LINUX Unplugged - 373: Your New Tools

Episode Date: September 30, 2020

We embrace new tools to upgrade your backup game, securely move files around the network, and debunk the idea that Windows will ever be based on Linux. Chapters: 0:00 Pre-Show 0:29 Intro 0:46 SPONSOR: A Cloud Guru 2:31 LVFS Hits 20 Million Downloads 4:10 Dell Precision 5750 Review Unit Coming Soon 6:27 LVFS Continued 7:29 Xen Hypervisor is Porting to Raspberry Pi 4 12:09 New Dell XPS 13 Developer Editions 14:56 Lenovo Expands its Linux-Loaded Selection 16:48 SPONSOR: Linode 19:31 WSL to Support GUI Apps 24:09 Will Microsoft Switch to Linux? 33:18 Fedora 33 Beta is Live 35:13 Housekeeping 36:13 Exploring Send and Receive 38:06 Send and Receive: Backups 39:37 Send and Receive: Setting Up the Volumes 41:00 Send and Receive: Rsync Comparison 43:40 Send and Receive: Data Retention Tests 48:10 Send and Receive: Comparing Performance 50:09 Send and Receive: Right Tool for the Job 55:29 Send and Receive: Rivaling NTFS and APFS 57:39 Feedback: Todo Apps 1:01:33 SPONSOR: Unplugged Core Contributors 1:02:30 Outro 1:04:17 Post-Show Special Guests: Brent Gervais, Drew DeVore, and Neal Gompa.

Drew, you're just so good. So handsome. Oh, why thank you. It's true. I know. Have you seen the way his hair sort of flops to the front? I was just going to say, he's got that adorable hair thing going on. I was just, you read my mind. You read my mind. Like a giant bunny ear, but for a person. Hello, friends, and welcome into 373 of your weekly Linux talk show.

My name is Chris. My name is Wes. Hello, Wes. You brought a posse with you today. I don't know if we have enough seats for all these fine people. Well, we have a huge show, so, you know, we need a crew. It's time for an in-studio audience for this one. This episode is brought to you by a cloud guru, the leader in hands-on learning. The only way to learn a new skill is by doing. That's why ACG provides hands-on labs

for cloud Linux servers and much more. Get your hands cloudy at acloudguru.com. Drew, hello. I'm impressed you managed to find a seat in here today. I know. It is so crowded in here. I think we're managed to find a seat in here today. I know. It is so crowded in here. I think we're breaking some ordinances, aren't we? Oh, for sure. But, you know, ever since we hit 370, Wes has always rolled with the posse now. This is the new Wes, so we're going to make it work.

But you like my hat, so no complaints. That's true. And I like that you guys are all color matched. That does actually look pretty good. Before we get into the stories, though, we have to bring in our virtual lug. Time appropriate. Greetings, Mumble Room. Hello. Bienvenido.

Hey, rocking 27 people in that virtual lug today. Good to see all of you. Hello. I am thrilled about today's show because not only do we have some fantastic community news that we're going to get into, but then Wes and I had a lot of fun messing around with Linode last night and setting up a couple of servers to just hammer ButterFS Send and ZFS Send versus rsync. And today what we want to do is just give you a plain explanation of what these technologies are and why it's an example of something that these newer file systems offer and how you can use it. But does it really matter when you have rsync and you can use rsync with any file system? We'll talk about that. And it was fun last night. I mean, Wes stayed up a lot later than I did. I'm an old man. I had to go to bed by 11. But we were building these systems and then attaching these disks and then setting up ZFS and then setting up ButterFS and then

throwing data between them and all in the name of podcasting. And let me assure you, all of our pools were named after Star Trek engineers. All right. Well, let's get into the community news first, though. This is just great for the LVFS project. A certain kind of milestone. I mean, numbers are arbitrary, but early this week, Richard Hughes from LVFS announced that they have provided the 20 millionth firmware update out to a machine. Now, here's the thing, though, is it actually

could be a lot more than that, Wes, because there's systems that are behind Nets that just entirely mirror all of LVFS. So that's all one update, yeah. Yeah, yeah. And you know, it really kind of all started, Richard writes on his blog, because of the ColorHug firmware. And if you have one of those ColorHug devices, in a way, you kind of helped start the whole LVFS thing off. Yeah, I mean, he writes,

my experience building open hardware devices really pushed me to make the LVFS free for all. Mostly on the logic that I wouldn't have been able to justify even a $100 per year subscription. And certainly, making the service free in all respects meant that it was almost risk-free for companies to test the service. That's key, right? Because now the LVFS analyzes uploaded firmware for security problems, and it keeps millions of devices up to date.

And here at Linux Unplugged, we keep reporting on how many new hardware manufacturers are adopting LVFS because of this open model. Yeah, it's pretty neat. And it's not without some lacks still still i think i'd still there's still some devices i have that i'd love to see updates but it's just an assumed now at this point when i get a new device which i have by the way i i don't i don't think i've mentioned this on the show have i west that i have this no you have not so uh i don't know if dell just

listened to last week's show and realized that they could probably get me, or if they just happened to send this anyways, and me being in the laptop market right now is just a coincidence. They have sent me a review unit of a Dell Precision 5750. Oh, that's a mean-looking rig. Look this thing up. As configured, it's $4,900, i'll in the review i'll get into why wait wait what yeah uh-huh yeah and dell has put some serious r&d into cooling so they have laptop xeons in this thing that they can run at high speeds because of the cooling system they

put into this as well as a super high-end graphics card i'm trying it out still i'm going to try it for a little bit get past the honeymoon stage throw a few different distros at it, run it through its paces and all of that. And I'll give you a report. But this is so clearly aimed right at whoever would be like kind of that developer market that might be buying a MacBook Pro. that it's thin, it's got chamfered edges, it's got four USB-C ports, plus an SD card reader, and it's got much upgraded speakers compared to what are normally shipped in a Dell. It clearly seems like they're trying to match what the MacBook Pro has in audio quality. And then just the specking on this thing, which I'll get into in the review, is just absolutely insane. And it's obviously Dell saying, hey, by the way, if you heard about this ARM transition and everything that you target in the cloud runs on x86 and you'd rather just get a Linux box that could replace your MacBook Pro, have you seen the Dell Precision 5750? It's really crazy.

I mean, between the like just the really hot RAM and the screen, there's a lot to like here, right? Yeah, yeah. It has screaming RAM in this thing too that's yeah that's one of the other things they did you know i just know that this thing it's going to be supported by lvfs now and that's a pretty sweet position to be in because i could put fedora 33 on that thing and i'm set i could put pop os on that which is going to be kind of particularly ironic or i could leave 1804 is what it ships with right now. And they're all going to get supported with LVFS. It's pretty nice.

There's other things in the works. A few hints. He also gives a notable mention to Logitech, who's done literally millions, shipped literally millions of firmwares. Kind of an early adopter, too. Mm-hmm. And he gives specific thanks to Lenovo. They've really worked well with the project. But there's other details in here, too,

that I think are worth the click in the show notes if you guys care about the project. Some good stuff in here. It's really neat to see this. What I really liked about this post by Richard was just last year he presented a talk, and his slide was well, LVFS is just a website

that runs cron jobs, right? At the end of the day, it has, you know, it's basically new packaging infrastructure. But someone in the crowd told him, you didn't just create a website. You changed an industry. And that's it, right? Like, LVFS feels like a pipe dream. It feels like the kind of firmware updates that we wish we could have in the free software community. And yet, it's 2020.

Here we are. By and large, we have them. Now this next story, I don't know, maybe it's going to appeal to millions. Maybe it's only going to appeal to just me. But the Zen project is officially porting its hypervisor to the Raspberry Pi 4. You got to, you know, you got to have the 4. So to be precise, it's really... 4! Yeah. Well, it comes down to some of the hardware components in there. And this idea kind of bubbled up, what, not too long ago?

Yeah. Well, the idea to do an official port bubbled up from the Zen community, you know, as it will, and then reached the desk of George Dunlap, chairman of the Zen Projects Advisory Board. Dunlap mentioned the idea to an acquaintance who works at the Raspberry Pi Foundation, okay, friends in high places, and was told that around 40% of pies are sold to businesses rather than hobbyists. And Chris, I mean, I think at this point, the number of pies in your RV alone, well, that's a small business, right? Hmm.

You know, now that I'm independent again, I wonder if I could claim it's a business expense. Yeah, I mean, I'm doing some serious business with my pies, and they're now actual production devices, so much so that I could even eventually see myself being convinced to run a Raspberry Pi in the studio, which is my highest level of production grade. We're getting close. Yeah. Now, here's the nutty thing, though. 30 million ARM-based pies have been sold as December 2019, and it looks like if you go by what has been released by the foundation, the Raspberry Pi

sales are running at like a brisk 600,000 plus a month as of April 2020. The register has a really good piece on all of this that goes through a lot of the details. And the thing is, Wes, there is some similarities to the Raspberry Pi and the x86 systems that Xen has run on that made it possible, but then there was also some complications as well. Yeah, I mean, the Raspberry Pi 4 system on a chip uses a regular GIC 400 interrupt controller, which Zen supports out of the box. And I know, I know, especially here at Linux Unplugged, we're all about KVM, but Zen has been there for literally decades. And it's, I mean, it's

fair to say it's huge in the hypervisor space. So the fact that you can just support the interrupt controller on this board, you can use it for virtualization. That's no small feat. Yeah, it's a pretty big deal. Now, the thing is, it's not just that easy. They had to do a deep dive into the belly of the memory allocator and Linux's address translation layers. Better them than us. Well, no kidding, right? And then later on, they had to figure out how to handle DMA and physical address conversions. They had to solve all kinds of problems, but you can read them for yourself.

But the upshot of it is that they're all kind of coming together to land in Linux 5.9. So by Linux 5.9, when that ships, we could have a fully working Xen hypervisor on the Raspberry Pi 4 just out of the box. Now, don't get too excited, though, Chris. Linus Torvalds, in his infinite wisdom, has released Candidate 7 and suggested maybe another cycle of testing will be needed before a full version release. So it might be just a little while before the official Xenon Pi is able to be deployed. But it's coming. Yeah.

And for those of you that can't wait, Project Eve, which is the Linux Foundation's OS for the IoT network edge devices, will soon publish details on their GitHub page detailing how you could kind of get it going sooner. But at the end of the day, I think this is actually a bigger deal for industry that are developing applications than it really is for guys like us. That said, I have given virtualization a bit of a poke on the pi 4 and it's horrible horrible right now under kvm i mean just because the thing is possible does not mean you should be doing it and it would not be usable in any sense of the of the meaning but if they could get something with zen and i could run a super low end base OS with Zen on there. And then I could have a few VMs that would give me a sense of security because being production grade also means easy to restore, easy to back up,

or at least there's a, there's a straightforward process. And I don't really have that with the pies right now beyond DD and the SD cards or, or something like that. Yeah. Something grows. Yeah, something or else. Yeah. So anyways, it's great to see it. Well, let's talk about a little more powerful hardware. We've got ThinkPads and Dell's Oh My this week. Let's start with the new XPS 13 and the developer

edition, which is what Dell calls the version preloaded with Ubuntu, has been announced right alongside the new XPS 13. Now, this is based on the Tiger Lake 11th Gen Intel laptop CPU. They're going to ship with Ubuntu 2004 and Thunderbolt 4. How about that for fancy? What? How does that work? And isn't it just USB? I mean, I think what happened is the USB became Thunderbolt. Essentially,

they've merged into one Ultra Bolt, maybe. Either way, if it's fast, I'm happy. There's some interesting details in here, like fast RAM and going up to 32 gigs. It'll be available in the US and Canada starting September 30th. And then they're going to roll it out across European countries that participate with Dell. The price was not yet announced. But here's the thing in this announcement that I thought was noteworthy. The hardware, it looks great. It's just it's what you expect now from the XPS 13. And they just impressively keep delivering every single iteration.

And that remains true. But this thing in here I thought was cool and perhaps a good data point for us to observe this is dell they write based on their input from our developer community we've also added functionality to allow windows xps 13 users to switch their system from windows to ubuntu 2004 with a free software tool that dell is making available, even if they did not originally purchase the developer edition. Wow. You can set it up so that both operating systems run side by side or that just Ubuntu takes

over completely. That's got to be a big deal, right? That they've made a tool to do this? That is something. I mean, they're invested at this point. Linux is a thing they're shipping and representing in their hardware. I mean, they start with based on input from our developer community. So I guess this is something that they've gotten from their customers.

This is enough customers either requested or attempted this and contacted support or whatever the story is. The number, the data metric was high enough for the leadership or whoever it is at Dell to decide this is worth investing in. That's got to be noteworthy. And that's pretty exciting. So, I mean, the machines look great, too. But I think that's, I mean, like, that's a nice note to, you know, for our audience, just in that consumer choices matter.

You know, if you and your purchasing department perhaps are interested in hardware like this, Dell seems to be listening. Yeah. And our original XPS 13 that I bought forever ago, still going strong. I'm still going strong in the Fisher household. And that's not to be outdone by Lenovo. Lenovo is increasing the number of PC and laptops they sell preloaded with Ubuntu. We talked about how they recently announced some Fedora systems. And at that time, there were Ubuntu systems on Lenovo devices that you could buy if you had like a business sales rep window that you could go through to get those machines. But it seems that barrier is being removed and then they're just cranking up the machines they support. Oh, yes.

That news builds on, you know, that certification program for big companies. But now there's a comprehensive range of ThinkStation and ThinkPad series laptops that are available to buy, or at least will be, from the Lenovo website, most coming preloaded with Ubuntu 20.04. That's a lot of machines and desktops. It's not just laptops. So that's pretty great. A lot of nice work systems to choose from right there. And I hope that means that maybe more businesses could offer some of these systems to their staff. Maybe they're already a Dell shop or a ThinkPad shop. It's just nice to see that integration with IT departments,

especially in the age of Apple embraces ARM and all kinds of custom system-on-a-chips like, you know, the bog-standard x86 platform where you can just run whatever you want. It's still going surprisingly strong. Yeah, and in a way, when the hardware kind of reaches a point where the CPUs aren't changing much and all of that, a way that they can start to differentiate and reach a new market is by offering Linux. And it's not like a huge market, but it is something they can do to the machine that

differentiates. And something tells me the people who are interested in that, well, they're willing to spend the money. Linode.com slash unplugged. Go there, sign up and get a $100 credit towards your new account for 60 days. Linode is the world's largest independent cloud for developers founded in 2003. And they're in 11 global markets with over 800,000 customers. 100% independent. They launched three years before AWS. Linode.com slash unplugged. Get a $100 credit. Wes and I use the heck out of Linode for the unplugged program.

And that's actually kind of how it started for me. Two years ago, I thought I need to have a cloud for my own personal stuff that's separate from my work stuff because I needed to have that divider now. I no longer owned all that. I needed to have my own account. So I knew of Linode, of course, because I've seen them at many of the events I go to. They're active participants in the community. They have developers on staff who contribute upstream. So I'd seen their name a lot and I knew about Linode. So I thought, OK, I'll go try them out. I had actually known a few developers who had written into Coder Radio who told me they host their stuff on Linode.

So I'd gotten over the years a picture and a snapshot. But I was using a different provider. You guys might be familiar with who that was. But I thought for myself, I'll try it Linode. That was two years ago. And then it just kind of built from there. I started doing it for my own systems, you know, when I needed like an SSH jump host or something. But then as we started to go independent again, it just was obvious to build out the infrastructure on Linode. We just prefer it.

The Linux ethos is it's baked into the way you just configure the systems and the access you get and the options you have there, but just all of it. Like I created a volume last night, just as an example, I created a volume last night and attached it to one of our systems. It immediately just comes up with a nice little hover over of, here's the command you run on Linux to mount it with the exact path. Here, if you want to format it to a different file system, run this command. They just, they know what the next step is because they use these systems themselves. They make it simple to deploy infrastructure and you can rely on it. We've been running a brand new matrix server on it flawlessly,

and with 11 data centers worldwide, you're going to find something that works for you. They have shared plans that start as low as $5 a month. Then, of course, they have dedicated CPU and GPU systems that are fast and powerful, and every system has crazy fast networking. That's one of the reasons why we're doing this test that we're about to talk about on Linode

because we wanted to just blast files as fast as possible at each other. And so we set up two Linodes to do it. And you get root access to your server along with their API, so you get full control. Just start by going to linode.com slash unplugged. That gets you a $100 60-day credit for your new account and it supports the show.

Linode.com slash unplugged. And thanks to Linode for supporting the Unplugged program. We had dropped some hints in the past that Microsoft was attempting to support Linux, GUI applications via WSL. And now we've got some interesting details on how X11 and Wayland applications are going to work. And now we've got some interesting details on how X11 and Wayland applications are going to work. And we even saw some demos of good old GIMP running under WSL on Windows 10.

And we have a picture, a literal picture, which we have linked in the show notes for you, that shows how all of this is going to come together. And the secret sauce seems to be that Microsoft is doing what just about every other open source desktop is doing. And yeah, get this, they're writing their own Weston implementation. And it's actually kind of clever. In a nutshell, what we're doing is we're taking Weston, we're taking this RDP backend, and we're extending it to teach it how to basically do remote application. So instead of remoting the desktop, we remote individual window.

And then we can take those window and integrate them with the rest of the Windows desktop. And of course, they have to do the good old GIMP demo, which is always sort of cliche, but proves a good point too, that, hey, look, complex multi-window application is working on desktop. Like, oh, okay.

All right. Well, there's no arguing. You obviously have X apps going. Right. I mean, can you edit images? Well, then you pass the test. Now, how are they doing all of this? This is the interesting thing.

And they say, and we'll have a link to the X.org Developers Conference presentation. It's a long presentation. It's like a big old one. The X.org Developers Conference ran September 16th through the 18th, and this was one of many presentations that was given there. XDC? Yes, that's right. XDC. Thank you. Yeah, it's got a cool logo too, which looks like the X logo, which is cool because it's got an X. They are modifying and upstreaming changes to free RDP to do all kinds of cool things, including create Pulse Audio syncs, pull that in, and push it all to the standard Microsoft RDP client. And inside of GAS, we're also running a Pulse Audio server that's hosting our RDP sync and RDP source for audio in and audio out.

And we forward those over to our compositor where we're running the RDP backend, but we're running a greatly enhanced RDP backend. Like, so we added support for things like multiple multi-mon. So now, like, you can go run all of this in multiple monitors. And as the monitor comes and go, like, the application will see the monitor comes and go.

I did support for things like cut and paste and things like this over like the RDP channel. So we start essentially to have an RDP server that is like fully functional, that allows you to do like almost the same thing that you can do, like say, remoting into a Windows box. And if you go by the diagram that they have, it's free RDP that they are doing this to, and they say they're upstreaming the changes. I think this is so cool in a number of ways. Okay, there's the large Microsoft sort of, you know, standout in the corner. We'll get to that later. But just that RDP is the thing that they invented, and Weston is from the Linux world with these like abstractions first design and that you can merge these two so nicely and have an upstream first mentality to make this all work.

I think that's Linux winning to quote Linus, right? Yeah. I mean like that's Microsoft having to give over and say like, look, this is the way we should design this. This is how we should tie our system in. Even if you guys over in the Linux world haven't even decided that you're all ready to move over to the Wayland-Weston world, yeah, we're there.

Yeah, it's remarkable the effort they continue to put into WSL. I mean, they could have stopped at WSL 1. I remember a time when there were other layers you could put on a Windows system that are like WSL. They actually had a Unix compatibility layer in the past that was on NT4 that you could install. Right. I mean, that's kind of what became WSL v1.

When they announced WSL v1, I thought, OK, that's what it's going to be like based on what I've experienced with their compatibility layers in the past. This is where it'll stop. And, you know, they'll patch it and keep it up to date and et cetera, et cetera. Just be a compatibility layer into their own NT proprietary world. Boy, is that not what's happened? Is that not what's happened? It's like not what's happened at all. No, it's 2020. And Microsoft is shipping like a regular new Linux kernel. Yeah. So let's jump ahead to an email we got in from Chris Thompson.

He says, is Microsoft going to ditch Windows and Linux wins the desktop war? One can dream, true or not, but the day may come. And he links to a register article, which we have in the show notes. But that register article is really about a story that Joe and I touched on in Linux Action News this Sunday. And that was that Eric Raymond predicted that Windows would essentially co-opt Linux and that Windows would turn into a compatibility layer, a Proton-esque compatibility layer that runs on top of a

Linux kernel. And that eventually developers would start targeting Linux and the compatibility layer would become just a option for old applications or old third-party applications. And that essentially you'd see a flip of what we have today. That you'd have the reverse. Instead of a Windows desktop with a Linux compatibility layer, it's a Linux desktop with a Windows compatibility layer. Boy, I don't know. I mean, I think there's still a lot to be said about controlling your own kernel, right? I mean, they've got a lot of say here. And as much as you hear the Microsoft folks playing very nicely with upstream, And again, we should emphasize that. We really should.

But there's also still a lot in the proprietary world when you're running your own cloud and when you control the dominant desktop operating system, you can choose when or when not to play those games because it's your own kernel. They've got their own abstractions and they've figured out how through Hyper-V, et cetera, to host Linux really well. So I think they're in, surprisingly, a very good point.

And if anything, they're in a better spot now from rejecting Linux as a cancer to embracing Linux where they need it, right? And where not, they've still got their own whole NT stack for the next generation of Xbox. And when they want it in Azure, they've still got their own whole NT stack for the next generation of Xbox. And when they want it in Azure, they've got Linux-based Switch hardware. Yeah, that's how I see it too, Wes. Eric Raymond makes the argument, though, that on a longer timetable, maybe a decade or something,

the economic pressures will just kind of force Microsoft to shed their historical tech legacy just kind of force Microsoft to shed their historical tech legacy and just enjoy the cost benefits of having other developers create software. And he makes the case that you're already seeing this right now. There's evidence today for you to witness this. And that is the release of Edge for Linux. He says that that shows you what Microsoft is doing is they're taking free or open source software and they're packaging it up and saying, look, world, we've made this business safe. The adults have come into the room. They've taken this and they package it up.

Yeah, those kids and crazy doctors over at Google came up with this web browser. But what we've done here is we've gone and given it the Microsoft audit. We've gone through it and we've added some features that we think make it better for your business. And oh, by the way, you can centrally manage this via group policy and actor directory. We think it's going to be great. And now you can run it on Linux too. And Microsoft saved themselves a decade of work in investing in a new web browser engine

like Triton. And instead, we're able to adopt Blink and Chromium and just essentially switch over to Edge in like a year and hit the ground running. And he argues that as a company, they're going to look at that return and say, well, that's clearly what we have to just do for the desktop now, too, because Azure is really our long play. I really don't see this argument. For one, Edge is not exactly the first thing that Microsoft has put out, even with a GUI. No kidding. Like we've had SQL, PowerShell, VS Code, Teams, all kinds of things have been coming over to Linux. So it's not like

this is some, oh my God, they're suddenly releasing software for Linux. No, they've been doing it for a while. This is not really something to start shouting from the rooftops about. And I really don't see the momentum going towards having a switch to Linux as a base. I just don't see it. There is way too much investment in the enterprise space. And yes, I know that they're pushing people towards the cloud for as many services as possible, including Active Directory, including Exchange. All of it is trying to go towards the cloud. But I want people to keep in mind that we have entities like the US government who are still paying them to support things like Windows 95. That's not going away anytime soon. These big on-prem deployments are huge moneymakers for Microsoft, and they're not just going to turn

away from that. And they're not going to tell their clients, hey, switch over to this thing where Linux is now your base. And we're going to support all your stuff via a compatibility layer. It's fine. That doesn't fly for mission critical. It just doesn't. I completely agree here. I think Microsoft has never been in a better position not to need to do this. Microsoft has never been in a better position not to need to do this.

And additionally, there's going to be one clear tell that you have to see before this is ever going to happen. And if you don't see this, then you don't need to worry about it. And that is, in some shape or form, DirectX is either open source or released for Linux or whatever. Because there ain't no way Microsoft is abandoning DirectX. It's a huge part of their strategy. And the only way they could do a transition like this is if they battle tested DirectX in a transition like that. All the hardware vendors would be utterly upset if Microsoft switched from a Windows NT kernel to Linux for two reasons. The first is switching to a Linux kernel means that they would have to constantly work with their drivers and have to potentially mainline it, which... That's a lot of work. Well, we've been lucky so far with the way that things have gone for the vast majority of hardware vendors. We were bootstrapped by people reverse engineering drivers, and then it got to a point

where there was this kind of critical mass in certain sectors that led to people starting to do it kind of first, then putting it, and then other people taking that work and then mainlining it themselves. But you rarely ever see the actual vendors themselves directly contributing to the Linux kernel to add new drivers. That just doesn't really happen very often, even today. And basically, the surefire way for Microsoft to ruin all their relationships with all of their hardware partners is to say that we're moving to a Linux kernel and you have to do this. Yeah, they'd be asking them to flush years of work. It's not just the years of work. It means that they actually have to care. See, the biggest problem with Windows drivers, and one of the reasons why it was so notorious for,

what, 15 years or so, is that when you make a Windows driver, you're building it against a stabilized interface that Microsoft guarantees for X period of years. Well, once they build the driver, they don't have to care about it. With the exception of graphics drivers, almost no drivers get updates after they're initially made and released. Yeah, fair. That does not work in the Linux world at all. It seems like every time

some big new piece of Microsoft, something XYZ on Linux happens, this conversation comes up again. And Edge, I think for those of us who have been around a long time, represents a pretty strategic move because we're familiar with how Microsoft leveraged Internet Explorer in the past.

And so it invokes maybe even a deeper analysis of this topic. But I think the fundamentals remain the same. I was pretty optimistic. When I read that article, I was pretty optimistic and I said, why not? But now I hear Drew and also the arguments of Nate, I want to say, these are really good arguments. And if you know that thing behind

Windows, all the relationship they have with their hardware vendors and everything, that sounds pretty obvious. So it was really interesting to hear these arguments right now. So thank you, Neil and Drew. And I think maybe the argument comes up because there is some logic to it. Like it would make it would be great for Linux in the sense that a lot of games would come and it would mean that the Linux user base and the underlying utilities that were all GPL would remain. And it would mean the unification of developer support. And that could be great. Like there's a lot of pros to it, if not a lot of risks.

But it seems obvious that Microsoft would also reap from a lot of benefit of not having to do the plumbing. And the other elephant in the room here that we all are not really saying out loud is no doubt Windows is a horrible rat's nets. It's just got to be just a mess of technical debt. Sounds like you've been looking at the recent Windows XP source code leak, Chris, because, oh boy, there's some embarrassing comments. No better way to fall asleep at night. But let's talk about something that we can get

our hands on today. Fedora 33 beta landed a little bit earlier today. And the nice thing is it was a little delayed, but there's a pro in that the final version of GNOME 3.38 landed in this beta. So we can actually get some beta testing on that. Gives a little more time for us to kick the tires on ButterFS as well, which is, I think, probably the most notable new feature landing in Fedora 33. And it seems like things are looking pretty good. It's in good shape. I noticed that you haven't mentioned, you know, Nano as the default text editor. You bastard.

That's my silent win right there, Wes Payne. That's my silent win. The Nano army is out there, and we just do favors for each other in different parts of the community silently. Really, we should say, you know, things like that, swap on ZRAM, link time optimization, these are all things that Fedora always does and advances the state of the Linux desktop because they care, right? They integrate upstream changes between Linux, you know, the kernel, between GNU and GCC, between LVM. They make it all work and we all benefit. between LVM. They make it all work and we all benefit. I've been doing like Linux podcasts for like 14 years. So it kind of takes something special to get me to look forward to reviewing and testing a distribution. And some of them, it happens still, thankfully. And this is definitely one of them. 2004 was one of them. I'm looking forward to 2010, but I'm really looking forward to Fedora 33 because they're pushing the envelope, much like 2004 did with CFS.

This is here now with ButterFS and just equally exciting. And there's something so great about that to be following this every single day on a weekly basis for 14 years and still get excited by this stuff. It shows you how hard these teams really are working on this. So we're definitely going to review this one on the show. I can't wait. How about a little housekeeping? Tidied up around here, a few things you should probably know about.

I want to suggest you consider upgrading to the Jupyter Broadcasting All Shows feed. You can search for the All Shows feed in your podcast catcher, or we have a link in the show notes or a link on the JB site. There's things in the works. A lot going on now that we've gone independent. I don't want to say specifics, but there may be some new stuff,

and I don't want you to miss it. The All Shows feed is a great way to get this show, get Linux Action News self-hosted when it comes out. There's Coda Radio back on the air that lands every single week. There's good stuff in the All Shows feed and one spot to get it all. And when something new comes out, if it does come out, I mean, who knows? But if it does, that's where it would be. And if you're looking out there to reach the perfect audience, Jupiter Broadcasting has some sponsor opportunities in October.

Could be a great fit. Email me, chris at jupiterbroadcasting.com, and let's talk about it. There could be something there. I think this could be an opportunity for the community to reach out to other community members. chris at jupiterbroadcasting.com All right, we got really nerdy for this one, Wes. We got really nerdy. And in a way, it was really great to once again go through the process of setting up a blank system. This is a temporary system, guys, so we went with Arch because it's simple, clean, lean, and mean. Okay, wait, wait.

What? Can I just say, this time, this one time, I didn't pick Arch. That was you. I SSH'd into this box, and you had picked Arch. Yeah, and also, I kind of freshened up the place for you and put on Fish Shell for you, so, you know. Which was so nice. Yeah, I take care.

But don't you think for this kind of build where you're trying to get what you want, modern file system version, you just create a couple of devices, one's ZFS, one's ButterFS, one's Extended 4, so there's a few devices. I mean, I think we've accepted Arch as the modern Gen 2 slash, you know, Linux from scratch alternative where you want to integrate all the upstream changes all at once without the, you know, Fedora important people like Neil and Carl who tested it first will just test it and do a worse job. I've got a new thought technology for you when

it comes to Linux distros. Are you ready for this one? Arch is my learn and burn distro. Learn and burn. You set it up, you update it, you make some mistakes. Well, like after today's episode, I mean, we're not going to keep these two systems running. We'll just, we'll literally just delete them. So, you know, for this, it kind of makes sense. Anyways, back on track. What we wanted to play around with was a feature that is truly fantastic about these new modern file systems. And that is either butterfs send and receive or zfs send and receive. You may have heard these mentioned before. We wanted to talk a little bit about what they are, what they can do for you and kind of the differences between those and, say, rsync. So how about in the context of backups, Wes? Maybe that's a good place to start.

We've been playing with ButterFS. We've been playing with ZFS for really a long time. And we have a lot of small files that change. So we started thinking about what's a good test case to use for backing these things up? And right off the get-go, Chris, you set these systems up over at Linode. ZFS had a bit of a detractor, right? Yeah, so looking at this from a standpoint of one disk or two disks, and we want to use this send and receive to do backups, there is the gotcha that you don't really just format a disk with ZFS. Like you don't just do makefs.zfs and point it at your disk or partition and say, go for it. Oh, no.

With ZFS, at least you have to create a pool to get started. Honestly, this has been a really fun example of when and where do you use file systems, right? Because, okay, so we want to talk about send and receive, right? We want to talk about backups, talk about copy and write, incremental backups, and where it made sense. And I think we've been flirting with, at Linux Unplugged, some of the advantages or disadvantages

of ButterFS and ZFS combined, right? Because we live in a Ubuntu 20.04 or honestly 16.04 world where you have ZFS. But we also live in a world where Chris sets up a Linode Arch server and you have to compile the ZFS module from scratch. And in that world, ButterFS is a lot simpler, right? I mean, okay, did you not tell me last night, I'm going to set up the ButterFS file system. Can you set up the ZFS file system?

Yeah, I just didn't want to mess it up for the test. And with ButterFS, I just did the makefs btrfs, pointed at the device, and I can essentially take advantage of things like compression and snapshots immediately without having to create a pool and setting up any kind of form of RAID. While both file systems offer some of the features we're about to talk about, that's one thing to consider. And I think the point we're trying to drive home is if you're on a laptop or you're on, say, a VPS with a single disk, it may be a more appropriate use case

for ButterFS, despite what everybody's going to say about it being scary and dangerous. If you're in a scenario where you have a large pool of disks and you have essentially this desire to have the file system and the operating system be separate and they don't intertwine, ZFS is a much better solution there. Well, maybe not much better, but it's definitely maybe a better use case for ZFS there. But at the end of the day, they both have this butterfs or zfs send and on the other side, receive command

that allows you to essentially shoot snapshots over the network to another system running butterfs receive or zfs receive and then replicate that snapshot onto the file system, which once it's there, you could treat as a mount and you could actually mount. I mean, there's a lot of options you have there and it's rsync at the actual file system level. Is that

too crude of a explanation? Well, maybe rsync++, right? How so? Well, okay. So we started things off with a split partition setup on our two Linode machines, where we had an EXT4 partition, a ButterFS partition, and a ZFS partition, right? And we copied the same data set over to each. And okay, yeah, this was a super nerdy data set. It was for machine learning. It's basically a bunch of JPEGs of fruit. Yes, honestly, it's fruit, it's apples, it's kiwi, it's pineapple. It's all the things that you might want to eat. Is that a banana or are you just happy to see me? Oh, I'm very happy to see you. Continue on. Sorry. And we set up a script to maybe make this a little more obvious than it needs to be. We had a Python script set up that would just modify a few pixels

in each picture. You know, they have white backgrounds. We changed that background to black. And that's where rsync kind of fails. And ZFS and ButterFS, with their copy-on-write technology, really come to the forefront. Because if you think about copy-on-write, really come to the forefront. Because if you think about copy and write, it's an advanced technology, but it makes a huge difference. Because a lot of our storage systems, well, you overwrite in place, right? You update a file, you just write over the same file. But with copy and write, you only copy if you have to make updates. And that means with Butterfest and ZFS, when you take snapshots, when you have new subvolumes, you only store the delta changes. Whereas rsync, well, it needs to go check all of the files that might exist. Right.

Rehash them. Right. And figure out what differences actually need to get sent over the wire. them and figure out what differences actually need to get sent over the wire. I've heard of use cases where you can use ZFS or Butterfest send over long distances. There's ways to accommodate that too. So it's not necessarily a replacement one or the other for rsync or for the file system send and receive functionality. But think of it as this way, everybody, is it's kind of just baked in now.

So when you have a system that has ZFS or ButterFS versus Extended 4, which I'm not trying to convince anybody to switch, but with Fedora 33 just around the corner, Canonical investing a bunch of work into ZFS, I'm trying to paint a picture for why it's necessary. Well, that's just it. one of the best aspects of this little experiment that we did, Chris, was kind of feeling out, like, what does it feel like to use a ZFS system or a Butterfest system for serious data retention? So here's the setup we had. We've got this large data set of images. We ran a Python script to modify one pixel in each image and then synced the changes both via rsync and ext4, zfs, and butterfs. And wanted to figure out how does that feel? What does it look like? And what are the actual transfer rates? And that actually was sort of where I think you and I did some of our most speculation is I honestly expected r sync to kind of own this, you know, because this is like its thing.

I thought for sure. It's its thing. Yeah. But what's fundamental about this is it's integrated the file system layer and it is aware of the blocks. It's aware of the snapshots. It's it's just a more integrated tool. And surprise, surprise.

I think in most cases after your initial sync, especially right, it sends much faster via file system than our sync. No kidding. Okay. So we should clarify here. We're using bitmaps. So we're updating exactly one little pixel in that bitmap. We're just changing it from a white background to a black background. Not a big deal.

But you could imagine that you've got a large corpus of legal texts or backgrounds for your website. It doesn't really matter, but you've got a small incremental update that you need to sync to a large number of data sets. A good real-world example is Facebook, right? They've got a huge number of containers they run in production, all powered by ButterFS. You might update a few packages as your Fedora release comes due and you've got a few more updates, right? You don't want to download each of those individually on each server. You'd like to just diff the delta, send the delta, and update it.

Yeah, and to that point, Wes, as we saw in that talk that was given by a Facebook developer, or no, it was the Butterfest developer who works at Facebook. He mentioned that they distribute updates to the Facebook website via Butterfest file system send and receives. I mean, they can actually just replace portions. Well, yeah, right? I mean, to give Docker due credit, they support ZFS and Butterfest as backends because they support copy and write fundamentally. as backends because they support copy and write fundamentally. So when you want to make a copy, instead of actually overwriting in place, you make a

new copy and you change it there. And that means these file systems can figure out which blocks have changed without having to do a full differential sync like rsync might do. This really showed up when we were testing this out because rsync took a full 20 seconds to go figure out like, oh, all of the images that you had, well, they each had one tiny bit flipped. ZFS and ButterFS, they took less than a second to figure out the same thing because they've got file system smarts. And of course, that speed advantage just scales the more you're trying to move around and the

more that it's on the file system. Exactly. I mean, really, we're doing the minimal level here, right? We've got like this machine learning initial training data set where it's just a lot of really nice looking strawberries. But once you start doing some real stuff, or you've got backups going for years, right? Where, okay, yes, the presentation from your college days doesn't change a lot. But maybe you've got some to-do lists, some stuff that does have a little more churn,

but isn't a complete overwrite in the last five years. Project files. Yeah, exactly, Chris. We know you haven't still updated those servers. That's true. That's exactly the stuff. It just sends in the most efficient way possible. And honestly, it was a really good comparison between where ButterFS is as a enterprise file system and ZFS.

Yeah. And I think maybe the folks in the enterprise, I could see how people with an enterprise mindset would look at ButterFS if they're not familiar with it and go, oh, that's more of a hobbyist file system. I don't think it's a fair accusation anymore. Really, you know, it's funny how many times we've been working with both these file systems for this show. And now I just clearly walk away from this going, oh, yeah, okay.

I see this use case for ButterFS that I use it for, and I have a use case for ZFS too. And that's why I just use both now. But I'm curious about the differences that you notice in terms of, did you really notice any speed or any kind of real performance difference between Butterfest and ZFS? I know it's not a huge workload, but it's a couple gigs. It's... Yeah, I think our training set here was approximately 2.4 gigs of just lovely fruit bitmaps that we had going. ZFS, by default, uses LZ4, at least if you, like, install on Arch ZFS on Linux with LZ4 compression.

But RFS, well, you've got to enable that. Yeah, that's kind of what I would expect, though. I was actually surprised to learn that it was turned on by default. But I think that's just because upstream has kind of determined that that's the more. But here's where that doesn't matter. You've still got to do your initial sync, right? So if you want to go listen to self-hosted, you should, because Alex has some great stories about syncing, you know, some of his backups over to his parents' place across the pond in London. I've done similar. If you've got big data sets,

the initial sync is still going to take a long time because you've got a lot of bits to send. What matters is the delta. And when you've got a fancy copy-on-write file system, you just don't have to worry that much. So the crux of our test here at Linux Unplugged was that we modified every file with a few pixel changes. rsync, it basically resynced the entire dataset.

Both ZFS and ButterFS, despite the initial slowness maybe that you might contribute to ButterFS on the initial files end, neither of them took longer than a second to update all files. That's a pretty big difference then. So there's a huge speed difference.

They both offer compression. What about encryption with ButterFS? Okay, so now we get into real-world administration details. Chris, I think it's fair to say that setting everything up, you kind of like ButterFS a little bit more. As a traditional Linux sysadmin who just over time I have learned I like things that are built into the OS that are tools like just kernel-supported stuff. And ButterFS to me works like a file system I'm familiar with.

I can just go make FS on a, on a, on a disc and it, it just is there. And it's already installed. When I, when I get going, all I have to do is just install the programs if they're not already installed. And it's just supported in the kernel. There's no DKMS module I had to worry about. So those things, and actually not just those things. I got to be honest, the other thing that really appeals is I'm very commonly working on VPS now. And a VPS system generally has maybe two disks at most, but often just one disk.

And you're pretty abstracted from like whatever that real hardware is, right? You're not talking to the disk. You're talking to like a network interface. And ButterFS just seems perfect in that scenario where ZFS, if I if I owned the box and I had direct access to the physical machines and I was writing ZFS to the actual disk and ZFS was managing the actual physical disk and the RAID and it was it was all top to bottom controlled by ZFS. I think that would be a pretty good ZFS scenario, especially more than a couple of disks. But with ButterFS on my laptop or on a VPS, it just seems to work a lot simpler.

And while I don't have a solution for ButterFS encryption today, I know it's coming and I know I prefer the ButterFS administration style. It doesn't mean our NAS here in the studio is gonna switch off the ZFS, though. I'm not doing that. No, of course not.

Well, so that's what's interesting is that it really feels like ButterFS is a great default, right? You have a mixed usage file system. Maybe you've got some subvolumes that, you know, they're archival, they're backups, they're things like that. And then you've got other ones that are just sort of EXT4++. Whereas ZFS, you've got to commit, right? It doesn't just slot into the sort of Unix layering system. It doesn't slot into how you manage any of the other volumes. When you do a DF-H grep CFfs, well, you get the dataset name, not the actual

partition number, right? So it just doesn't play nice with some of how you might use the Linux system as an admin. But at the same time, it's a standalone enterprise system. And some of the biggest wins it has are resumable send and receive, built-in support for compression, and recently native encryption. So if you just want to send some snapshots, you want to catch things up on the other side, ButterFS is the easiest way to go. And I didn't actually expect this conversation to kind of go into this, but it is sort of the logic for ButterFS is the easiest way to go. And I didn't actually expect this conversation to kind of go into this, but it is sort of the logic for ButterFS. And maybe with Fedora 33 around the corner, it's on our minds. But I'll also say this, and it's not as silly as it used to sound, but you're not going to run ZFS on a Raspberry Pi.

And remember that number we just talked about have been 30 million Pis sold. And that number is based on April data. So it's more than 30 million. And you can run ButterFS like a champ on my Raspberry Pi, at least. You sure can. And so I think there's a lot of use cases for it still. And so I think we've made our case. There's obviously a lot of reasons why it's pretty great.

The send and receive functionality, though, is something that I really encourage people play with once they start switching over to these file systems, whichever one it may be, because this could change your backup game a little bit. And think about it from an application restoration standpoint. You could have a container host or a VM host and you snapshot and you ButterFS or ZFS send that and put it on another machine, it's a hot standby. Mount it as a sub volume, get going, spin the containers right back up, zero downtime. It was kind of fascinating to see how much of a difference it was between ButterFS and ZFS. So like on our little test setup over at Linode, I was taking snapshots

on ButterFS, I was taking snapshots on ButterFS, I was taking snapshots on ZFS. And in the ZFS world, it was like sort of a logical entity, right? I had a data set, I would take various snapshots, they wouldn't show up on the file system. No, of course not. Those were like logical entities that existed within ZFS. But in the ButterFS world, those all existed, like you had to pre plan stuff, so that you had like a top layer namespace that you would take snapshots underneath. And if you want to do a send or receive in ButterFS, you got to make sure that those snapshots are read only. In ZFS, well, they're always read only. That's good and bad because you get a little more flexibility depending on how you think about these things.

It's just fascinating. You can do the same thing on both. I think ZFS has more maturity by far in the sense of encryption, native compression, and resumability for send and receive. But ButterFS has the advantage that it's just sort of built into a modern kernel you might use. That's true. And you could turn on compression and take advantage of that. And you could just not use any of these features. You don't even have to use copy on write and you just get a pretty decent file system for your laptop. I think that's what's what has me so excited about this. This is what an enterprise-grade file system does,

is it has this kind of functionality. And I've watched over the years as it's sort of been lacking on these Linux workstations. We sell these machines supposedly for developers, but they have a file system that has the functionality of the 90s or actually even the 80s. Meanwhile, Windows has shadow copies and NTFS and with granular permissions. I'm not a huge NTFS fan, but I have to admit they seem to have kept it in decent shape over the

years for what it is. Apple has APFS, which offers not as many features as these nicer file systems that we have, but definitely more than Extended 4 has. And it's just kind of expected on a workstation for the file system to have a workstation-grade functionality. And now, just by installing Ubuntu 20.04 or by installing Fedora 33 in the near future, you get this stuff. And you never have to touch it, but it's there if you want it, along with a bunch of other really cool stuff that's built into these

that we haven't even touched on today. But the send and receive functionality specifically is worth a little more attention. Maybe it's a little early now, but we want to get it in front of you so that way when you start using a distribution with these file systems, if you don't already, you can kind of see some of the advantages.

And we also want to arm you with actual knowledge about this file system, ButterFS, that's built into Linux, and try to give you some counterpoints to what you'll always hear about ZFS and how great it is. It's not that it's a bad file system, or it's not that it's a good file system. It's different use cases. And there are advantages to both. And I hope we've made that point. And I hope that we've given you kind of that ammo. So when you're in those conversations, at least for yourself, you know what the reality is, is that not one file system is the answer to all. But the situation's getting better for all Linux users over time. And pretty soon, I hope that it's the default on all Linux desktops.

But that's just crazy, Chris. In the meantime, we have some feedback we should probably get to because we're going to be going on forever if we keep talking about file systems. I don't know what it is, but we love talking about file systems. Thomas writes in with a question about to-do apps. And this is something we need to talk about because otherwise we just talk about file systems all day and I got to keep on track. So to-do app is perfect. And he says, you recently covered GNOME, or I'm sorry, GNOME to-do, which got me going down a rabbit hole. Thanks so much. I'm looking for more To Do applications.

I've set it on DTask for now, or DSTask, after trying TaskWarrior and the GNOME To Do list. And I've also tried ToDo.txt for the CLI stuff, and I've tried TaskBook. Wow. I'd love to hear more about To Do apps. Is that something you've covered before on Self-Hosted or LUP? If not, I'd like to hear more about to-do apps. Is that something you've covered before on Self-Hosted or LUP? If not, I'd like to hear more about it. Well, I decided, Thomas, we'd probably cover this one because I just kind of went all in on Todoist.

And so I thought maybe I'd talk about Todoist a little bit. You know, I've been using Todoist as well. I started looking at it again when they announced a, quote-unquote, native Linux app. That's always kind of, okay, it also works great on my phone. It works on my watch. It works on the web. And now it works on desktop Linux. And they added support for Trello like project boards along with to do lists. And they have support for sharing those to do lists with other folks. for sharing those to-do lists with other folks. Like the wife and I have one for the RV that we just put stuff on individually. And, you know, I have a bunch for work on there. And I found it to work pretty well for me.

It's not the prettiest to-do app, I gotta be honest, but it is super functional. It has an API for days. The integrations with Todoist are just mind-bending. It's just mind-bending. And on top of all of that, it seems like it's going to be around for the long haul. I've watched them for a couple of years. They seem pretty solid as a company. They seem to be iterating at a pretty decent pace and adding decent features. And the

more I drill into the application, the more I find that it does like I have this situation where I have to, again, call back to Verizon Wireless. And so I just keep all of these notes for every time I call these people and I try to get them to cancel my line. I get it. I have it right in there, and I have all the details in my call notes, and I find that to be super handy. Plus, it supports a lot of plain English stuff for scheduling and whatnot. And the mobile app's decent, too.

So I recommend Todoist. Anybody in the mobile room have suggestions for a to-do app that works for them? Super Productivity is one that I've started using. I got a reference to that from another podcast, and that was, I'm kind of amazed at how it makes things better. It's not just doing to-dos and note-taking and stuff like that, but it also integrates with managing issue tracking workflows and stuff like that. It might be a little bit more developer-centric than Todoist is, but it's certainly starting to help me out a little bit in balancing what I'm trying to keep track of and what I'm supposed to do and things like that.

Nice. And the name was again? Super Productivity. Nice. That sounds name was again? Super Productivity. Nice. That sounds like one worth looking into as well. Drew, do you have one that you use any kind of like to-do list management at all? Yeah, I do. But I like mine to be super duper simple. I don't think that a to-do app needs to have a whole bunch of bells and whistles. So I actually just use Google Task and it works great for me. Wow. I thought it'd get

killed, but it seems like it's sticking around. They seem to be integrating it more and more with different aspects of the Google app suite. So I think it's in for the long haul. Well, there's a couple of ideas right there. And if you have any others and you're screaming them at us right now, you could always throw them in the telegram group, jupiterbroadcasting.com slash telegram, join up and tell us about them.'d like to know i'd like you know we get a we get a batch of them we could always do a a little to-do pics special i suppose i want to say thank you to our core contributors those of you that are supporting the show we really appreciate it you're making the show possible and you get more show you have a choice you get the fully uncut live stream with all of our mistakes, but also a lot more content.

Too much content? Is there ever too much Wes Payne? Really? No, probably not. No, no. You can never have enough Wes Payne. And there's two options for you. You have the limited ad feed,

which is fully produced, the fully edited show with the nice mix and all of that, just with limited ads, or you get that full live version. Thank you for making the show possible. And if you're interested in supporting the Unplugged program, if you want to become a member, you can go to linuxunplugged.com. And if you look right along the top there, we have the links to send you over to our core contributor page. And you can sign up, and then you get the feeds for the extra content and support the show. Also, over on the website, along with all the show notes, we have our Matrix server info, if you've been wondering about that, and our Mumble server info. If you'd like to join us on a live stream, we do this show on Tuesdays at noon Pacific, 3 p.m. Eastern.

But get it converted at your local time at jupiterbroadcasting.com slash calendar. All right, Mr. Westpain. See you next week. Same bat time, same bat station. Wait, do we have a bat signal? I thought it was a penguin. I think it's Twitter. You know, when we just tweet that we're live.

You're right. It's at Linux Unplugged, right? Yeah, that's the new bat signal, I think. It's got to be. Or the Telegram group. Because I often post it in the Telegram group. So that's another way to know when we're going live. Thank you, everybody, for tuning in to this week's episode of the Unplugged program. Links for the topics and stories we talked about

and information about the file system features we've been going on about linuxunplugged.com slash 373. Thank you so much for tuning into this week's episode of the Unplugged program, and we'll see you right back here next Tuesday! Thank you. So, hey, Drew. Hey, Chris. review is it possible to script a reaper series of actions so that i could import audio process audio render audio or encode audio and then just run it over and over again uh maybe you know getting into the scripting thing isn't something that i've done a ton of i picture you sitting in front of reaper and you're kind of like just like executing all kinds of automations and you've got like a VR headset on and like a Nintendo Power Glove and stuff is just like flying on the screen.

The Power Glove is key. I was actually picturing Drew with like one of those duck hunt, you know, those orange guns that you see. The light gun. Oh, yeah. And he's shooting at mistakes and they're getting cut. Yeah.