Transcripts

Untitled Linux Show 279 Transcript

Please be advised that this transcript is AI-generated and may not be word-for-word. Time codes refer to the approximate times in the ad-free version of the show.

Jonathan Bennett [00:00:00]:
This week was wild, and we talk about it. Not just Office three sixty five, or is it Office three sixty four now? But we've got a system d and Rust crossover. There's a Blender with Gaussian splats even. We crashed Linux on purpose. We celebrate Ardor's next release, and we talk about how terrible Raid is and what may be coming to save us all. You don't wanna miss it, so stay tuned.

Rob Campbell [00:00:25]:
Podcasts you love.

Ken McDonald [00:00:27]:
From people you trust. This is Twint.

Jonathan Bennett [00:00:34]:
This is the Untitled Lenox Show, episode 279, recorded Saturday, November 14. Slow day at the office. Hey, folks. It is Saturday, and you know that means it's time to talk about Linux. It's the Untitled Linux Show. I'm your host, Jonathan Bennett, and we've got the guys, the crew, the group. And, it's an it's an interesting day. I don't know about you guys, but I was earlier today glued to my TV watching a starship fall from the sky and get snatched out of the air.

Jonathan Bennett [00:01:05]:
That was pretty cool. Did you guys see it?

Rob Campbell [00:01:09]:
No. Missed it.

Ken McDonald [00:01:10]:
Busy doing other things.

Jonathan Bennett [00:01:11]:
Oh my goodness.

Rob Campbell [00:01:12]:
I've been too busy with the email servers and other things falling from the sky this week. But

Jonathan Bennett [00:01:17]:
Yeah. Yeah. That's true. I, I... I'm I'm gonna channel Jeff Gierling here, and I'm gonna say that we need to we need to rename it Office three sixty four or maybe Office 363.5.

Rob Campbell [00:01:28]:
I've seen a name already, three sixty three.

Jonathan Bennett [00:01:30]:
Yeah. Yeah. It was a it was a it was a it was a bad day for certain places and people.

Rob Campbell [00:01:36]:
I found it to be a very relaxing afternoon. You know, when you can't do anything about it, there's no emails coming in. It's like... Oh, yeah. Well, it's

Jonathan Bennett [00:01:46]:
a slow day at the office.

Jeff Massie [00:01:48]:
No email, no Teams, no Copilot. No. It's pretty much Yeah. I'm just going home.

Rob Campbell [00:01:54]:
I don't get to keep hearing that. Teams worked fine for us.

Jonathan Bennett [00:01:57]:
Oh.

Rob Campbell [00:01:57]:
I was communicating, but that's what I heard. Teams wasn't working. People... I've heard that multiple times. I don't work... Had no problems with that. Yeah.

Jeff Massie [00:02:05]:
Most everything for me just wasn't working.

Jonathan Bennett [00:02:08]:
So it was...

Jeff Massie [00:02:09]:
I I literally was like, wow. I might as well go home, and maybe I'll log in later to see if it's still working. But

Jonathan Bennett [00:02:15]:
Yo. Yeah. It was it was pretty, pretty interesting day with... Between that and the ground stop from the FAA.

Rob Campbell [00:02:20]:
Yeah. That was the next day, though. But... Yeah.

Jonathan Bennett [00:02:22]:
It was all yesterday.

Ken McDonald [00:02:24]:
That was yesterday.

Rob Campbell [00:02:25]:
Email was the day before.

Jonathan Bennett [00:02:26]:
Day before. And yesterday.

Rob Campbell [00:02:28]:
It it was fixed by, like, 6AM yesterday.

Jeff Massie [00:02:31]:
Oh, okay. Yeah. It was totally totally working for me. But I know there was a lot of memes running around with the, and for people who are curious, look up John Malkovich Microsoft. Oh. And there's there's a YouTube little short, and there's also memes out there.

Ken McDonald [00:02:48]:
You know? But do that after the show.

Jonathan Bennett [00:02:51]:
Yeah. No. Well,

Jeff Massie [00:02:52]:
it's it's not, family friendly. So Oh. Say what it is.

Rob Campbell [00:02:56]:
I'll pull it up for everybody to share.

Jonathan Bennett [00:03:00]:
Well, that's great. Alright. We've got some we've got some stories about things that are not Microsoft and not the FAA. At least I don't think any of them are. We're gonna... Well, we've got some stories that are about Microsoft, actually, because Rob's here, and he just can't help himself. Really, Rob,

Ken McDonald [00:03:16]:
this isn't about Microsoft. Yes. It is.

Jonathan Bennett [00:03:20]:
It's for... Yeah.

Jeff Massie [00:03:21]:
It is.

Jonathan Bennett [00:03:21]:
The the the Linux... Wait. The Linux subsystem for Windows. Wait a second. What is this?

Ken McDonald [00:03:30]:
Yeah. Let's ask John... Let's ask Rob so we can find out.

Jonathan Bennett [00:03:34]:
That's what I'm doing. Rob, tell us all about

Ken McDonald [00:03:36]:
this one.

Rob Campbell [00:03:37]:
It's not exactly what you first thought. You know? It's not. This is LSW, not WSL. So here's the backstory. Microsoft has spent years making it easier to run Linux applications on Windows through what, Jonathan was thinking I was talking about, WSL or the Windows Subsystem for Linux. I know the backwards stuff, it's confusing. It makes sense if you really think about it, but it's just confusing. But anyway, now there's a new open source project that flips that idea around, and it's called LSW or Linux subsystem for Windows.

Rob Campbell [00:04:14]:
And it can make running Windows applications on Linux a whole lot more practical. Now, before you say, we've been running Windows applications on Linux for years, you're absolutely right. But what makes LSW different? Let's start. Let's start with WINE. WINE is compatibility layer that translates Windows API calls so applications can run on Linux without, actually installing Windows. It's impressive technology, but compatibility isn't always perfect. Some applications, particularly, newer versions of Adobe software, can be difficult at least or impossible to get working, reliably. And then there's the traditional virtual machine approach.

Rob Campbell [00:04:59]:
You install Windows inside something like VirtualBox, KVM QEMU, whatever, and run your applications there. Compatibility is generally better because you're running actual Windows, but you're often stuck working inside a separate, Windows desktop. And getting a good GPU acceleration can require additional configuration or even a GPU pass through. More recently, you know, we've seen projects like WinVote, you know, where WinVote improves the experience by running Windows in a virtual machine and using Microsoft remote app technology to display individual Windows applications directly on your Linux desktop. But WinVote currently has a significant limitation. No GPU acceleration inside the Windows guest. And that's where w... Or LSW gets interesting.

Rob Campbell [00:05:59]:
Almost said the wrong one. LSW also runs a real Windows 11 installation inside QEMU, KVM. But instead of relying on traditional remote desktop technology, it uses its own integrated system and a, paravirtualized GPU. That allows Windows applications to access your Linux system's graphics hardware without dedicating an entire GPU to Windows. We're talking about the potential to run applications like Photoshop, Affinity, Office three sixty five alongside your normal Linux applications with GPU acceleration. Now there's a catch. So so it's, you know, very similar to how Windows w... Does WSL, you know, because that's actually a Linux virtual machine and it does things very similar.

Rob Campbell [00:06:54]:
I don't know. It's not open source, so I can't see exactly what's doing. But, anyway, there's a catch. W s to l, LSW is still very early in development. It currently targets Wayland, which is good for our feature, with k d plasma as its tested desktop environment does. I mean, it won't work anywhere else. That's where it's just tested, which is good for this panel here. You guys all love KDE.

Rob Campbell [00:07:21]:
You'll also need a Windows 11 installation, so this doesn't eliminate Windows entirely. And compatibility and performance are are still a work in progress. But I think this project represents an interesting direction for Linux desktops. For years, one of the biggest arguments against switching to Linux has been that, that one Windows application you simply just can't live without. LSW might eventually make that argument a little harder to justify. You may still need Windows to run that application, but you might no longer need Windows on your desktop. You know, you can hide that little nasty thing in some environment you don't have to look at.

Jonathan Bennett [00:08:09]:
Yeah. It's, I've played around with... I think VMware had, one of their virtualization solutions that let you sort of do the same thing where it would hide most of Windows and let you just map individual applications onto a Linux machine. And, you know, that was that was cool and all. I eventually just sort of kicked the habit and gave up on Windows altogether. But...

Rob Campbell [00:08:30]:
And I s... Well, yeah. I don't know. I won't... I don't know how the GPU acceleration would have been on that. That probably would have been lacking. Yeah.

Jonathan Bennett [00:08:37]:
That I'm not sure about. It is actually... It's pretty impressive that that this project has gotten GPU acceleration, working in the point it has.

Rob Campbell [00:08:47]:
Yeah. Through your own parallel, paravirtualized technology instead of, using the the traditional Windows app or RDP remote desktop really is what that is.

Jeff Massie [00:08:59]:
Mhmm.

Jonathan Bennett [00:09:00]:
Yeah. That's cool.

Ken McDonald [00:09:02]:
Now does it say how much of the, what, portion of your systems,

Rob Campbell [00:09:13]:
resources?

Ken McDonald [00:09:15]:
Yeah. Memory and, percentage of the CPUs are

Rob Campbell [00:09:19]:
used in VM? You're running Windows. So whatever it would take to run Windows, that's about

Ken McDonald [00:09:25]:
the VM. Gonna grab at

Jonathan Bennett [00:09:27]:
least. Yeah. Yeah. Yeah. It's a Windows VM. Oh, interesting. Alright. Yeah.

Jonathan Bennett [00:09:37]:
Let's let's move on. Jeff, you've got, you... You've got some system d news and, some Rust news all rolled up into one. Handsome

Ken McDonald [00:09:46]:
love for both?

Jonathan Bennett [00:09:47]:
Maybe. Yeah. Maybe.

Jeff Massie [00:09:50]:
Well, you know, people talk about controversy and everybody goes, oh, AI? No. No. No. This is where the really controversy is. Because if you spend any time on Linux forums, you know the two most reliable ways to start a fight. Is say something about system d or you say something about Rust. So naturally, this week, I'm going to talk about both of them. So system d is now officially supports programs written in Rust.

Jeff Massie [00:10:18]:
Now the code landed in late October in a pull request from Christian Bruner titled Rust support program written in Rust. Then on November 3, Bruner and Leonard Pottering published a blog post explaining the rules. The short version, Rust is welcome, but only on a short leash. Here is where Rust is allowed. System d's own coding style documents now say Rust is only used for new leaf components, I e programs that have... Programs that nothing else in the tree depends on. So leaf here means the end of the branch. You know, a standalone program can be written in Rust as long as nothing else depends on it.

Jeff Massie [00:11:02]:
And where is Rust not allowed? Well, again, from the document, you know, like Lib System D shared, lib system d, the NSS, and PAM modules and everything else other components depend on, they stay on c. So basically, the foundation of system d still stays on c. You know, Rust gets the new rooms on the, you know, the foundation stays in c, Rust gets the new rooms on the edge of the house. So why draw the line there? The big reason is shared libraries. System d has more than a 180 separate programs, and they all share one library of common code that only works if programs can call into the library through a stable, predictable interface, you know, called an ABI. C has had that for decades. Rust still doesn't have a stable ABI, and there's a Rust issue about about it where it's been open for more than ten years. So Rust needs to settle down a little.

Jeff Massie [00:12:09]:
If system d put Rust code into the shared library, the Rust parts would have to get copied into the program after program resulting in duplicated Rust code across multiple libraries and executable... Executables. Bruner and Potter Pottering said that's just not practical and static linking a 180 plus programs isn't either. Then the... There's another rule in here. The next rule, you know, might surprise some Rust fans, no cargo. Cargo, you know, for those who don't know, is Rust build tool and package manager, and most rod... Rust projects live and die by it.

Jeff Massie [00:12:45]:
System d builds everything with Mason, and that isn't changing or miss... Nissan?

Jonathan Bennett [00:12:53]:
Something like that. Yeah.

Jeff Massie [00:12:54]:
Mason. Okay. The blog post gives two reasons. Two build systems would make life harder for distribution maintainers and pulling packages in from the cargo ecosystem raises security and supply chain concerns. So it's it's not just, you know, we don't want to. It's like, no. There's some real reasons here. The restrictions do go farther.

Jeff Massie [00:13:17]:
The system d's rust code is no underscore, underscore STD, so no standard. So it doesn't use Rust standard library at all. The coding document says only the core and Lib C crates are permitted. And, you know, from what I'm told, if you've ever written in Rust, that's like being allowed in the kitchen, but only with a butter knife. So you're you're definitely, handicapped a little there. Now the reason is is Rust standards collections abort the program when memory runs out. System d wants every allocation to be to be able to fail gracefully the way the kernel does it. So it it provides its own versions.

Jeff Massie [00:14:03]:
Basically, you can't just kill something that other things might depend on or run. They want graceful exits. There's another rule in the document says a program written in Rust must not look or behave any different than one written in C. So same help text, same option parsing, same error messages. It should be indistinguishable. There's even a test that captures a Rust program and a C program byte for byte. So when it it compares that output and it had better be exact per byte. As a user, you're not supposed to be able to tell which language, tool is written in.

Jeff Massie [00:14:48]:
So that's one of their founding reasons for this rule. They want it to be as seamless as possible. So what's the first Rust program? It's called system d keyring setup and it was merged on October 27. At boot, it loads security certificates into the kernel's key strings for DM dash verify, FS dash verify and BPF and then seals those key rings so nothing can be added later. In other words, the first Rust code in system D is a security component. Now size was a concern too. The poll request says that rust or says a stripped rust test program comes in at 32 kilobits or kilobytes against 20 kilobytes for its C twin. So it's bigger.

Jeff Massie [00:15:34]:
Not by much, but it is bigger. What about, excuse me, rewriting the old C code? Not happening, at least not just for the sake of it. Bruner and Pottering point out that many of system d security risks aren't memory bugs at all. They come from how files and paths are looked up and switching languages doesn't fix that. There's also note that Rust doesn't run on every CPU architecture this system d supports, so Rust stays optional. You can still build system d with no Rust at all. So if you're running some hardware and you're thinking Rust doesn't work on my hardware, this is going to break everything. Nope.

Jeff Massie [00:16:15]:
Nope. Optional. You're you're still okay. There there was already one close call. A change in the Rust compiler would have broken system d's Rust support and it was reverted on the Rust side in late October. The Rust folks reverted it because they said some no underscore standard projects like system b define that symbol themselves, and this would break things in that case. And and I do wanna point out, I don't have it... I didn't really capture it in show notes, and it's kinda in other articles I was researching for this.

Jeff Massie [00:16:50]:
They're not officially... System d and Rust aren't necessarily officially working together, but it seems like there's some communication going on because there are things like that where changes in the Rust compiler or in the Rust language could break what system d is trying to do, and they kind of, are working together on some things even if it's not in a official sense. But, you know, overall, who wins? Well, the Rust fans, they got Rust in system d. The the skeptics got a written promise that the core stays c and there's no cargo and the code goes through more checks than airport security. So the short version, Rust is welcome. It's on a short leash. Take a look at the links in the show notes for the blog posts and other articles talking about this change.

Jonathan Bennett [00:17:39]:
Yeah. This is something that I've I've thought for the longest time with the idea of using Rust as a system language is as cool as cargo is for putting things together quickly, it's not great for systems programming for some of these same reasons. So I, you know, kinda nice to hear somebody say it out loud.

Jeff Massie [00:17:56]:
Yeah. And and, I mean, a system d, you're at the core of every... You know, that's the, man, that's the keystone for a lot of Linux stuff. So it's like, yeah. You better be sure that that's solid. And I I found it interesting when they said, yeah. Memory was not really a big issue with system d. That was not their biggest concern.

Jeff Massie [00:18:13]:
Mhmm. So switching to Rust, you know, doesn't fix it.

Ken McDonald [00:18:17]:
Probably because system d does what Linux does, treats everything as if it's a file.

Jonathan Bennett [00:18:23]:
I mean, there's still gonna be memory stuff going on behind the scenes.

Jeff Massie [00:18:27]:
Yeah.

Jonathan Bennett [00:18:27]:
But it... And I think it's...... That solves that particular problem. Yeah.

Jeff Massie [00:18:31]:
Well, like I said, it's it's the path lookups and things like that where they're they're fighting a lot more than, you know, memory out of bounds and whatnot, read after free stuff.

Rob Campbell [00:18:41]:
The thing with cargo I've always questioned and never never enough to actually look into it. It's never made it past my head to my fingertips. But the repository piece of it, like how safe and trustworthy is that when you cargo some random thing from the repository?

Ken McDonald [00:18:58]:
That's totally safe. You can trust it as much as you trust the Java repository. It's

Jeff Massie [00:19:05]:
just right up there with the AUR?

Rob Campbell [00:19:06]:
I mean......

Jonathan Bennett [00:19:07]:
At NPM?

Rob Campbell [00:19:08]:
Yeah. I... Yeah. I I I understand, you know, all these repositories, the app and Ubuntu's, all of them, there's a risk. But is it more like, you know, Ubuntu's app repository, or is it closer to NPM and and those? I don't know. I've never looked into it, but I've always had a little slight concern about it.

Jonathan Bennett [00:19:34]:
Yeah. I mean, so the question is, can anyone publish to crates.io?

Rob Campbell [00:19:37]:
Really? Yeah.

Jonathan Bennett [00:19:38]:
And I think I think the answer is generally yes. Whereas with something like Ubuntu, there is a process where someone already on the inside of the Ubuntu system, you know, has to sponsor your package and be like, yes. We actually want that. And so it's more of a, it's more of a pull with with some discernment happening as opposed to just a push whatever you want to.

Jeff Massie [00:20:00]:
That's a little too.... Ring of trust or levels of trust.

Ken McDonald [00:20:04]:
Mhmm. Yeah.

Rob Campbell [00:20:04]:
Yeah. Yep. That's somewhat somewhat what I suspected, but, I was too lazy to look at it myself, which is why I asked you guys. Yeah.

Ken McDonald [00:20:12]:
Rob, what I've seen over the years with Ubuntu is if if it's not already in the Ubuntu's repositories, people will put up their own PPA. Mhmm. They... And let you point to it and download it. Even What's a bunch who trust it, then it gets moved into their repository.

Rob Campbell [00:20:34]:
Yeah. I've even made my own, Debian repositories for myself in the long time ago.

Jeff Massie [00:20:40]:
Well, that's why they say don't trust PPAs unless you really trust the person. I've known people who have been... Done packages and, yeah, you you start building trust. So you get on and you maybe help with documentation. You do tiny little fixes or little mismatches. You present it to somebody who then can go over it and has the power to then submit it to the repository. And to get to the point, you can actually submit things to the main repository yourself is a years long process, and you gotta have a lot of continual involvement with the with the community. Debian in

Jonathan Bennett [00:21:19]:
particular is the one where they they have quite the quite the process to become a a a maintainer in the Debian package or the pros. In the Debian ecosystem. I I think it used to be the case, and I think it still is, that to become a Debian maintainer, you had to physically meet another Debian maintainer in person to do an identity verification Yeah. And then get your keys signed. And that was the only way to get into the ecosystem. And so, you know, you were googling your... Or, you know, searching asking people on the mailing list. Like, here's where I'm at.

Jonathan Bennett [00:21:50]:
Who's closest? And you would just go do a road trip to get your your, your PGP key signed.

Rob Campbell [00:21:57]:
Yeah. You know, even, even if the code is all safe, the maintainer will package it or, you know, compile and package it. You never know if, they use the right, you know, hardening, flags to to make that binary secure. But, you know, I'll show you later on in the show, a way to check that.

Jonathan Bennett [00:22:16]:
Oh, cool. Alright. We're actually... Before we get to any of that, we're gonna take a quick break, and, we're gonna come right back with some Blender news, so don't go anywhere. Alright. Ken, what's up with Blender?

Ken McDonald [00:22:31]:
Well, I'm glad you asked, Jonathan, because this week, Bobby Borisov and Marcus Nister wrote about this powerful, free, cross platform, open source. We don't wanna talk about that one yet. But...

Jonathan Bennett [00:22:49]:
We don't talk about Blender?

Jeff Massie [00:22:50]:
I'm sure you cut there. Talk to us as a commercial....

Ken McDonald [00:22:52]:
To the next article. This week, in addition to Bobby Barsoff and Marcus Snyder, we also have Michael Larebel. All three of them wrote while Jonathan Lampell and Wayne Dixon from g g cookie released a twenty eight minutei video talking about everything new in blender five dot three. According to Michael, blender five dot three introduces NVIDIA DLSS four dot five ray reconstruction as a new denoiser and add support for NVIDIA RTX spark devices for those that can afford those. According to Bobby, Blender 5.3 includes native Gaussian splatting support with Vulkan as the default graphics backend on Linux and improvements to the rendering and video editing tools. According to Marcus, Blender 5.3 also introduces a projects feature that lets you group related files, share asset libraries, manage color configurations, and set custom variables across an entire project. Asset libraries are now also distributable as extensions directly from the extensions platform. According to Jonathan, Blender 5.3 adds support to find any property or operator in the sidebar across all editors and supports basic markdown formatting.

Ken McDonald [00:24:29]:
Write markdown directly in the new comment node or in regular layouts for scripts or addons through the label underscore markdown widget. You can now arrange your objects freely with custom sorting in the outliner. As I said earlier, the video is twenty eight minutes long, but we have other news to cover. So follow the show note links to read Bobby, Michael, and Marcus's articles for more details. And, of course, there's a link to the video if you have the time to, watch it, demonstrate almost everything new.

Jonathan Bennett [00:25:04]:
Are there demos of the Gaussian spotting?

Ken McDonald [00:25:08]:
They're going through, actually, showing up pretty much everything.

Jonathan Bennett [00:25:13]:
Gaussian spotting is very cool. That's that's a technique where you take a picture of something from multiple different angles, and then you you feed it into this algorithm, and it builds essentially a three d point cloud and then sort of renders the texture onto each of those points to rebuild the thing that you were looking at.

Ken McDonald [00:25:31]:
Yeah.

Jeff Massie [00:25:32]:
It's a good I

Ken McDonald [00:25:33]:
was watching the video. It was neat seeing it switch to the Gaussian splouting and then back to, the three d image.

Jonathan Bennett [00:25:39]:
Mhmm. Yeah. Yep. It's cool stuff. It's it's, it's neat tech. So, yeah, very cool. If we don't have any more to say about Blender, Rob, you, you you you crashed.

Rob Campbell [00:25:53]:
You crashed.

Jonathan Bennett [00:25:54]:
You crashed?

Rob Campbell [00:25:54]:
Linux is crashing. We're crashing Linux.

Jonathan Bennett [00:25:57]:
Some... Something's crashing.

Rob Campbell [00:25:59]:
Apparently. So, yeah, apparently, you know, Linux has become so difficult to crash, actually. The developers have decided to add a new tool to make crashing easier for everybody. Now I find this a little strange. You know, I I thought not crashing was supposed to be one of Linux's selling points. That's what I tell all my friends. Are we trying to keep up with Windows here, maybe? You know, after all, Windows is still more popular on the desktop. You know, maybe the Linux developers figured we need to offer a few more crashes to attract those Windows users to us.

Rob Campbell [00:26:37]:
That's... I don't know what else they're stuck over there for, so it must be crashing. Here's the real story. Coming to Linux 7.5 is a new tool called hwpoison panic. And, yes, its entire purpose is actually to intentionally crash your Linux system. But there's actually a very good reason for this. Modern computers, especially servers with ECC memory, can detect certain types of memory errors. Sometimes those errors can be corrected.

Rob Campbell [00:27:11]:
Other times, the memory is considered corrupted or what Linux calls hardware poison. When this happens, Linux tries to isolate the damaged memory and continue running. And, you know, that's great when it works, but what happens if that corrupted memory contains something critical to the kernel itself? Well, Linux might detect the error, log in, and continue operating. Unfortunately, that corrupted memory could eventually cause a crash anyway, or worse, silently corrupt your data without anyone realizing there's a problem. And if the system crashes twenty minutes later, figuring out what caused it becomes much more difficult. That's where a kernel option called Panic on Unrecoverable Memory Failure comes in. When enabled, it tells Linux that if certain critical, kernel memory becomes corrupt and, cannot be recovered, the safest thing to do is to immediately trigger a kernel panic. Well, I don't know about that.

Rob Campbell [00:28:15]:
I kinda think the safest thing to do would be to send an alert to administrators first and and let us decide how to take action. I don't know. Maybe we're bad at deciding. Maybe, you know, but maybe we wanted to gracefully fail it over to a different server or something before a crash hits. I don't know. Anyway, I get they have a lot of good configurable options, they could add to this. But their thought is that this instant crash gives administrators a better chance of capturing, diagnostic information while the original error is still identifiable. And in high availability environments, it can allow another system to take over rather than letting a damaged server continue to operate, which is fair, you know, when works quickly.

Rob Campbell [00:29:03]:
Could also consider adding some kind of call to trigger the rollover right before it crashes then, you know, rather than waiting for the quorum to detect the system is down after a crash. I guess it's a start. The new h d... H w poison dash panic tool lets kernel developers deliberately deliberately simulate these memory failures to verify that everything works correctly. To run it manually, you actually have to supply an argument called, dash dash yes dash panic dash my my dash kernel. So, yes, panic my kernel. You know, at least they're making sure you know what you're getting into. Now now this isn't something that will suddenly start crashing everybody's, you know, Linux computers, you know, at least not until hackers find the vulnerability in in the call and and do whatever they do.

Rob Campbell [00:30:02]:
The panic behavior is optional and this testing tool is really intended for controlled environments, particularly virtual machines, I guess. Sometimes, the only thing worse than a computer that crashes is one that keeps running while quietly destroying your data.

Jonathan Bennett [00:30:24]:
Yeah. So, you know, what comes to mind with this sort of thing is not just ECC. I wonder if this sort of panic would also catch, like, row hammer attacks where someone, is intentionally going on there and trying to get bits to flip.

Rob Campbell [00:30:37]:
Bits to flip. Mhmm.

Jonathan Bennett [00:30:40]:
But the reason the reason why panicking is probably almost always the right call is, when your when your memory is changed, like, when your when your RAM bits are changing out from underneath you, it's like the... You don't know anything. Like, the the very ground that the system is standing upon is is no longer solid. And so there's no... Like, you've got a lever, but you've got no, no safe place to stand on and try to use it. So I I think, I think just the immediate panic probably is the right thing, because anything else, you... You're gonna use those bad bits, and then who knows what what you're gonna jump off to and do.

Rob Campbell [00:31:14]:
Yeah. Panicking is always the right call,

Ken McDonald [00:31:17]:
especially if you're in quicksand. Yeah. I

Rob Campbell [00:31:22]:
don't think so. It doesn't really work like the cartoon. Sorry, Ken.

Jonathan Bennett [00:31:27]:
It's true. We all grew up being morbidly terrified of quicksand and

Rob Campbell [00:31:32]:
thought I'd play a much more important role in my life.

Jeff Massie [00:31:34]:
Yeah. Yeah. I did too. Well,

Ken McDonald [00:31:38]:
so the description Jonathan gave made it sound like quicksand.

Jonathan Bennett [00:31:43]:
Not exactly.

Rob Campbell [00:31:45]:
Alright. Is this a a segue to something?

Jonathan Bennett [00:31:47]:
No. It's not. I was talking about the ground not being solid under your feet. So

Rob Campbell [00:31:51]:
ground. Oh, that... Alright.

Jonathan Bennett [00:31:53]:
Yep. Sounds like quicksand. Okay. Well, anyway, I I guess it's time for another break. And, then Jeff is gonna come back and save us from the bad puns, talk about some banjaro. So don't go anywhere, and we'll be right back.

Jeff Massie [00:32:09]:
Most forum announcements start with, hi, everyone. This one started with, dear friends, sympathizers, relatives, in laws, naysayers, foes, detractors, tax collectors, YouTube talking heads, Reddit trolls, and other potentially sentient beings monitoring this forum. Now that greeting comes from Frank VanderMeen, known as in the... On the Monjaro forums as Aragorn. And on November 13, he announced that the Monjaro project is now an officially registered and fully legal nonprofit organization in Germany, completely independent from the commercial Manjaro, GmbH, and company KG. In the past, Monjoro was one of the most popular arch based distributions. You know, for quite a while, it got a lot of attention there. And it started as as a community project.

Jeff Massie [00:33:08]:
And in 2019, founder Philip Mueller set up a company, the Manjaro GmbH, to run it. The company would start, you know, could sign contracts, hold money, pay developers. And for a while, it looked like that was a pretty good working model. But this year, the volunteers saw it differently. In March, the Manjaro team published a document called the Manjaro 2.0 Manifesto and did not pull any punches. It said the project has been declining over the past decade and that it had lost almost all of its contributors and even became a laughingstock for repeatedly making the same mistakes. Like one example it gave was the TLS certificates not being renewed in time. We covered this a few months ago on the show, so if you feel like you've heard some of this before, you have this little bit of a recap.

Jeff Massie [00:34:03]:
The central complaint was about control. The manifesto said Manjaro was being run as one individual's personal project centered on someone who refuses to share essential access to the project and its infrastructure. It's also said the project money had run out causing Manjaro's only full time developer to lose their only source of income. The proposal was to split the community project off into a German nonprofit association. The new association would take over the GitHub organization, the GitLab server, the manjaro.org domain, the forum, and other community donations. The company could keep doing its commercial work downstream of the community project. The team also said, what would happen if they were ignored? Stage one was a gentle strike where the moderators stopped approving new forum members and stopped, you know, handling flagged posts. That was kind of the polite version of the strike.

Jeff Massie [00:35:07]:
Stage two was considering a a complete fork. Developer Dennis Hoove tried to calm things down and he wrote, the goal is not to remove anyone from the project. He wrote, we value Philip. He's a genius in his field, but his lack of communication skills and the inability to share anything is destroying the project. So how did Mueller respond? In his own words, it seemed a bit like the mutiny on the bounty, but he didn't say no. He wrote that he wasn't against an association in 2019 and I'm not against it now either. He... But he did set a condition.

Jeff Massie [00:35:47]:
Build the thing first. He's... Mueller said the ball's in your court, he wrote. Go ahead and set the new entity up and then get in touch with me. Otherwise, business as usual. Well, eight months later, the volunteers have done just that. The association is registered in Berlin with seven founding members from five countries, The Netherlands, Belgium, Germany, The United States, and Switzerland. Two of them, Mark Waggie and Roman Gig...

Jeff Massie [00:36:15]:
Gilg, g I l g, signed the original manifesto and Gilg did so while serving as the company's CTO. And there's another detail in the announcement. Van der Muren noted that the most recent testing and stable updates were put out by the Monjero team without supervision from Philip Mueller. His point was the team can keep the distribution running on its own. He also mentioned they're rebuilding parts of the infrastructure partly because of what he called ongoing bot attacks. Now one important caveat, this is not a done deal. In a follow-up post, Dennis Hu wrote a transfer of assets to the EV, basically the nonprofit, has not yet taken place. The next steps are an opening...

Jeff Massie [00:37:03]:
Are opening a bank account, inviting more members and cleaning up the bylaws. Only after that will they quote reach out and engage in a dialogue with the company, meaning Mueller. So right now, the association is seven people and a notary stamp. The domain, the repositories, the trademark, and the domains are all still owned by the company. Now if you run Mon Jarrow, nothing changes today. So, you know, updates are still coming and there's no new release. But this is the first time in years that Manjaro's future has had a real fork in the road that isn't an actual fork. Either the company and the association negotiate a handover or the volunteers who actually ship the updates end up with the people and the company keeps the name.

Jeff Massie [00:37:55]:
Take a look at the links in the show notes for full details.

Jonathan Bennett [00:38:01]:
Interesting. So I I... On on on one hand, like, sounds like things were coming to a head and something had to change. On the other hand, this this is not necessarily the sort of thing that you wanna end up in the news for. And if I were running Minjaro, I would have some serious considerations about whether that was continuing to be a good idea because this this gets a little worrying.

Jeff Massie [00:38:30]:
Yeah. And we did cover his... Basically, the biggest problem with Mueller is nobody can get a hold of him. He just doesn't reply. He doesn't respond. He does Yeah. But he holds and gives everything.

Rob Campbell [00:38:40]:
And Jaro's been kinda known to be a troubled distro for some time now.

Jonathan Bennett [00:38:48]:
Yeah. Yeah.

Jeff Massie [00:38:50]:
I mean, it used to be so popular and then it really fell off. And and and they're trying to bring it back to its glory by saying, hey. We're gonna have a committee. We're gonna run it. And, you know, if if Mueller wants to keep the corporate stuff, yeah, okay.

Rob Campbell [00:39:04]:
I think when I tried, it was right about, like, at its peak right before it went over the cliff.

Jonathan Bennett [00:39:10]:
Over the cliff. Yeah. We'll we'll have to continue to watch this. Sometimes sometimes these restructurings work and they work well, and there's a path from the other side of them. And then I've seen other times that you you try this and even the best of intentions, and it's it just... It never works. It never gels under the new, under the new structure, under the new new leadership and ideals. Just give me trying something.

Rob Campbell [00:39:34]:
I I

Jonathan Bennett [00:39:34]:
know you That's fair. That's fair.

Jeff Massie [00:39:36]:
I think it'll work because they're the ones doing the work and keeping it running now. Mueller Mueller is not doing it. He's... He just owns it, but he's kinda in the background. And it sounds like from what they... Everything to transfer, you know, if they get it completely set up Mhmm. It's gonna go.

Rob Campbell [00:39:54]:
The the problem or concern is that it's not working all that well now either. So is this going to be enough to make that difference?

Jeff Massie [00:40:02]:
Well... And Yeah. That's where they'll just probably fork it. If if he's really thrown a fit, they'll say, alright. We're gonna name it something else and take off.

Rob Campbell [00:40:13]:
Yeah. I mean, I guess, really, the question is is the problem having him have in the pitcher, will have... Will getting him out of the pitcher actually restore it? Can can they restore it?

Jonathan Bennett [00:40:24]:
Is this is this actually a path to fixing things?

Jeff Massie [00:40:27]:
I I think it is. Because it... Because... So

Rob Campbell [00:40:31]:
do you...

Jeff Massie [00:40:31]:
When we did the story...... Like... Yeah. When we did the story last time... Or I think I I... Actually, I was the one that did it last time a few months ago or several months ago. Yeah. It's it's just that he's not renewing leases, you know, the the certificates.

Jeff Massie [00:40:46]:
He's not

Rob Campbell [00:40:47]:
Mhmm. Yeah. Yeah.

Jeff Massie [00:40:47]:
Right. Answering mail. He's not doing any Okay.

Rob Campbell [00:40:49]:
I mean, there's that, but, I mean, the problems have actually been with the distro itself. It... Has he been holding them back from making it... Keeping it in its glory?

Jeff Massie [00:41:03]:
That's kinda what it sounds like. Yeah. Just he's blocking so much stuff that they... It... You're trying to circumvent him to get things done.

Jonathan Bennett [00:41:11]:
Yeah. Yeah. It... It's it's amazing how many things become much more difficult when you don't have the keys and when your TLS certificates go up. And and you gotta spend so much time. It's kinda like when instead of doing your regular work, you are troubleshooting the Microsoft downtime, how you didn't get anything done this week because

Ken McDonald [00:41:30]:
of that.

Rob Campbell [00:41:30]:
It's like when you have a problem to fix that you don't have access to.

Jeff Massie [00:41:34]:
Yes. Well, and he's he's got, like, a few different

Ken McDonald [00:41:38]:
raise funds for something else.

Jonathan Bennett [00:41:41]:
We're just doing the work.

Jeff Massie [00:41:43]:
Yeah. He... And he... You know, he's got the GitHub as well, the GitLab, the... There's a lot of that stuff that Mueller does control that you're you're kinda having to go around him. Mhmm.

Jonathan Bennett [00:41:58]:
Yep. Yep. Absolutely. Alright. Let's, yeah, Ken.

Ken McDonald [00:42:04]:
Just gonna say, let's hope they can come up with some way to raise funds similar to the way Ardor does.

Jonathan Bennett [00:42:11]:
Ah, I I see I see what you did there. Alright. So how does Ardor raise funds, and what, what have they done with that recently?

Ken McDonald [00:42:19]:
Well, you know better than I do how they raise funds since I think you actually support them. But what's nice is what they have recently come out with. And we can thank both Bobby Barsoff and Marcus Nestor since they wrote about this powerful free plus platform open open source digital audio workstation for GNU Linux, macOS and Windows now being updated to Ardor nine dot 10. That's... This is gonna be its latest stable version. Now, what immediately caught my eye is one of the plugin improvements. Linux users get fixes for v s t three plugin interface resizing under, you ready for this, Rob? X 11. So, yes, there is still support somewhere for x 11.

Ken McDonald [00:43:14]:
Now, Ardor now sends corrected parameter values and time stamps at the end of processing blocks addressing compatibility issues affecting some VST plugins including those from fabfilter. According to Marcus, Ardor 9.10 release introduces optional support for audio waveforms that reflect the region's gain envelope. So the waveform of an audio region updates as you drag the gain envelope nodes around. It also adds an option to move the regions gain envelope along with the contents of audio regions when slipping. According to Bobby, midi editing has received several usability fixes. Users can make rubber band selections of velocity markers, drag them in draw mode, and add to existing note selections using ctrl plus drag. Now the middle click dragging is supported for several control points while shift plus right click deletes notes and points. The midi editor also gains missing context menu actions including quantize, legatize, transform, and remove overlap.

Ken McDonald [00:44:34]:
There is also a new wall clock transport master that syncs our door's own generated LTC or MTC to the system clock so that you can run or record a show time stamped with the current time of day. I also noticed evidence that the release notes were entirely human created. Now as always, more details can be found in Bobby and Marcus's articles. I have provided links in our show notes.

Jonathan Bennett [00:45:04]:
Yeah. You know, I I have come up with kind of a... Just a riff off of that that the the shown... The, the release notes were human created. I kinda have a, a policy on that. If I think real humans are gonna take the time to read something I write, I'm gonna write it by hand. But, you know, if it's a poll request and you're using your clinker to write the poll request and respond to the criticism, then I'm just gonna turn Claude loose on and say, go go find all the things and tell them to fix it. And, in fact, I I hunted somebody down on Discord just the other day, and I'm like, hey.

Jonathan Bennett [00:45:36]:
Is this actually you doing this, or are you using the LLM tooling? And he's like, I'm using LLM tooling. Oh, okay. Fine. I'm gonna go tell the call to nitpick if we got stuff fixed.

Rob Campbell [00:45:45]:
I like, I think release notes are perfectly, great use case for it. I mean, I don't know. It it it it get... If if it's something monotonous and time draining, you know

Ken McDonald [00:45:59]:
prone to typos?

Rob Campbell [00:46:01]:
Sure. Whatever. But, you know, they could be spending more time coding.

Jonathan Bennett [00:46:06]:
Yeah. Well, so, I mean, there's there's then a hybrid workflow that you can do where you say, hey. You know, here's the here's the the old commit. Here's the new commit. Pull all of the all of the commits that we made in this. Just give me the ones that you think are important, and, then you rewrite it. Or, you know, you you distill it down to something that's more palatable to a human.

Ken McDonald [00:46:23]:
Intentionally insert one typo somewhere.

Jonathan Bennett [00:46:26]:
Insert a typo somewhere just so people know that, oh, a human wrote this.

Rob Campbell [00:46:32]:
Oh,

Jonathan Bennett [00:46:32]:
that's terrible. Shit. But, yes, the things we do. Yeah. I need to I need to go and pull our door, actually. I was... I I used it for the first time in a while because we finally had a Flos Weekly episode this past week. But, anyway...

Jonathan Bennett [00:46:47]:
Yeah. Yeah. Yeah. So let's take a, let's take one more break, and then we've got, we've got some more news to get to. Rob's gonna talk about Wayland, one of his favorite things to talk about. It's gonna be great. Don't go anywhere. I'll be right back.

Rob Campbell [00:47:02]:
Alright. So of all the Linux desktop environments I expected to make expected to make the jump to Whelan, mate was probably near the bottom of my list. And I don't mean that as an insult. I've I've always appreciated what mate represents. So when I first tried it, I thought it was great back in the day. You know, when GNOME moved away from its traditional desktop, you know, with GNOME three, mate carried on the GNOME two experience. It's it's never really been about chasing the latest and greatest technology with them. And frankly, you know, I was under the impression that mate was barely being maintained these days.

Rob Campbell [00:47:43]:
You know, after all, earlier this year, a long long time Ubuntu mate leader, Martin Wypress, stepped down, and the distribution even skipped its twenty six point zero four release. But to be fair, a new team stepped up and Ubuntu mate returned with version 26 dot 10 in October. So I I actually wasn't even aware of that. Didn't catch that. You know? So maybe there's more life in this old desktop that I gave it credit for. Now, I was going to say MATTE was beating other traditional desktops to Wayland, but after doing some digging, you know, that's not entirely accurate either. So ranting, raving, you know, maybe I'm making some corrections. I'm on assumptions here today.

Rob Campbell [00:48:28]:
It's a cinnamon, another GNOME derivative desktop, that took a long time to get to Wayland, has actually made significant progress. Its new 6.8 release, that just came out, considers Wayland fully supported rather than experimental. Even XFC has offered, experimental Wayland sessions since version four dot twenty, and, developers are working on their own composite compositor there. Trinity, if anyone remembers the old KDE three fork, is still primarily living in X11 World. And you know what? That's probably gonna be one of the last hangouts... Holdouts. I I don't think, I don't think they're going anywhere. But, here's where MaTe gets interesting.

Rob Campbell [00:49:17]:
A developer named Kareem Soke has, released, version one point o of polo. I know most people think polo is not, you know, it's a shirt you wear, but here, it's a new Whelan compositor designed specifically for, mate. I should have wore a polo shirt today, but, I'm not. Anyway, now, mate, yeah, I guess I didn't know this. Maybe I knew and I forgot because I haven't really paid attention to mate in quite some time. But, mate, I guess they already had an experimental, Wayland support using Wayfire. But just a caveat, you know, that required adapting mate to work with a compositor that, wasn't originally designed, for it. Polo takes a different approach.

Rob Campbell [00:50:09]:
It uses West... The Weston compositor library while recreating the behavior of, Marco, mate's traditional window manager. That means preserving familiar windows management, keyboard shortcuts, themes, workspaces, and desktop settings. It even supports session management, multiple monitors, screen sharing, and older X11 applications through X Wayland. And, you know, for those of you feeling nostalgic, the developer even brought back desktop effects like wobbly windows and the desktop cube. Anybody remember those? Actually, I think Ken's been using a cube on something recently. But... Yeah.

Jonathan Bennett [00:50:54]:
KDE can do it. They brought it back to KDE too.

Rob Campbell [00:50:57]:
Yeah. It's... You know, the old is new again. But now with x, x is still old. Anyway, now that's a feature I wouldn't mind seeing, make a comeback. You know, the wobbly windows, it's fun. Of course, polo isn't, ready for everyone. It's a independent project, still needs more testing, and parts of Mate haven't been fully converted to Wayland.

Rob Campbell [00:51:24]:
You'll also likely be compiling software yourself, if you want to try it today. But I think there's something encouraging about this. As, Linux continues to move towards Wayland, we don't necessarily have to abandon the desktop environments that made Linux great in the first place. You just have to wait a little while for them to, you know, catch up, come back. And sometimes moving forward doesn't mean changing everything. Sometimes it just means making sure the old favorites could come along for the ride. Ain't that right, Ken?

Jonathan Bennett [00:51:56]:
Yeah. So it's, it's an unofficial port. Ah, gotcha. Interesting. So I guess time will tell if the if the folks at Mate are interested enough in it to, pull it forwards and make it the official fork.

Rob Campbell [00:52:14]:
Yeah. Hopefully, they want to move forward somehow. But

Jonathan Bennett [00:52:18]:
Yeah. That's a good

Rob Campbell [00:52:20]:
We could fork mate in

Ken McDonald [00:52:23]:
Though, I think the three d effect that KDE does is slightly different from that cube because it has three sides. Well, I guess it depends on how many virtual desktops you have.

Jeff Massie [00:52:35]:
Well, a cube wouldn't have three sides. Not to be pedantic.

Ken McDonald [00:52:40]:
Just Well, I've got three virtual desktop set up on here. So that if I go to here... Let's see. It's more

Jonathan Bennett [00:52:50]:
of a triangular prism.

Rob Campbell [00:52:52]:
Yeah. That's the problem, Ken. You don't have enough desktops. I don't know how you get anything done.

Jeff Massie [00:52:57]:
He's got scripts.

Rob Campbell [00:53:00]:
Who needs a desktop when you have the

Jonathan Bennett [00:53:02]:
You've got scripts. That's correct.

Rob Campbell [00:53:07]:
Alright. Alright, Ken. I don't think we have all day to wait for you to show us your cube. Move

Ken McDonald [00:53:12]:
on. It's still...

Jeff Massie [00:53:16]:
No. I do I do remember the cube being cool back, you know, quite a while ago when you're running comp is, and you could have a video plan on a couple of them, and you're rotating things, and the video would be playing while you're spinning the desktops around and the windows around. Yeah. That was that was pretty cool.

Jonathan Bennett [00:53:37]:
Yep. There you go.

Jeff Massie [00:53:39]:
Yeah. It's like a triangle. That's not a cube, Ken.

Rob Campbell [00:53:42]:
Does the old cube have a top and a bottom?

Jonathan Bennett [00:53:45]:
The compass, you could set it up with a top and bottom, I think. Yeah.

Jeff Massie [00:53:48]:
I can't remember.

Rob Campbell [00:53:48]:
It's been so long.

Jeff Massie [00:53:50]:
Yeah. If you had if you had all six desktops for... Or four, you'd have a, you know, two dimensional

Ken McDonald [00:53:57]:
square. Like I was saying, I've only got three virtual desktops set up, or four spaces. So that's all it's, giving me is that triangular setup.

Rob Campbell [00:54:08]:
Like I was saying, you need to add some more desktops for better productivity.

Jonathan Bennett [00:54:14]:
I I

Ken McDonald [00:54:14]:
find three is adequate.

Jonathan Bennett [00:54:18]:
Three is adequate. Well, there you go. Alright. Let's move on. And, oh, Jeff, what is X RAID? I get excited about a new RAID. What do we got here?

Jeff Massie [00:54:30]:
I I know that's kind of been my thing lately. Right? I seem to be talking about RAID quite

Jonathan Bennett [00:54:34]:
a bit. He's the big data guy.

Jeff Massie [00:54:37]:
Chris... Christopher Helway gave a talk in August with a title that doesn't leave much room for interpretation. And it was it was titled RAID sucks and what we can do about it. So not much, leeway in that one. On November 13, we got the what can we do about it part. So Hellwig announced the first public release of x raid, which is described as declustered parity RAID that is integrated into XFS. Now, Hellwig is one of the most pro... Prolific contributors to the Linux kernel and, you know, he...

Jeff Massie [00:55:14]:
He's one of them that does also does maintenance on the NVMe driver and, you know, he... He's... He cares enough clout behind his name. When he said RAID sucks, people listen. So first, let's look at, you know, what he thinks is wrong with it. Problem one, the right hole. So if you take RAID five, you've got say four drives, data is split into chunks across three of them and a fourth chunk holds parity. If any one drive fails, you can rebuild its contents from the others.

Jeff Massie [00:55:45]:
The catch is that changing data means updating two things, the data chunk and the parity chunk. Those are rights on two different drives and they don't happen exactly at the same instant. If power goes out between them, the data and the parity no longer agree. So the Linux kernel's documentation describes it this way. After a dirty shutdown, parity of a particular stripe may become inconsistent with the data on the other member disks. On its own, that's fixable, but if a drive then fails while the parity is wrong, the rebuild uses the bad parity and produces garbage silently. The kernel doc... Documentations call this the RAID five right hole, and it's been known for decades.

Jeff Massie [00:56:33]:
This this is not anything new. Now MD RAID has fixes. You know, when we talk about MDADM, that's part of that. You can add a journal device or use something called a partial parity log, but those cost you. The Linux kernel documentation says that with the partial parity log, write performance is reduced by up to 30 to 40%. So you can have it safe or you can have it fast, but not both. Now problem two, SMR hard drives. Now in case you don't know, SMR stands for shingled magnetic recording.

Jeff Massie [00:57:13]:
Now in a normal drive, data tracks sit side by side with little gaps between them and SMR drive overlaps the tracks like shingles on a roof, which means more data on one platter. That was the whole reason they did this. The catch is that riding one track partially overwrites the track next to it, so you can't change a sector in the middle. The drive is divided into zones and inside a zone you have to write in order front to back, just like you're writing on tape. Reads are still fast and random, but writes have to be sequential. Now traditional RAID hates this because RAID updates small pieces of data and the parity all over the place. So many people found this out originally in 2020 when some Western Digital Red NAS drives turned out to be SMR without saying so on the label. In Serve the Homes testing, three regular drives finished a ZFS rebuild in under seventeen hours.

Jeff Massie [00:58:14]:
The SMR drive took nearly two hundred and thirty hours. That's almost ten days with your array one failure away from losing everything. You know, my personal opinion, a rebuild that requires a calendar is not a good rebuild. SMR isn't going away. Now Helwig's talks, abstract says zone storage makes up an ever increasing share of ship storage capacity, and he says no existing solution handles its sequential right rule. So then we're kind of come back and around say, so well, what is x-ray? X-ray is built into XFS itself on top of zoned XFS. A mode of XFS designed for these write in order zones. WIG says X RAID's two main goals are to support parity RAID zone devices like SMR hard drives and to provide a fast RAID implementation without inherent data integrity problems.

Jeff Massie [00:59:16]:
The trick is splitting the data into in two. Zoned f f x s already keeps your data file apart from the file systems metadata, so file data is written front to back in the zones like tape and it's never overwritten in place. X-ray uses parity raid only for that file data. The metadata is different. It's made of small pieces that get updated constantly, which is exactly what parity raid is bad at. Helwig sends that to a plain mirror instead using Linux's existing MD raid, which he calls more suitable than parity raid for small objects that are updated in place. Now think back to the right hole. It happens when an existing stripe is being overwritten and the power dies halfway.

Jeff Massie [01:00:02]:
X RAID's parity data is never overwritten and the data that does get overwritten sits on a mirror with no parity to fall out of sync. It also works on regular hard drives and SSDs. Helwig says SSD speed hasn't been a focus yet, but even without that, rights are up to 20 times faster than MD RAID. That's his claim and the test details haven't been published in anything I could find. So it's promising but not proven. Can you use it? Not for anything important. It supports zero to zero to two parity drives, which is roughly like RAID zero, RAID five, and RAID six territory, and they're more planned, you know, more RAID levels planned. But...

Jeff Massie [01:00:51]:
And there's a new x-ray management tool called XFS Progs. But Helwig is cleared to say that after a year and a half of work, we aren't anywhere near anything ready for an upstream submission. What he wants right now is testers. He wants people just to hammer it with data you can afford to lose and, you know, find find those issues and bugs that are obviously still in... Inside the, X RAID. Take a look at the show notes for full details and further further links for more information.

Jonathan Bennett [01:01:29]:
That is actually really interesting because RAID and DM RAID really are kind of terrible. They've got some problems, and they are really slow, especially when you're on old spinning Rust disks. It's true.

Jeff Massie [01:01:44]:
Oh, yeah.

Ken McDonald [01:01:44]:
Old spinning Rust disk.

Jonathan Bennett [01:01:46]:
That makes it even worse. Stacking problems.

Jeff Massie [01:01:52]:
Yeah. And and I... You know, and this... There were there were some discussions about... Well, what about ZFS is... Handles this? And I actually looked all the details up and it doesn't. Now there is some where it does like the right hole ZFS does take care of that. But declustered RAID, which is core in X RAID, is an option in ZFS.

Jeff Massie [01:02:24]:
You you can use it, but it's not default. And there's stuff like the SMR, which X RAID handles, but ZFS doesn't.

Jonathan Bennett [01:02:34]:
Mhmm. And there's also the whole the whole problem with ZFS that it can't be upstream in the kernel, and therefore, you're sort of in this weird on your own territory to run it.

Jeff Massie [01:02:44]:
That that is true as well. So, I mean, there's... If people said, wow, ZFS, they're not... There are differences between the two, maybe not enough that somebody worries about it, but they don't fully overlap. And like I said, as SMR drives are more and more increased, it turns into kind of a nightmare.

Rob Campbell [01:03:06]:
I mean, even even butter f s does RAID, but it's probably less... Even less feature, rich.

Jeff Massie [01:03:13]:
Yeah. And and butter f s raid five and six is still a... Don't go there.

Jonathan Bennett [01:03:18]:
Yeah. Do you guys know where XFS came from? I couldn't remember, so I looked it up.

Rob Campbell [01:03:25]:
Germany. I know.

Jeff Massie [01:03:27]:
I don't know. I know Red Hat uses it.

Jonathan Bennett [01:03:29]:
Red Hat uses it. 1993 STI in their old IRIX systems.

Jeff Massie [01:03:34]:
Wow. Oh, Silicon Graphics.

Jonathan Bennett [01:03:37]:
Yeah. Oh, there

Ken McDonald [01:03:40]:
you go. New system.

Jonathan Bennett [01:03:41]:
Yeah. Alright. Let's see. What's next? LibriBoot. Oh, Ken, what's up with, with LibriBoot? I've tried this. I think I've told you guys my story about almost breaking my wife's computer with trying LibriBoot, Or or was that, coreboot? One of those one of those UFI replacements. Yeah. I

Ken McDonald [01:03:59]:
Well, this is just a distribution of coreboot.

Jonathan Bennett [01:04:04]:
Gee, that's a little more than just a distribution of it.

Ken McDonald [01:04:08]:
Well, let me give you a little more background.

Jeff Massie [01:04:10]:
Take it away.

Ken McDonald [01:04:11]:
Well, thankfully, this week, Bobby Borisov and Marcus Nestor also wrote about a hefty update to LibriBoot '26 dot 11. And as I said, I wanted to give you a little bit of a background. Libreboot replaces proprietary BIOS or UEFI firmware offering users greater security and configurability like when using Linux instead of Windows. You become the true owner of your computer. LibriBoot is a coreboot distro that patches and builds upstream sources in many configurations much like Linux distro package managers. This along with regular stable releases serves LibriBoot's main goal make coreboot easy to use. LibertyBoot 26 dot 11 transitions to full UEFI support through the micro... Mister Chromebox edk two payload.

Ken McDonald [01:05:16]:
And nearly all supported boards now ship exclusively with edk2 firmware. According to Bobby, the project removed GRUB as a coreboot payload along with the previously included memtest86 plus and csm wrap payloads. Simplifying firmware configuration and reducing maintenance. According to Marcus, the libreboot 26 dot 11 release enables TPM two dot o by default on Topten x two e n one fifty, the Supermicro x 11 SSH f, ThinkPad X280, the ThinkPad X270, and the ThinkPad T480 or T480S and ThinkPad five eighty devices. Now it also enables EDK2 by default in Coreboot's MK helper and adds a new ROM naming scheme. Now this release simplifies several aspects of LibertyBoot's build system including firmware payload configuration and source handling. Custom firmware images now use clearer BIOS and UEFI prefixes. Improvements to the LBMK build utility.

Ken McDonald [01:06:49]:
Making it easier to insert additional EFI payloads. And work with older firmware images. Leroux, Librebooter's founder and lead developer also owns MiniFree Ltd which funds Libreboot. MiniFree's new business model is to provide only paid Libreboot installation send in service going forward. Lee says this enables a wider variety of hardware to be worked on while greatly reducing administrative overhead by no longer managing stock. This frees up more time to focus on LibriBoot itself. And, always, you can get more details from Bobby's and Marcus articles that I do have linked in the show notes.

Jonathan Bennett [01:07:40]:
Yeah. I I do I do take one one bit of umbrage with the Libre boot versus coreboot. And even the the whole, the... Is it the free software foundation that does the whole respects your freedom pieces of hardware? They they they draw this weird line in the sand that if you have to touch the binary blob to get the thing to work, then it doesn't respect your freedom. But if the binary blob is baked into the piece of hardware where you can't get at it, then it's good to go and it respects your freedom. And I always just thought that was the weirdest the weirdest this this... Discernment to make. Like, that's...

Jonathan Bennett [01:08:19]:
It it... Did... It never it never clicked for me. I never I never appreciated that argument. So you will never... You'll never find me running any of the Libre, kernels or distros or or any of these. I think, I think it kinda goes to a silly place.

Jeff Massie [01:08:34]:
If the binary was compiled on a Tuesday, no. But Wednesday's okay.

Jonathan Bennett [01:08:41]:
Yes.

Jeff Massie [01:08:41]:
It just seems the printer won't It seems very arbitrary. Yep.

Jonathan Bennett [01:08:44]:
But the printer won't print on Tuesdays.

Jeff Massie [01:08:47]:
Yeah.

Jonathan Bennett [01:08:47]:
That's the downside. This is one thing you have to live with. Alright.

Ken McDonald [01:08:52]:
It was interesting. Some of the hardware that this is actually capable of being installed on.

Jonathan Bennett [01:08:59]:
What was the what was the newest piece of hardware? Was it... Did they have pretty recent stuff, or was it all pretty old?

Ken McDonald [01:09:08]:
There was actually some Dell hardware, when you look over at the, mini free, website.

Jonathan Bennett [01:09:17]:
Yeah. It was Coreboot at the time. So years ago, I I teased this. This is where I almost bricked my wife's computer. It was, I was trying to get a coreboot install running on my desktop, actually, because I I don't remember which generation of AMD chip it was. It may have been, like, bulldozer when bulldozer first came out, which... That's a long time ago now. But I was excited, and I found one motherboard that, you know, coreboot advertised support for it.

Jonathan Bennett [01:09:46]:
I'm like, I'm gonna buy that one motherboard. And, of course, the thing would not boot after I after I flashed the, the the ROM chip. But then it's like, how how do I make my computer work again? I don't I don't have a replacement ROM chip. I may not have thought this through. And, I discovered that my wife's computer had the exact same ROM slot on it. And so I had hers turned off, yanked her BIOS chip out, put mine in there, and ran the commands to put my BIOS back on it, and then swapped them back, and everything kind of worked.

Rob Campbell [01:10:18]:
Oh, wow.

Jonathan Bennett [01:10:18]:
And I was a very lucky man that day because I could have bricked two computers.

Rob Campbell [01:10:24]:
So hold on. So you you pulled the chip out. Like, the one wouldn't work because it was bricked. You pulled that chip up, put it on the other one. How did the other one boot? Or what

Jonathan Bennett [01:10:36]:
I did it... I I did the chip swap while it was booted.

Rob Campbell [01:10:39]:
Oh, jeez. Live.

Jonathan Bennett [01:10:40]:
That Yes.

Rob Campbell [01:10:41]:
I... That's what I thought you're getting at, but I wanted to make sure because that's crazy.

Jonathan Bennett [01:10:46]:
Yes. It is. Oh, no. But it worked.

Rob Campbell [01:10:48]:
I would've I would've had it.

Ken McDonald [01:10:50]:
The the one thing that the... They do rec... Do require is that it not have an NVIDIA card in it.

Rob Campbell [01:11:02]:
I require that of all my computers.

Jeff Massie [01:11:04]:
So I...

Jonathan Bennett [01:11:04]:
No. I didn't have an NVIDIA card in mine either. It's just... It... It's just that at the time, Corbit was very optimistic about what computers that they could run on. I haven't jumped on their, what was it at the time? I think it was their IRC channel at the time. I jumped on their IRC and, you know, started haranguing some guy in there. And he's like, yeah.

Jonathan Bennett [01:11:22]:
Well, I don't know if that's actually ever been tested. I'm like, oh, okay. It it it may have been a, it may have been a RAM training problem. Right? Like, that's what gets us a lot of times. There's some secret sauce inside the official BIOS that makes makes RAM work, but, yeah, I never I never could get it working.

Ken McDonald [01:11:42]:
What I found interesting is that it can, also be used on, some ARM based motherboards.

Jonathan Bennett [01:11:48]:
Mhmm. Yeah. Absolutely. Because we... Do we now have UEFI on ARM, which makes those ARM machines way more usable, than having to fight with U Boot. And that's a whole different topic, but that is also a painful experience.

Jeff Massie [01:12:04]:
At least at least a few I looked up, there's talking about machines that are probably 15 years old That... You know, circa something....

Jonathan Bennett [01:12:12]:
About right now. Twenty twenty twenty. I could probably get that old bulldozer laptop... That'll bulldozer motherboard to work now.

Jeff Massie [01:12:20]:
That... And, you know, I'll be honest. That's the one thing that kept me away from a lot of those core brute and Libre boots and two is just the age of the hardware. You know? Mhmm.

Jonathan Bennett [01:12:30]:
The the... It's gotta be soon enough for the thing to be usable or else there's not really any point to it. Yeah.

Ken McDonald [01:12:37]:
Would, full... True full disk encryption be one of those?

Jonathan Bennett [01:12:42]:
You can do that. You can do that without coreboot or libreboot. So

Rob Campbell [01:12:45]:
Yeah.

Jonathan Bennett [01:12:46]:
I'm trying to I'm trying to remember if... There's some of the laptop companies. I'm trying to remember if Framework is do... One of the ones doing it, that are that are looking at trying to get coreboot running. I think system 76 does work on this.

Jeff Massie [01:13:02]:
System 76 is doing that.

Jonathan Bennett [01:13:04]:
Yeah. So that's that's one of the few places where it probably does make sense. I

Jeff Massie [01:13:09]:
don't I don't know if they've got it working yet, though.

Jonathan Bennett [01:13:12]:
I don't I don't

Jeff Massie [01:13:13]:
know working on it, but...

Jonathan Bennett [01:13:15]:
Yeah. I don't I don't know enough about that. I just know that it's it's a... I've heard the two things mentioned together in the same sentence before. That's about how confident I am.

Ken McDonald [01:13:26]:
Yeah. You've gotta know... If you've got a Chromebook, Liberty Boots supports Chromebooks now.

Jonathan Bennett [01:13:33]:
Some of them. I'm sure. If if you have them From

Ken McDonald [01:13:37]:
the sounds of it, vast majority of them.

Jeff Massie [01:13:41]:
I know when I

Ken McDonald [01:13:42]:
As long as they're not too old.

Rob Campbell [01:13:44]:
When I put Linux on my Chromebook, and that's probably a ten year old Chromebook by now, I had to... You had to replace the BIOS. I think it was, like, c c BIOS or something. I don't remember what my directions gave. Something like that, I feel like it was called. But, yeah, the original BIOS doesn't even work.

Jonathan Bennett [01:14:03]:
No. System seventy six has some, they have some open firmware models. So several of their pieces of hardware, you can purchase, and it comes with an open firmware option. So... And that is based on coreboot. So if you if you really wanna go play with something that's officially supported on coreboot, system 76 is gonna be one of the places to go look at. Alright. That is it for the, the news stories.

Jonathan Bennett [01:14:33]:
We've got some some pretty rad tips that we're gonna get into right after a break. So don't go anywhere, and we'll be right back.

Ken McDonald [01:14:40]:
Not even to get snacks?

Jonathan Bennett [01:14:41]:
Nope. Don't do it. Stay right here. Rob is up first, and, you've got a, we're we're checking something. What what you got for us, Rob?

Rob Campbell [01:14:53]:
We are checking sec or checking security. So I do have a, an article so you can get, all the details. And, but I ran a few commands on this too. So, after installing it, you know, I did with, this is on Ubitoo with the app install, check sec. I think it's on multiple distros, so not hard to get. You know, we can examine common commands like ls bash and sudo. So what... That's what I did here.

Rob Campbell [01:15:26]:
I did check sec files=binls, and then I did the same thing for binbash and, slash user slash bin slash sudo, actually, I did a slash bin slash sudo. But anyway, after installing, or, you know, the output here shows several security features. So what this is, this is a way to check on how something is compiled if it's if it's hard, and if there's security features in place. So what you see here, ReLo, or R E L O, protects certain memory structures from being overwritten. Stack Canary, helps to detect buffer overflows. NX prevents execution of code in memory areas like this, the stack. Pi allows the operating system to randomize where the program loads into memory making attacks harder. R path and Run path show embedded library search paths that could introduce security risks.

Rob Campbell [01:16:35]:
Symbol tells us whether, debugging symbols are present. So they compile with debugging symbols, it's gonna tell you if they're there. And Fortify adds extra checks to certain functions to help prevent memory, related exploits. So look at our three examples, ls, bash, and sudo. All have the, major security protections enabled, which is good. One interesting difference is that sudo has a run path flagged as potentially writable, which, might deserve further investigation in this situation. You know, this is... It's especially useful when evaluating, like, third party or vendor supplied software, something that somebody else can plot...

Rob Campbell [01:17:17]:
Compiled, you know, allowing you to see whether developer compiled it with modern security protections enabled. You know, one thing to remember is that checks sec... It doesn't tell you whether software has vulnerabilities. It just tells us what protections are available to make those vulnerabilities harder to exploit. So as I said, you know, there's a lot more to go into and I'm only gonna go to the basics to to get you started, to really dig in if you're interested in finding out how your things are compiled, check out the article in the show notes for more details how to use that command even further.

Jonathan Bennett [01:17:59]:
Very cool. I like it. Alright. And, Jeff, what do you have for us?

Jeff Massie [01:18:05]:
So here's a here's a tool you most certainly have installed but probably have never typed, Upower, u p o w e r.

Rob Campbell [01:18:13]:
Me power.

Jeff Massie [01:18:13]:
It's the background service that tells your desktop or your laptop or your computer, I guess, about power. It's, you know, like, where the battery icon in your panel, if you're on a laptop, gets its numbers. KDE, GNOME and most other desktops ask Upower over D Bus. Well, the Upower command on the on the command line lets you ask it directly. A good place to start is Upower space dash e. That's for enumerate. It lists every power device u power knows about. On a laptop, you'll see something like, you know, ending in battery and probably something like line power AC for the charger and, you know, you can even see things you might not expect like your wireless mouse, keyboard, Bluetooth headphones, you know, it...

Jeff Massie [01:19:00]:
Because there's many devices that report their battery levels too. And, you know, there... There's always, just so you know, there's always one entry called display device on no matter what system you run it on. That... But that's not real hardware. It's kind of a combined summary. Like, if you have a desktop and there's no wireless things recording battery and things like that, it might be the only thing you see because it doesn't know about any other, batteries. So if you wanna see more details about a device like the, like, your battery on a laptop, you use the u power space dash I and then give it the path to that device, and it will give you a whole bunch of stuff like, you know, what state is it in? Charging, discharging, how much energy, you know, and you'll see energy, energy full, energy full design, energy rate, time to empty, percentage capacity, and charge cycles.

Jeff Massie [01:20:01]:
Now it might, you might be thinking, why so many energy? Well, energy full design is what the battery had when it was new and energy full is what it holds now. So you can see how much your battery is degraded over the life of your system. The capacity is just one divided by the other. So that's the percentage of your energy full number. So if, you know, energy full design is 100 watt hours and your energy full now is at 75 watt hours, well, you have 75% of your battery left. So you've lost 25% because of aging and whatnot. You know, energy rate is what you're currently drawing, so how hard are you running everything. You know, you can you can see that if things are really pulling a lot of energy or not.

Jeff Massie [01:20:55]:
Charge cycles is, you know, tells you how many full charge cycles the battery has been through. So you, you know, it, gives you a better idea of what's going on and where your actual power is at, and it's really useful for batteries. Some more laptops kind of stuff. There is to see everything at once. You can got a dash d for dump, so that gives you every device and every detail of what you can see and it, it... It's depending on what you've got connected, it might be more than you wanna see. But, you know, it's there if you wanna play with it. Recent versions of U Power, you can limit charging to help your battery last longer.

Jeff Massie [01:21:42]:
So you can, you know, slow things down if you've got a super fast charger just to keep your battery a little healthy. Now they they also have a Nomad at a battery preserve health switch, which on new power that you can stop charging at about 80% while you're plugged in. Now that that one is dependent on if you have firmware that supports it. But, yeah, there's a lot of information and take a look at the link in the show notes and the man pages for more switches, more information that you can pull while using U Power.

Jonathan Bennett [01:22:22]:
Very cool. Alright. And, Ken, looks like we're back to ebooks.

Ken McDonald [01:22:28]:
Previously covered the ebook dash meta command available when you install Caliber. Another command is fetch ebook, metadata. Now it will query online metadata sources such as Amazon, Idlewise, Google, and the open library. And while I'm explaining this, I'm gonna go ahead and bring up my, command terminal. And is that gonna be big enough for reading, Or do I need to go... How about right there?

Jonathan Bennett [01:23:02]:
Right there. Perfect.

Ken McDonald [01:23:04]:
Okay. And the, now you may also have other sources. In my case, I do, have some custom plugins for using Goodreads as a source for metadata. The basic command itself is fetch ebook. Let me go ahead and get that out of there. Actually, this is gonna take up a full screen. So, it's fetch dash ebook dash metadata. That's the command.

Ken McDonald [01:23:41]:
And let's go ahead and start, of course, with the auth to use dash dash help. And, now you see why I went ahead and cleared everything. For those of you all listening, when I typed that, it took up the full page, for my terminal. And, of course, with the dash dash help, it gives you how to use it. And simple description effects... Fetches book metadata from online sources. You must specify at least one of title, authors, or ISBN. So ads, you've got a lot of options here.

Ken McDonald [01:24:20]:
The one I'm gonna demonstrate is gonna be the dash dash title here. And it's gonna be a simple title. But if you look, down the... There's an option where you can specify the plug in you wanna use as well. And in my... It does list Goodreads and Goodreads More Tag as two of the options that I have available. But I'm gonna use the dash dash title equals, and the name of the title is Aesop's Fables. And, smart quotes trip you up every time.

Ken McDonald [01:25:16]:
Yeah. Let's put this f back at the beginning. Now what it's gonna do is it's going to, go out to all the sources. And with recent updates to, Caliber, I'd mentioned that it, now uses Camofox as, like a built in web browser. And it's giving you some warnings here. One is that warming up, CamoFox, and that it failed to get the list of fonts for CamoFox. But CamoFox is ready for metadata downloads, and it only took twenty one point three seconds. And there we go.

Ken McDonald [01:26:02]:
And now it gave us a human readable option. And by def... That... By default, that's what the, output is going to be, and it is going to go to standard output. Now, you can use the, dash o. Let's get the dash in there. And that will change that output to an XML formatted output. Again, it's just gonna go to standard output, but you can redirect it to a a file if you wanted to.

Ken McDonald [01:26:37]:
And we've covered that in, previous episodes. I didn't think to look up which one. But you can always go through the, command line tip spreadsheet to, find where we've covered doing that. I think you did it, didn't you, Jeff?

Jeff Massie [01:26:55]:
I don't remember right off the top of my head.

Ken McDonald [01:27:01]:
But it's basically going through doing the same thing it did before, accessing it. And now the disadvantage to this is it does only come up with whatever, it finds first. If you're using the graphic user interface for Caliber, you can actually choose between all the different ones it finds. And, for those of you all listening, it did come up giving you an XML version. It starts off with, the, greater than and less than signs for enclosing, question mark XML version equals one. And goes through and the all the information is still there. But as I said, you could use this, to, download, information to, attach to your ebooks. Or you could even use it, say you've, got a audiobook server like audio bookshelf to, create a metadata file for those.

Jonathan Bennett [01:28:08]:
I've got a I've got a command line tip that you may have used. Not that one. That's the camera. This one. You may have used this one somewhere along the way, but you may not realize that you can install it on your machine, and that's BusyBox. And, we've talked about this before. BusyBox is an all in one application so that you can put, you can you can put one binary and have access to a whole bunch of different programs. And, it's used in places like OpenWRT and a lot of those embedded systems.

Jonathan Bennett [01:28:44]:
But you could install it right on your machine. And why would you want to do that? Well, a couple of reasons. One, I think personally, it's useful just to have it there for reference. Like, what will a BusyBox install be able to do? But you can also you can also run BusyBox sh, and then you have a you have a a a shell right inside BusyBox here. You can run the help command and get information about what it's actually got built in. And they're they're simpler versions of the commands that you, you already use. And so if you want, if you want something that's just a little easier, fewer fewer flags, fewer things, shorter, shorter help pages, those sorts of things, BusyBox is the way to go. But in reality, the the reason this would really be useful is for checking out how a script is gonna run on OpenWrt or something like that.

Jonathan Bennett [01:29:42]:
But it's just... It's interesting because I I asked myself, well, surely you can't actually install Busybox on a desktop? And the answer is yes. Absolutely. You can. And, in fact, the article that I've got linked also includes instructions on how to set up the SIM links so that your other applications like LS and Edit and all of the things that it has here, a whole bunch of, you know, n c and netstat and nice and kill and log read and log in and all of the different applications that can run inside of BusyBox. You can sim link them to where when you run that, it is actually the BusyBox command that runs. So, yeah, just a... Almost a bit of trivia here to know about, but pretty cool that you can do that and, that, yeah, it works almost certainly on your desktop if you wanted to.

Jonathan Bennett [01:30:32]:
So there you go. Alrighty....

Ken McDonald [01:30:35]:
Install that on my ebook reader.

Jonathan Bennett [01:30:38]:
It's probably already on your ebook reader, actually. Chances are if it has a shell, it is BusyBox. I would I would guess. So... Alright. I think that is it. We've made it through the, we've made it through the stories and the tips. I'm gonna get each of the guys a chance to plug what they want to.

Jonathan Bennett [01:30:56]:
We'll let Rob go first. He's gonna tell us about his grandiose plans to become a number one times New York bestseller or something like that.

Rob Campbell [01:31:05]:
Something like that. I hope so. But, I'll start out with, you wanna find me, keep track of me, find out about the exciting things I'm doing. I see my green screen is messed up here on the side here, but whatever. Anyway, the way to do that is go to my website, robertpcampbell.com, and then follow me at, you know, scroll down below you got... Or I just looked down. You don't have to scroll. Links to my LinkedIn, Twitter, Blue Sky, Mastodon, places to donate a coffee.

Rob Campbell [01:31:39]:
I've been posting on those places regularly, and I'll be posting updates too. So first, reminder, my Trailer Park to Technology Leader book is still there. 4.7 stars on Amazon. Got another review that's bumped it up a little bit. But you're gonna wanna read that because the sequel is coming. The sequel's a smaller book. This is, I forgot, three fifty some pages or whatever. The sequel is gonna be, I think, around around 150.

Rob Campbell [01:32:11]:
It's well in the works, and maybe coming out very soon. It's called Something to Prove, and it has it's a different kind of story. It's, not so much about my arc from being born in a trailer park to a leader. It really, it kind of pulls a little more emotional and self reflecting on my life. And and and there's some new stories on my on on my past that weren't in there. I I, did a match... Fixed a criticism that, Jeff had that didn't talk enough about my family and stuff. So you can learn about them a little bit.

Rob Campbell [01:32:57]:
But that's not all. Another thing I started doing is, like me, if you like time travel books... I love time travel books, but there's not enough out there, which is why I have written my first book. It's not it's not ready yet. I mean, it's not ready for you to buy yet. My first time travel book. Forgot the name right now. And I already have another one in the works.

Rob Campbell [01:33:24]:
I am outlining the, I'm I'm doing the outline of it right now. And and and I'm excited for these ones because, if you like time travel books, I... They're some good stories, I think. I mean, I I I love to get someone else's opinion once I get them out there, so follow me and and, you know, you know.

Jonathan Bennett [01:33:44]:
We can be we can be Rob's advanced readers. He'll get... He'll he'll need it. Alright. Cool.

Jeff Massie [01:33:51]:
Yeah. Page A 150. I don't know as I can afford that. That's a pretty expensive book.

Rob Campbell [01:33:55]:
150 pages, I I said. So it's actually gonna be a a bit cheaper than the first one.

Jeff Massie [01:34:02]:
Oh, nice. I don't have a lot, but... And this this isn't really a haiku. It's a little more free free form poem, and this is kinda with the Microsoft crash and the airline issues. It seemed fitting. The network is down, a silent empty highway. No data flows by. The disk is spinning, but nobody is reading.

Jeff Massie [01:34:31]:
A lonely platter. Have a great week, everybody.

Jonathan Bennett [01:34:35]:
Alright. And, Ken, do you have anything to end the show with?

Ken McDonald [01:34:40]:
I just wanna remind everybody that with all the excitement coming up with the upcoming holidays, do take a few minutes to back up, back up, and then back it up all side. Beep.

Jonathan Bennett [01:34:57]:
Now that kind of backing up. All right. Cool. Thank you guys for being here. It has been a blast once again diving into the stuff that's going on and the Office three sixty four outage. With that, we just wanna say, appreciate everybody that's here listening and watching. If you want more of me, we did we did do a Floss Weekly episode this past week, and, that was a lot of fun. We have another one scheduled for an, a bit earlier recording time, and I don't know when the next one is gonna go live because I've got some traveling going on soon.

Jonathan Bennett [01:35:27]:
But, still looking forward to being back into the swing of things with Floss Weekly. Other than that, just, again, thank you to everybody that's here. Whether you get us live or on the download, watch the audio or just catch the watch the video or just catch the audio. We appreciate you being here, and we'll see you next week on the Untitled Linux Show.
 

All Transcripts posts