you know, DockerCon with their hallway tracks and, you know, like I've already talked about OpenStack, really kind of give you the tools and then say, okay, go over there and we'll teach you what you're going to do next. Is that your impression too, Wes? Yeah, I mean, it seemed like they were talking the right talk. They had the right intentions, perhaps. They want to find ways to have more contributors or find users and turn them into contributors. But it's not very fleshed out yet.

So unless you already have enough motivation that maybe you're using one of these tools at work so you're familiar and you started sending pull requests, that sort of person, I'm sure, could make their way. But if you're just an interested party coming in, you would feel pretty disconnected. Can we talk for a second about that message of don't stop believing? Seems like they sort of hit on that a couple of times. Like that was the key part of the keynote there.

So yes, Kubernetes in 2019, don't stop believing. What's that about? Why drive that message home? Well, there's, you know, there's just been a lot of change, and it's not that old of a project. So you're just now seeing things like financial firms or telcos or large scientific institutions moving over to Kubernetes. There's been a lot of gains.

There's huge adopters that all the cloud services, all the cloud providers provide these Kubernetes services. But we're not even, I mean, to be fair, we're not even fully out of the hype cycle yet. Yes, exactly, right? So there's a long way to go and there's a lot of work. I liked he had a slide, Brian had a slide there, where he said, this is old advice.

And it was talking about stateful applications on Kubernetes and you shouldn't do it. And he goes, well, here's my new advice. Slide is replaced just with the word maybe. And so it's not there, right? Like Kubernetes isn't finished. There's not, it's not some ultimate solution, but there's, along with Kubernetes

and the whole suite of cloud native applications, they're all communally building building blocks they can further their businesses with. And I think that's been a lot of it is leaving, you know, so we saw the merger of two projects, OpenCensus and OpenTracing. They've become OpenTelemetry. They were very clear when talking about it that they're not trying to provide an end-to-end solution like, you know, some like APM monitoring tool. No, they just want to provide the building blocks.

And that leaves a lot of room for businesses on top. Maybe that seems weird from the open source, but the results, I think, are that all these businesses feel comfortable sharing and working on the same lower-end toolset. Well, because of that, they're all going open source by default. In fact, an observation you made while walking around the streets of Barcelona. But open source is front and center in almost every product, whether it's a paid service or not. Open source is one of the first words you hear. Now, you get some interesting reactions when you start asking about the license. But several of those interactions end up being with one of the lead programmers or architects of the product.

And usually, they just pull up the GitHub page and show you the license directly. To be clear, many of these services are paid commercial offerings. Some are totally open source and provide support. Some are more of an open core model, although I'm not sure they would like to be called that. It is very apparent that this is a community focused on open source as a balancing layer. So I guess if you take nothing else away from all this talk about Kubernetes and containers and all of that is at the very high level, it is resulting in a ton of new open source code. Some of these little companies are going to go away or get consolidated, but they're all still producing open source code that will be useful.

I think many of these companies, I went to a great talk by some Spotify employees talking about how they had reinvented so many of these cloud native projects internally. I think that happened to a lot of companies. Data centers were set up just enough differently that there wasn't this motivation for that shared internal coding and infrastructure. And as one outcome of the cloud is, you know,

everyone was using the same APIs and had sort of the same set of problems. So there became this base that you could easily build open-source tools on top of. It becomes, you know, everybody was super, super worried for a very long time that the cloud was going to mean like all this really crazy vendor lock-in. And it still does. It still will for some people. But Kubernetes is sort of creating this common platform that is vendor neutral.

And that's why I think it has so much excitement around it. Yeah, one of the things that was a big takeaway for me that this was the first time that I've been to any container event where we actually talked about the concept of in production. You know, one of the things that we kept joking around is like, yeah, Kubernetes, everyone's doing it in development, but who's doing it in production? And that was always the naysayer. And this was the first conference where companies are like, you know what, we're running thousand level nodes in production. Here's our failover rate. It was actual information, actual statistics of companies doing it live. Real lessons learned in production. Yep. That is nice. Yeah. They're not

there, right? I mean, it's usually the easy ones, but there is real stuff. It's getting there, right? The first workloads. They talk about reliability as one of the factors, right? So they are seeing in those scenarios where it is well developed and makes sense, better results. You know, it's not just that it's lower maintenance or whatever. The applications are running better. Brian, that senior software engineer at VMware, I have a link in the show notes to his keynote. He also talks about what Kubernetes is not doing well right now. It's a bit of a hype buster keynote, but I really appreciate that.

Yeah, I mean, so the keynotes were at times awkward, but the upside of that is, I think, a reason is it was mostly engineers, not business executives. Sure, there were product managers and people, but these were mostly nerdy people directly involved in these products. Maybe they weren't coding anymore, but they'd been coding for years. Elle, did you get any step tracking? I would imagine you must have gotten. I did. I think we averaged about 20,000 a day, except for the day that we went to the after party and Wes says, we should just walk back to the hotel. And then we were almost at 30,000 steps because, you know, why look at the

map before you leave? No regrets. Oh, man. That's the only exercise I get is going to cover these events. So yeah, right. I shouldn't I shouldn't complain. We were in the same venue that they hold Mobile World Congress, which, you know, that's that's way more attendees. And we weren't using all of it just a little bit. It just meant that between if you wanted to go look at, which, you know, that's way more attendees. And we weren't using all of it, just a little bit. It just meant that between, if you wanted to go look at, you know, all the sponsors being displayed out in the expo hall, and then you wanted to go to a talk, well, you have to walk literally between buildings. It's fenced off in a way, kind of like the Boston Airport, you might say, Chris.

Where you just, there's no choice but to go the long way. Oh, man. And the long way is like a 20-minute walk with actual escalator, or, you know, with like moving walkways like you find in an airport. Yeah, you should note that the only place that they had coffee located at the venue was next to the showcase showroom. So if you were at a talk and you wanted coffee, you had to make the walk all the way through.

That's brutal. Well, I know we've got even more stories that we want to share in the Friday stream when we do that again. I'm looking forward to some of that. And I got a couple of things I think I might ask you in a post show. We should probably keep going. Any other thoughts, Elle, before we wrap up? I think that we are going to see bigger things coming out of Kubernetes. I almost hate to say that because I did want to be a person that say, you know, it's all hype, but I'm really changing my mind on that and really seeing what it can do for businesses. So you know that I work day in day out with OpenShift, which is the red hat version of Kubernetes, if you like. And I agree that it's

not hype. There are huge, huge companies using OpenShift to run their entire websites, shipping millions of parcels every day. You know, there are huge businesses using it. So it's not just hype. I know. And it's making a maturity transition now, which you, I think, pointed out pretty well in your clip about the fact that really we're talking like 2018 is when this thing kind of hit gold master status.

So very good. Well done, you two. Thank you for going over there. How long was the flight? I mean, pretty long. I think I set the record getting back from leaving my hotel door to my apartment in Seattle, 23 hours. I thought Elle would have set the record back since she had all those delays. He beat me by one hour.

Oh, my God, you guys. Nearly 24 hours of traveling. The way there was pretty smooth for me. I got a little layover in Heathrow and then just a short little hop to Spain. But the way back was a bit more complicated. Next year, why don't you consider sending somebody from the JB crew you've already got in Europe? You know, save the bodies. I think maybe they wanted to go, but I'm not sure.

Oh, maybe. I'm even more impressed the tech snap went out then. That's really something. Well, glad you guys are back and glad that the jet lag is beginning to fade. I did not have to go as far for my event. I was lucky enough to notice that the folks at Intel had launched a meetup to explain what they're doing with Clear Linux, and I was all about it. I felt like I had not really been able to they're doing with Clear Linux. And I was all about it. I really, I felt

like I had not really been able to wrap my head around Clear Linux. And this was an opportunity to finally write that. Well, Intel Campus number two for me this time at the Hawthorne Farm Campus, and it's just outside of Portland, Oregon. And they had an event all about Clear Linux with an architecture overview and then some demonstrations. And now they're actually doing a hands-on lab right now. They just also gave out dinner. And I wanted to step outside with Brent and kind of share our thoughts about the presentation,

what they're kind of offering. Yeah, well, first, I think showing up to the campus was really special. It was really neat. The people were really nice here. Going into the actual auditorium to take in the presentation, for me being pretty visual, I thought, okay, there's, I don't know, maybe like 60 people in attendance or so. I think the meetup page at 63 or so. And then, but I felt, at least initially, it was quite technical right away. Really technical. And the people that were talking obviously knew some of what they were talking about. And with the focus on basically the update mechanism and security were what I saw as the

overarching theme really. Some concepts that they're doing sounds really great to me. I think a lot of distributions can put some of what they're talking about in play but in other parts I was a little underwhelmed like they were fumbling a little bit in answering some of the questions from the crowd. So I just wondered, you know, they've been working on this, what, for four or five years now. And I just wondered why the push all of a sudden to push it out and wondered, maybe it needs a bit more polish. But overall, I was impressed. Aha. So that's an interesting thing you picked up on there. So I asked about that. I said, why the push now? And the answer essentially came down to, for the last two years, we've

really been cranking to make this thing good. And now we think we've gotten there. And so this is the first of what's going to be a quarterly kind of event that they're going to hold to tell people about Clear Linux and do demonstrations. And they're refining it with this one. I mean, well, not even refining it. It's really 1.0, this whole thing. It's the first time some of these people have ever talked in front of people before. And so, yeah, sometimes they couldn't answer like the stateless question. They had a hard time explaining how Clear Linux did stateless versus other distributions. And then after some back and forth in the conversation, people were able to kind of grok

that essentially to Clear Linux, stateless means you can install the application and not have to configure its state. It's ready to go. Nothing has to be in slash Etsy at all. And you could delete slash Etsy and the applications will still start and run. That's what their version of stateless is. But that's not everybody's definition of stateless. And they didn't come out with a clear answer to that right away. And I think that's what you picked up on was one of those instances. And that is them working this out. But that message of we've been really working at this thing for two years and now we're

just going to start talking about it like crazy, like they're ramping up a whole messaging program around this. I think she said next month, I talked to Chris, one of their head open source community organizers, and she said that they're doing essentially this same presentation with this same crew, but they're going to take another crack at it in California next month. I think maybe it's their strength is iteration.

If you saw the version numbers on ClearOS, they're pretty high. They're very high. So maybe that's one of their strengths. I mean, a lot of these people are really technical, so iteration is a good way to push forward, right? So maybe this presentation, it's a first go. They're here. They're doing it. Smaller group, maybe in California, they're going to have a bigger group,

some practice and things like that. So maybe it's a good approach. Yeah, I'd be curious to circle back to this in a quarter and see what the presentation is like and see what they're going to have a bigger group, some practice and things like that. So maybe it's a good approach. Yeah, I'd be curious to circle back to this in a quarter and see what the presentation is like and see what they're talking about. And, you know, I'll say this. I learned a few things in the presentation that I didn't quite understand before about how they control versions, about how the updates, how bundles work. Like that was pretty good to wrap my head around.

And I think it makes it worth the trip down here alone. But then to get the bigger picture where Intel is going with this, they're clearly trying to build something for the entire stack. So you have an OS that's optimized for your containers. You have OSs optimized for machine learning. You have OSs that are optimized for developer workstations. Because their whole approach is you can use them as upstream and then bundle up and mix your own software in and make it your own distribution with them as sort of an authoritative upstream. It could be a compelling sell to some enterprises. So we'll see how they do.

I'm glad we made it. You, are you glad you came down here? I'm definitely glad. I mean, it's been great to be more part of the community, and having someone like Intel pushing something out, of course we're going to have to be here and cover it. I'm really glad we made it. And like I said, I got an opportunity to better understand

a couple of aspects of Clear Linux. And I thought maybe I'd share a few of those insights with you. The thing I didn't really understand before was bundles. So I grabbed a little audio, the entire presentation where the audio is, the recording the audio is actually much, much worse than

what I'm about to play. Wow. It's like done over a phone line. So it's quite bad. But I have the link to a video version and a link to the slides if you want to know more and if you want to see the slides to this guy's presentation. But here's a little tiny bit about what bundles are in Clear Linux. So the idea behind bundles is to simplify discovery installation of the software that you need to satisfy certain use cases. If you're familiar with the set of bundles that Goins provides, right now there's around 900 or so, a lot of them, especially the ones that have been around

for a long time, are very use case focused. Some of them are providing, you know, like interpreters for various languages like Python 2, Python 3, et cetera. But others are a bit more use case focused, like application server, cloud control, some of the ones that are shown in the diagram on the right. Bundles are, you can kind of think of them as similar to package groups

and other distros in a way. And they can be more work case specific. They can have a wider range of applications. are, you can kind of think of them as similar to package groups and other distros in a way. And they can be more work case specific. They can have a wider range of applications. Some of them are fairly complicated. And another aspect about packaging on Clear Linux that is unique to Clear Linux is when they resolve dependencies. It's not when you install a package. All dependencies are resolved at build time instead of at run time, which is what you usually see with package managers and other distributions. The main motivation behind this is to keep things simple for what's running on the end-user system,

and also just to speed up update times is another motivation for that. It's kind of neat, huh? I like that they're rethinking some things about how you put together a Linux distribution, and we've seen that in other projects, but few with backers who have the resources of Intel, right? Yeah.

So they might actually get away with it if they can get enough stuff running on their system. Yeah. And Brent, I don't know about you, but I had the sense that we were witnessing the beginning of something new for Intel. Yeah, I like that insight too. It felt like we were there at the start and that

there were some small group of people that were pretty jazzed about what they were doing. And we weren't the only ones trying to wrap our heads around everything they were doing. There was a lot of really insightful questions coming from the audience and some really good answers too. And they were also kind of upfront when they know, we're still learning about this too at the same time. So when we know more, we'll definitely communicate more about that. So it was kind of neat to see that back and forth happening between the audience and the ClearOS team as well. Can I ask, it seems like we're maybe skeptical is too strong of a word,

but we're obviously very curious about like why Intel's doing this. What are the motivations? It's funny, especially in the times, right, we've just lost Antigos, and we've seen many small distributions die off. Should we just be kind of happy that there's a new player in the field? Yes, I think that is a default. But I think we're underestimating them. Part of what I got a better appreciation of over the last couple of years,

because remember, it was not too long ago, I went to meet BSD, which was also at the Intel campus. And that's where I learned, like, the number of people in their open source division is over 100. Like, we're talking... That's some significant staffing, yeah. Yeah, I mean, so they have some resources here,

and they've been really heads down for about two years on Clear Linux. And I think they're kind of getting to a point now where they're pretty proud of what they made because they've built a couple of different scenarios where you can see there's some clear advantages, and they're able to tout those use cases. And it's compelling. In fact, they don't really talk about this very much,

but even on AMD systems, you can see clear optimizations and benefits of using Clear Linux. Yeah, it's not just Intel. So they've done a good job with it. Yeah, it's really anything back to Haswell, which is, you know, any modern CPU that is, yeah, it's going to work great.

And they have a very sharp community manager, Carrie, who I met. She's also, she's doing like a, do you remember the name of the podcast, Brent, that they're doing? Something like Open Source Voices. Or Clear Voices.

Clear Voices, something like that. We'll have to find it. Yeah, I'm going to talk to her more and see if maybe we can't be of assistance in some way. But she was really sharp, and I think they have a good plan of how to try to get the word out there. They're just getting started.

So I think our skepticism that we have today at mid-2019, at this time in 2020, will be gone. It'll be obvious what their value is. I think. And I think they're going to do a good job explaining it. Well, I've got it running on my system, and maybe I'll just let it stick around for a while.

What did you think of the install? Did you use the GUI installer? I did around for a while. What did you think of the install? Did you use the GUI installer, or did you use the... I did use the GUI. What did you think? For me, it was actually surprisingly smooth. Really? Yeah, I didn't mind it. And now you have GNOME?

Or what, what'd you do? Yeah, well, I mean, I'm back in Plasma for the show. I wasn't ready to do unplugged on it. Did you do it in a VM? I did it in a VM, but I booted into it as well. Did you keg-zack into it? I mean, maybe. I can't believe you.

Well, it's just the easiest way to try it out, okay? Okay, well, how come you don't show me how to do this? This is something I want to know how you do. Okay, that I can do. So you installed it via a VM, and then once it's a VM, you k-exec into it and run it like it's on your actual machine. Then at that point, it thinks it's running. I mean, it is running. Your kernel switches over to it, right?

Yeah, I mean, there's like a little overhead with the way the disk is set up right now. It doesn't have to be that way, but thinks it's running. I mean, it is running. Your kernel switches over to it, right? Yeah, I mean, there's a little overhead with the way the disk is set up right now. It doesn't have to be that way, but otherwise it works great. How does Kexec compare to Chirrut or an LXC or something? The biggest difference is just that it's booting into a new kernel, basically. So you could do it with Grub, too. The Kexec part is just a little smoother, or faster, I guess. You don't have to make

a modification to your bootloader. The host OS just goes away? Is it like just pulling the power point? It loads the new kernel in memory, and then jumps to the new kernel.

So as far as the processes are concerned, you've just pulled the power on your machine? Yes. System D has a nice command option, so it will do all

the shutdown stuff, and at the very end, instead of issuing a halt, it'll do a k exec. This has been built into the kernel for a while now. Yeah, there's a whole bootloader based around just using it, right? It's such a great... Between that and hardware pass-through with my virtual machines,

I just feel like I could make it an entire release cycle without having to repartition. I could try out different distros with full hardware support. At Clear Linux, though, as a workstation, that I feel like I would want to give a go on physical hardware for a while. To me, it's a clear winner on the server and as a container platform, depending on what your different workloads are, what your requirements are.

But as a workstation OS, maybe. They do have support for Flatpak, so that does open the door to a lot of apps. That's what I'm curious about is how well they've changed the idea, right? They've gone from packages to bundles. How well does that play out? And I'm not sure. On a server, I might not interact with that as much as I would on a

desktop. And it's a rolling release distro to the extreme. Like, they sometimes have three releases in a day. Yeah. And what's crazy is there's a way to build every version they've ever released and they keep on their server

the source code and compiled versions of every version they've ever released ever. Brent, do you remember what their version number that they're at now? It's something like in the thousands, like maybe the multiples of thousands. Yeah, they iterate on tens, if I remember correctly.

So, but the number was at like 2,500, something like that. It was ridiculous. Yeah. And so they keep a copy on their servers and they were just moving around cloud providers. And they say when they apply aggressive deduplication, the whole

Clear Linux source stack is about 16 terabytes, but if they don't apply aggressive dedupe, do you have a guess? You have a guess what it would be? 100 terabytes? Very close. 76 terabytes. Yeah. 76. So they said,

we might not do this forever. We might not keep every build forever. A giant immutable storage of all packages. I would love to know how big the Ubuntu package archives are, just for comparison. I wonder. You figure, Linux has been around for a couple of years

and it's not large. They can't do that forever. They are Intel, though. They're moving fast, so generating lots of versions. Yeah. Yeah, it looks like, I think it's a, I feel like it is a Linux distribution

that deserves to be here. All right, so, so when do you think we're going to see the first distro based on Clear Linux? Oh! But, you know,

with its own branding and website. It seems like that would be pretty doable because they have this whole mix system where you can mix your own Clear Linux that has stuff pre-built in and ready to go

that's still based on their upstream stuff. So it's almost like a distro re-spinner's dream in a way. They've already thought of that and made it easy for you. Yeah, yeah. Or does that just mean that they won't need it as much because you can do it yourself?

Have you thought of a timeline? Have you thought of a timeline? Because I had, you know, you just sprung this on. No, I just think that might be a sign of a success at some point. Hmm. So the numbers are in. Pixel Geek from the IRC says the version number current is 29,660. Oh! How about that?

That's their version number. Go, baby, go. 29,000. Version 29,000. Take that, Chrome. Wow, yeah. I just missed a decimal point. No big deal.

Oh, yeah. Jeez. Well, anyways, you're right. Anyways, it was a cool insight, and I feel like I understand them a lot better now. And I'm kind of rooting for them. Are they going to do more of these events? Yeah, frequently.

In fact, I think next month they have one in California or something like that, and then they're going to try to do one, I think, there at the campus, like nearly every quarter there in Portland. Wow. That's a lot of outreach. Yeah, it is. It is. They did a catered dinner, too.

Wow. First class. They really had a nice setup. You get there, and it's structured beautifully. You start with an architecture overview that was really well presented by, I believe his name was Patrick McCarty. And he did a great job of going through the overall architecture of it. Then you got a couple of use case demos. So they did some onstage demos of how Clarelytics can be kind of handy.

And they were legit demos. Then you break for dinner. And then you come back and it's hands-on. You set up Clearlytics. You go through some scenarios with the engineers in the room there. You can just bug them for any questions or tips. Yeah. And the Intel campus, they have auditoriums.

In fact, I encourage everyone to go check out the blog posts for, don't call it a blog. They're really articles. They're not blogs. They're articles. For both Wes and Elle's trip to Barcelona and Brent and my trip to Portland, they're both beautifully visualized. What's the word I'm looking for?

Come on, Wes. Help me out. Great photos. You guys, you took, did you take that photo of Elle? Yeah. That was a great photo. It worked out really nicely. I mean, and, you me out. Great photos. You guys, you took, did you take that photo of Elle? Yeah. That was a great photo. Yeah, it worked out really nicely. Elle, I'm just saying.

That photo that Wes took of you would make like a good Twitter banner, you know? It's a little scary because I never saw him take them. That's always the best ones. Those are always the best photos. In your element. Yeah, and you can see me being a goof. In your element. A hint of some of my big news. Those are always the best photos. In your element. Yeah. And you can see me being a goof.

In your element. A hint of some of my big news. Those are both posted. Brent wrote up the trip to Portland, and Elle wrote up the trip to Cubicon, and they're both posted now. They're in the show notes. But if you would like to just go over there, if you click on the little blog link, I don't know why they call it that because they're articles. Articles.

Obviously. Or you go to linuxunplugged.com slash articles, then you'll see them. If you click on the little blog link, I don't know why they call it that because they're articles. Articles. Obviously. Or you go to linuxunplugged.com slash articles, then you'll see them and many others, like my factory tour of System. Remember when we went to System Center? That was a lot of fun. It's on there, too. Great write-up, both of you, Elle and Brent. Thank you, guys.

Yeah. And Cheese's previous write-ups about LinuxFest are on there, so go check that out. We don't talk about it very much, but this is an audio podcast, and there's some great visuals right there on the blog, so you can go check it out all over at linuxunplugged.com. And while you're over there, go to linuxunplugged.com slash contact and send us your feedback. Of course, we've got this wonderful virtual lug,

but not everybody can make it on 2 p.m. We get it. It's a hard time. Tuesdays are work days. And 2pm is... That's Pacific time. We're at the end of the world. Who's in Pacific time? Right? Yeah. So I totally get it. So let us know what you think.

Go to linuxunplugged.com slash contact and also linuxunplugged.reddit.com We monitor both those and Telegram. jupiterbroadcasting.com. Now the show is live. I'll be doing it from Texas next week. We'll see how that goes.

Thanks so much for tuning in. See you back here next Tuesday. So So the one thing that we didn't talk about in the show, because it would be completely inappropriate, but I'm curious to hear your take and Elle's take on how the press aspect of KubeCon was. Did they have a media room? Was there media food and all that kind of stuff? Yes, there was a media room and they had like a little area

for quiet meetings. I actually saw some pictures. I guess someone had recorded what looked like a podcast in there. Although the room was way too reflective for my taste. We avoided it. But they had like spaces to sit

and access to power chargers and that sort of thing. What did you think? How did it stack up compared to other events, Elle? Yeah, the room was pretty barren. I would say that they were just kind of like, here's an extra leftover room that we had.

And I love that Wes says that they had access to power, you know, chargers and adapters because we couldn't use them. We forgot that not everyone runs the same way the U.S. does. Oh, I want to hear all about that. We're going to tell that story in the Friday stream. So I would say it was definitely, they had some stuff. So we got to see Jim Zemlin basically fawn over Google and just thank them

over and over for giving the world Kubernetes. So they had some people, and they had like a little panel we got to see, but that was it. It was like kind of media stuff on one day in a little room with terrible audio. They couldn't hear any of the questions coming back at them,

so nowhere near as good as the Red Hat Summit. Ah, interesting. How was the food? Well, it was all right. Al had some misadventures in food because her and Barcelona tastes just don't really align. That's another Friday stream. Yeah, we'll talk more about it on Friday.

Al lost six pounds. Oh, no. Well, that's just because Wes has you walking everywhere. He did the same thing at Red Hat Summit. Now, we did do our best to try to eat all of Amazon's money. They had one of the after parties we went to. So those snacks were pretty good. Were the parties all right?

I drug him to every single party, I think. Atta girl. Good for you. Did he complain, Elle, or was he a trooper? No, I think he made friends with my friends. That's Wes for you. You have nice friends. You know what?

That's one of the great things about hanging out with Brent, too. Wes and Brent, same way, it's like, you guys are pretty much down for anything. Except we hate each other for some reason. Yeah, right. Nice. Brent, I had such a great time. Again, I'm just so glad we did that. Brent's down for whatever you want to do.

You want to sit around and watch some YouTube? Brent's good with that. You want to go take a crazy-ass road trip to some coastal town so that way you can get tacos? Brent's good with that. You want to get pulled over? Sure. Oh, yeah.

Oh, my story about getting pulled over with Brent in the car, totally embarrassing, in episode five of the Friday stream. Were those prawn tacos? Yeah, they were. They were prawn tacos. Yeah, they were. They were prawn tacos. Yeah. Have I told you

about prawns? I'd like to take a moment and tell you about prawns. Oh my, oh my. It's something we're working on. I don't want to tell

you now, but I'm very excited about a new prawn opportunity. Project Prawn? We'll get there. You'll see. You'll see.