LINUX Unplugged - 361: Buttery Smooth Fedora

Episode Date: July 8, 2020

Fedora's getting to work and reconsidering some long held-assumptions. Plus the best tool for getting things done on Linux, we take a look at openSUSE Leap 15.2, and breathe new life into an old Pebble. Special Guests: Alex Kretzschmar, Brent Gervais, Drew DeVore, Jeff Fortin Tam, and Neal Gompa.

This is my Geiger counter project that I've been working on for about the past year. I decided to build this project as I discovered these cheaply available radiation monitor boards on Amazon or AliExpress. And although they, like as they stand, they give you a nice clicking sound, like the recognizable sound you hear on movies and TV, they don't really do an awful lot. So I decided to up the ante a bit by connecting this board to a Raspberry Pi, feeding the data into InfluxDB and showing it with Grafana. But then, of course, it needed a fancy case. So I set up on this expedition to build a steampunk style Geiger counter case, including custom machine brass parts, LEDs, and Nixie tubes. Yeah, it is a beautiful case. The input to the project is, of course,

generated by the radiation detector. This board generates a pulse for every count, which is then fed into custom circuitry to drive the LED displays and simultaneously into the Raspberry Pi, which, as he mentioned, not only allows the Nixie tubes to update, which look incredible, but that's also how the data gets into the Influx database. And it's kind of a neat little trick because the Raspberry Pi is locked inside this wooden box behind screws and adhesives. So there's no getting physically to the Pi anymore. Thankfully, you know, and this is a show and tell over on balina.io. So of course, the author here is using Balina Cloud to handle all the updates. They've got a neat sort of

framework where it's all still powered on Docker containers. Now they've got their own forked version from upstream Mobi so that it works well for, you know, IoT embedded use cases. But basically means you can just go, you know, push some updates to your code, push a container to a repository. And then the Pi just pulls those updates, runs the new container, and you don't have to worry about having to go take all those screws out, pull your thing out to replace the SD card with your updated, less broken software. Plus, it's all connected with a VPN connection to a public URL, so you can get to the dashboard from anywhere. That's my favorite part. They might be known for Etcher, but they come up with some pretty cool solutions.

It's worth checking out the link in the show notes. Hello, friends, and welcome into 361 of your weekly Linux talk show. My name is Chris. My name is Wes. Hello, Mr. Payne. 361 is packed full of community news. Our look at the new OpenSUSE Elite 15.2 that has been in the works. And Crazy Drew got his crazy pebble running rebel.

And we'll tell you about that experience. We have an app pick and a lot more. So before we go any further, let's say a hearty, time-appropriate greetings to Cheese and Drew. Hello, gentlemen. Hello, Internet. I'm not crazy. I'm eccentric. Fair enough.

And I agree in a good way. In a good way, Drew. Because now we get to hear about your experiences with Rebel, which as former um pebble watch owner which has been a while now i'm really curious to hear how these old devices which i loved mine had the l cars theme on there curious to hear how it holds up and also we have a duty right here at the top of the show a big shout out to our virtual lug time appropriate greetings mumble room hello hello hello the force isappropriate greetings mumble room. Hello, hello. Hello. Hello. The force is strong in the mumble room.

Some good voices, familiar voices, new voices. It's great. It's great. And if I'm not mistaken, I think one of those voices is Jeff, who is joining us from an undisclosed location, I'll say, because I actually don't know what it is. And he has a new project announcement this week. Jeff, welcome to the show, or a big update at least. Yeah, thank you for having me.

I could say I'm from a bunker in New Zealand, but that wouldn't be the truth. I like it, though. I like it a lot. So, Jeff, I grabbed a clip or two of you from a video today, but I guess I didn't need to because you're here and you have a big release. This is an app that I think a lot of us could use. It's Getting Things GNOME. Version 0.4 came out today as we're recording, just a little bit ago, and it's a big update. It's, if I'm not mistaken, the first update in a long time with 630 new changes, hundreds of bug fixes. It seems like kind of a big milestone. Yeah, it is. And it's an application that I've been using since, what, 2009?

So it's been a while. I've just decided to not let this application go the way of the Dodo because it has gone unmaintained for a while. So I just thought this application is too good. It's worth saving. And now it's basically the last six and a half years of development packed into one release, so pretty significant. It seems like there's a lot of focus here

on improving the contributor experience, with some new methodologies, some new structure around how the project's going to go. Is that making you hopeful that the next release won't be in another six years? Yeah, because I'm not going to try to fix everything under the sun. I mean, part of the way you get to release early and often is to try not to fix everything and just fix the most important things and then leave all the rest, what I call the low-hanging fruit to newcomers to tackle.

Now, getting things done apps are very popular. For those of you that aren't familiar with the getting things done methodology, it's a personal task management system. And I've seen hundreds on iOS and Android. There's websites dedicated to this. There's entire services and companies like Todoist where this is all they do. And I'm curious, Jeff, what motivates you to make this? Because it doesn't seem like we're lacking options for commercial offerings, but maybe we are lacking really nice, well-designed, beautifully designed, I should say, to-do apps for Linux. What's your motivation here? Obviously, I'm biased. I've been using

this thing for over a decade. But I do think that getting things GNOME has no equivalent on Linux, or even arguably on some other platforms. And it's really the, I mean, there's maybe one or two other apps, including Emacs, that you can use to do GDD on Linux, including Emacs that you can use to do GDD on Linux, but it's not the same user experience. It's not built to be not just as beautiful as in iCandy, but beautiful in terms of efficiency and in terms of workflow. So it has a bunch of little details, attention to detail,

that make it a wholly different thing from your regular to-do list. And someone might say, why are you obsessing about something that from your regular to-do list. And someone might say, why are you obsessing about something that's basically a to-do list? But the thing is, it's really easy to make a grocery shopping list to-do list application that's basically computer science 101. You hack that in two hours. But making something that is really well integrated and smart and that has all the crazy features that you need to do something as advanced as the getting things done methodology or some of the other methodologies out there is quite a different ballpark. And you touched on something there that I think is really important for a personal life management app like this. And you said, I've been using this for 10 years. Can you talk a little bit about how an app like this needs to be around for a very long time?

And that's why it's kind of risky to go with something like Todoist or a commercial application in this regard. That's part of the reason why I stubbornly refused to let this app go uh completely unmaintained is i basically invested my life and my my productivity and i centered everything around this app wow and so you don't really migrate that easily to another app for one thing you would need other apps to understand the file format and that sort of thing. Those apps usually aren't built to interoperate, especially in the commercial world. Well, they don't really care to interoperate. I refuse the notion of just going out there and paying like 30 bucks to buy a proprietary application that I am not sure will still exist in five years whenever the company gets bought out by another company, or when they change their terms and

condition, or they decide to stop supporting my platform. That's one of the risks you have with proprietary software, especially software that you really depend on for your daily life, work and play, is to depend on a system that might not exist next year. Yes, well put. Yeah, and of course, if you switch, I mean, all the time spent switching is time spent not getting things done. Arguably, that's also a disease that many GTD practitioners have. We have this shiny widget syndrome thing where we want to try yet another tool and

yet another to-do app. Oh, and this one promises to be just as more amazing than the previous one. And you spend more time playing with the tools than actually using them. And this tool, well, I mean, it's good enough. It works. So I haven't changed it and I don't want to change it. I just wanted to improve it. I'm curious about the experience of bringing it up to a modern

GNOME application because it kind of almost looks like an example modern GNOME desktop app. It's got a lot of the bells and whistles that you see in there. And I'm curious what that process was like to just modernize it. I can't speak for Isidor and the other people who spent a couple years burning out doing the actual port to GTK3. I just came in and finished the job, kind of, by just kind of focusing the project towards a singular goal of getting the release out and managing the project properly. But it's more of a technological problem and a problem of risking burnout

than a design problem. It's kind of special in the way that getting things GNOME was kind of so well designed from the get-go that it was almost a GNOME 3 app already in terms of how the UI behaved and everything. Of course, you need to change the code base port to Python 3 from Python 2 and to integrate things like the GTK header bar widget to have a nice cleaner UI that is now possible. It was not possible 10 years

ago. But the whole design process thing, that was not the hardest thing. The hardest thing is getting things GNOME done and getting it out of the door. And like so many free software projects, it stands on the work of those who came before you to get some of that work done to get it moved over to GTK3. Well, congratulations on the new release. We'll have a link to the project page and also the Flathub entry

if people want to get the flat pack of it. If you're out there looking for a native, local task management application, either for a beginner or somebody who's fairly advanced, I give this a big hearty recommendation. Jeff, thanks for making it on the show and telling us a little bit about it. My pleasure. Thanks for having me. Let's talk about the crazy cats over at Fedora now who are just working up a whole bunch of stuff for Fedora 33, including the end of mod PHP.

And perhaps more interesting to us is no more swap partition. Fedora normally installs itself with a modest swap partition to take some pressure off memory if needed. But a new proposal says to do away with the swap file and instead use a virtual swap device using ZRAM. The ZRAM device works by compressing memory contents and then storing them back in memory when needed and then retrieving them and obviously decompressing. When you're compressing and decompressing a RAM, it's super fast.

And swapping a ZRAM can potentially free up considerable space from a swap file or a swap partition if it's not needed. Yeah, and surprisingly maybe, ZRAM is often faster than swapping to a storage device, even though the CPU has to do the work of compression and decompression. But that's just really fast with modern CPUs. Yeah, that'd be interesting to see if that takes off. I could see it, especially since I think Fedora Workstation's really targeted a lot of laptop users with limited SSDs. There's also some stuff being considered at the lower levels of Fedora workstations really targeted a lot of laptop users with limited SSDs. There's also

some stuff being considered at the lower levels of Fedora. Yeah, some interesting potential changes on compiler policies. Now, longstanding Fedora policy as it is now says that the entire distribution has to be built with GCC, with exceptions only for packages that cannot be built with GCC. But the plan for Fedora 33 is to get rid of that requirement. Instead, the compiler used for any given package would be the one preferred by Upstream. Well, assuming that Upstream has expressed a preference, of course. That would open the way for building a number of LLVM-preferring packages without maintainers having to struggle to get a working build with GCC. And that seems pretty great if you want to, you know, maintain good relationships

with Upstream, especially if you've already been building with LLVCC. And that seems pretty great if you want to, you know, maintain good relationships with upstream, especially if you've already been building with LLVM. And here you go, just some downstream, you know, distro says, oh no, if you want your package here, you got to use an entirely different tool chain. Yeah, I hadn't really thought about that problem before. LWN.net has some great coverage of the reactions too, which we'll have linked in the show notes if you'd like to grab that.

Jeff Law is driving this change, and he said using the upstream preferred compiler would make life easier for Fedora package maintainers. Some of these proposals have strong reactions, but the one that I think has probably gotten the most strongest reaction is that Crazy Cat Neil, who is in the Mumble Room right now, has proposed that Fedora change the default file system for Fedora Workstation to ButterFS. And I'd say it's seen the whole gamut of responses, Neil.

You're here. Yeah, yes. Yes, it has. You sound tired, Neil. You sound tired. It showed up on Devil on a Friday. And thank goodness I had the day off that Friday. And so that Friday, Saturday, Sunday, Monday, and Tuesday, I was responding to things basically almost every other email and pouring over it and trying to figure out, you know, what's going on and the feedback. really been very balanced. The Fedora community took it seriously and responded with really useful feedback that helped us tune and refine the proposal. And also, I want to point out that this is not just Workstation. Every Fedora desktop flavor is switching. The only variants not switching right now are Fedora Server and Fedora Cloud. Those two variants are remaining with their current setup, which is for cloud, I think it's just a single big X4 volume. And for server, it's XFS on LVM. And Fedora Core OS also has its like strange X4 based thing. But

yeah, that's the change. That's just because it reflects the intentions of RHEL not to make a switch to ButterFS. Yeah, it's that. But also, one of the primary motivating factors for changing to ButterFS is that there's this space contention thing. I know you talked about this on air a few weeks ago. The space contention thing is essentially the biggest driver for this. The next driver for using ButterFS over alternatives is that Anaconda, the Red Hat Fedora installer, has special behavior when it comes to ButterFS subvolumes. If you wish to reinstall and you have home as a separate subvolume, it will not make you blow it away because it will just make a new root sub volume or erase the existing one

and then remount the new one that you make. And you can reuse the existing home sub volume. And that lets you preserve your data as you reinstall. So you get to have your cake and eat it too. Access to all the space while also being able to reliably preserve data across reinstalls, which can happen. There are, for various reasons, people may want to reinstall. Maybe they want to switch desktop environments. They want to change some settings or whatever. So this is about making sure that that's still possible because every other alternative blocks that. Now, on the other hand, there has been some pushback, and I think there's also a lot of folks still who are uncomfortable around the idea of ButterFS. They consider it unsafe.

But I remember a time when the newer versions of the extended file system, like 3 and 4, were considered a mess. I remember a time when RiserFS was a mess. I remember a time when a lot of file systems were rough and they got better as development continued. The file systems that didn't get better just didn't have active development. And I think the big difference back then was we didn't have all of these platforms to amplify the problems. You didn't have social media, Reddit and podcasts

and YouTube videos that were just amplifying the issues. You were just a person in a data center dealing with data loss by yourself, maybe bitching on a form or a BBS. And it was essentially contained. But now there's an amplification chamber where people who don't really know the cause of the problems, i.e. it seems like if you go by the data, a lot of file system issues that ButterFS run into today are actually caused by crappy firmware on disks. And this is something that's bared out at Facebook as well. About five, six years ago, Facebook started seriously ramping up

towards a next-generation storage architecture, and they hired on developers to work on ButterFS as their choice for this. And they've discovered over time that basically things that people expected it to be ButterFS' fault wound up being stupid things like the storage controller just can't handle massive IO load and started rejecting drives. Or the SSD just like chokes out because like it just does the wrong thing. And the problem is that ButterFS is sensitive to those things. Other file systems are not.

that Butterfist is sensitive to those things. Other file systems are not. And, you know, people do things like, well, let's intentionally punch holes through the file system and see if Butterfist can recover. The problem with that strategy, when you compare it against XFS and X4, is that when you do that,

the way that Fisk works is that it attempts to repair the mountability of the file system. It tries to restore the integrity of the metadata at all costs. This means that you could to repair the mountability of the file system. It tries to restore the integrity of the metadata at all costs. This means that you could entirely have the possibility of restoring an X4 XFS file system where the links to the data are fine,

but the data itself is bad. Right, and it just continues on as if nothing's wrong. Exactly, and in ButterFS's case, and of course, you and I both are aware, like we talked about ZFS before, ZFS also does this. If you repair the file system and it does a validation of the data as part of relinking the metadata to the data and it discovers that the data is broken, it will loudly tell you so and it will say, you've lost data. Good luck. Have fun. My final thoughts on this, because I know we've talked about it a lot, is essentially

in the Venn diagram of when to adopt open source or free software. I think ButterFS is now sitting in the sweet middle. It needs more deployments, but it has now been tested not only in scale with lots of NAS products that ship with ButterFS and obviously at Facebook that gets mentioned all the time, but also as the default file system on OpenSUSE for years. So we've had a lot of production testing, and now we could bring it to even more masses. And the reality is Fedora and Red Hat,

they're never going to ship ZFS as the default file system, but they need a competitive product. They need something that competes with modern NTFS and APFS. Things like snapshot-based backups and Windows shadow copies have been commonplace in the commercial operating systems for a while now. And we will never even get to the point of starting to build those features into things like our desktop environments and our file managers or even user land tools until it's shipped. Developers need the code shipping to use it. And there's things we need to work out, too, on how to properly ship all the updates and all of that that will only happen when we start to do it at scale.

And it seems to me that if you were to review the landscape of possible Linux distributions that this could happen in, Fedora is an obvious case. My old butt remembers a time when Fedora was so cutting edge that they adopted Pulse Audio before everybody else, and they shipped a volume applet in the GNOME menu bar that didn't actually change the volume. But they planned to make that work in the next release. And it was important to get Pulse Audio out in front of end users and make it reliable. And Fedora could bring

ButterFS to a whole new scale of users. And the ideal Fedora user from a project standpoint is a sysadmin DevOps developer type person who's using a laptop with an SSD in it that's probably 512 gigs or less. And that is such a perfect use case for ButterFS. Additionally, features like encryption are coming down the pipeline that already use the same mature data pipeline that compression uses on ButterFS, which means it's going to be a solid crypto using a solid proven data pathway to make it happen. And you'll be able to encrypt a single file, a single folder, or encrypt an existing mount point without having to wipe the data and set it up again. And that single limitation that exists now in Linux has forced vendors like System76 to create installer methodologies that wipe and reload your system after you receive it, which if you think about it, is mental that they have to do that because it just

enters so many possibility of mistakes and errors where it could leave the user with an unusable system. So they have to test the hell out of it to just work around this problem where they can't encrypt a mount point or a home folder that already exists. So they can't ship you a system that's set up and ready to go. They have to do that work. And that's crazy. And it seems like something from the 90s. And Linux should be a bastion of technical excellence. And we have a file system here that is GPL. It's free software. It's completely cool and okay to link it to the kernel. And we're just sitting on our hands going, I don't know, I've heard some bad stuff about it. Meanwhile, people are putting in production going, actually, it's gotten a lot better. Last but not least, even though I said I

know it was almost done, but now I'm going on a roll. I was a big critic. I said it was junk. And I was willing to revisit that and try it in production again. I even was willing to revisit it after I blew away data because I realized it wasn't Butterfest's fault. It was my fault. I screwed it up. I even I will now eat my's fault, it was my fault. I screwed it up. I, even I, will now eat my words and say, you know what? It's gotten better. Maybe it was crap when I tried it and my systems wouldn't boot in the past. But now here I am using it and it's been working fantastic for me.

And I just have to implore people to just give it another look because the reality about free software, and this is something that's great, is that as long as development continues, it's always improving. And if you haven't checked in lately, then maybe your opinion is out of date. And I'm not saying it's forever use case, but the developer workstation with a 256 gigabyte SSD in their laptop, I think it's probably the ideal case. End of soapbox. But I would love to see it happen in Fedora 33. And I'm curious, Neil, in your personal opinion, how close is that actually to happening? I mean, is this getting dangled in front of me and it's going to be yanked away? I sure hope not. I mean, like, man, I've been working on this personally for a little over three years. Like, I've been doing

integration at almost every layer of the stack for a few years now. And it's just nobody more than me would be sad if I had to feed from the jaws of victory here. I have no reason to expect it to not be approved and that we would not start going forward with this within the next day or so. Basically, once it's approved formally by FESCO, the Fedora Engineering Steering Committee, then we will go forward with starting to implement all this. And I have in my mind a mental roadmap of ideas of things that we can do to start leveraging these things as we validate and test these features and integrate them into our user experience. I'm going through this actually

fairly conservatively. There are many things that I am personally experience. I'm going through this actually fairly conservatively. There are many things that I am personally using that I'm not even including in this change proposal because I want to make sure that we have this rock solid as we go forward and adopt features. That makes sense. It's no fun to be the guy who has to say, well, I'm sorry, like your data's gone

because like you messed up here. It's Neil's fault. He got everybody to use ButterFS and you lost your data because of Neil. Yeah, no, no, no. The reason why I'm excited for it is I like to see this stuff happen. I like to be on the front lines and watch it and then hopefully see developers begin to integrate it. And that just won't happen because today I could install Fedora and I could choose butter FS or XFS and I have and it's fine but it's basically as functional as any other file system there's nothing integrated with the operating system there's no real advantage to it it's just disconnected it's

disjointed and we can begin to bring that together if we start to deploy it as default and still leave all the other file systems available for people to select. You could put Extended 2 on there for all I care. So keep an eye on it, Neil. Hopefully we'll see it develop. Now, we are going to be off next week. So if it does happen, I mean we're prerecorded is what I should say because we never really take a week off. It's the steady beat of LUP.

But if it does happen, we'll have to cover it in two weeks. It should be interesting. And then actually, you know what? If it does happen, we'll be talking about it in the LUP plug. So why don't we right here, this is probably the best time to do a little housekeeping. So the LUP plug happens every Sunday at noon Pacific, the regular broadcast time in our mumble room up in the lobby. I encourage you to join it. Now, I mention the mumble all the time and I in our Mumble room up in the lobby. I encourage

you to join it. Now, I mention the Mumble all the time and I want to let you know how to find it. I should probably put it on the main Linux unplugged website. It never crossed my mind. Wouldn't that be an idea? You know, I guess we've just been around a long time, Wes. We never thought about it because the traditional way is you just go Google search Jupiter Colony Mumble and the setup guide is there. You can follow that and get it set up in a few minutes. But I think we probably should have something on the main site, so we'll probably do that eventually. Go Google search Mumble Jupiter Colony, and you'll find it in there. You'll find the guide, get it all set up, and then join us on

a Sunday and chat with us and be part of a lug. We have 30 folks in here right now. If you think about that in terms of people showing up for a lug meeting, that's a really great attendance. Usually on Sundays, we usually have around 20 and we're just chatting about all kinds of stuff. So yeah, like I mentioned, we won't be live next Tuesday, but the LUP Lug continues every single Sunday. And then the live show will be back on the 21st of July, 2020. Obviously, if you're listening to this in the future, I'm going on a little family trip to Montana. And then Linux headlines will also be off back on the 20th because Drew's taking a little time off, too. That's great, huh? It's about time.

We're finally letting him. Finally. I mean, I'm going to make him set up cameras and wear a tracker, obviously. But we're going to let him take a little time off. And then you can always check the calendar. Yeah. Put that, put that tracker back on, Drew. This ankle monitor has given me a rash, dude. Yeah. Yeah. That's what my kids and wife complain about too. You can know when we're live at jupiterbroadcasting.com slash calendar. Hopefully we just keep that updated and it's all over there

and then you don't have to worry about it. So there you have it. That's it. That's all I got for you in the housekeeping. Nice and tidy around here, Wes. Doesn't take very long. This is my last podcast total from Texas. You sound a little sentimental about that. I am.

I've enjoyed my time down here. It's so much nicer to drive down here. They've got like side highways, Wes, that go along the main highway. And then they've got frontage roads. And then they've got toll roads. And so you can just zip around here. You just zip. I mean, I'm in Austin and it's still better traffic than you get 50 miles north of Seattle. So I will miss that. I won't miss the heat though. The feels like temperature is 108 degrees Fahrenheit today. So I won't miss that. All right. Well, let's talk about OpenSUSE Leap 15.2. A new leap is upon us. Speaking of ButterFS-based systems, I have not tried out SUSE in about 275 years. So I thought, why not give it a go and immediately install the server version, which I'll talk about in a little bit. But SUSE has

leaped again. I don't really know what to say about SUSE because to me, it's this Linux distribution that I try for a little bit and then I end up usually going back to other distros just because I've never really just gotten in the groove. But 15.2 contains a lot of nice stuff. They're shipping Sway for Wayland, which as you know is a drop-in replacement for the i3 window manager. They're also including a real-time kernel for some folks that need that, like embedded devices or audio pros, or if you're capturing a lot of data, real time can be very important. It's just nice to see this going in another distribution. Of course, Yast has gotten some improvements, as always.

And the other thing that's interesting about Leap is they're continuing to support a migration strategy to go from OpenSUSE Leap to SUSE Linux Enterprise, their commercial product, which I suppose probably happens. Well, right. I mean, you know, maybe you're testing it out, maybe you're considering switching, or you just, you know, you build out your prototype. Down the road, your business does well and you realize, oh, you know, we can afford support now. We need support now. And you don't have to re-architect stuff. You should just, you know, switch on over. Yeah, it's just all those people are using Ubuntu. Well, there is that, but maybe you're in Germany.

Yeah, I suppose. It's just, you know, in the Ubuntu model, and I don't mean to harsh on Seuss's vibe here, but in the Ubuntu model, it's just one Ubuntu. You just go get the LTS. There's no silly different versions. It feels like a Windows thing to do this. Yeah, or I mean, it is definitely an older style, I guess, right?

Not a hyper-converged. Which is okay. I mean, that still works for a lot of people. But, yeah, I could see it. I could see it being a possibility of try-before-you-buy sort of thing. So let's go around the horn and see what you all thought. Cheesy, let's start with you. Why don't we do the old say what you didn't like and then end on what you did like?

And so I'm going to look at this as kind of a new user perspective. That's the perspective I generally like to give someone that hasn't used a lot of Linux distributions. And Seuss is one of those that's kind of, like you said, kind of on the outskirts a little bit that a lot of people may not have. Yast, while cool, is still Yast. And it feels a little nostalgic to me. The installation was a little awkward. I didn't get, you know, like a simple partitioning method. It did select the disk that was empty by default.

So it was smart enough to know that this disk didn't have any other file systems on it. Use that disk. There were also some other little features in the installer, which kind of annoyed me. They had green buttons with gray text on them. Why would you ever do that? It's a horrible idea. Please don't. Because I didn't know when to click next. Was it done? Was it supposed to light up black over the green button whenever it was done? I didn't know. So that was a little awkward. But I mean, installation was pretty straightforward. No big deal there. The default installation and version of Grub recognize my Pop!OS stall on another disk and was fine.

I was able to switch into either one. However, after I updated it for the first time, it removed that entry from Grub. So, you know, not a big deal. Something I can get through. But, you know, if you are a user that's trying out multiple distributions on a system with multiple hard drives, that might be something you come across. And then after the first update as well, auto login was turned on by default. Yeah, I think it is in the installer too. I think during installation. Why would you, that's no? Well, you know, I think there are by far and large, my use case is, you know, the system I'm setting up at home where, yeah,

I never turned that on. But I imagine for a lot of people, it doesn't make sense to have to log in if you have one compute user on the account and you keep your computer in your house, especially right now. That's true. But I mean, I think it's always good practice to have a login. The thing that's annoying about that is then you just have to unlock the keychain anyways. You know, like KD wallet or GNOME keychain still has to be unlocked when you auto-log in. And so you end up just having to put your password in anyways. Right, right. They also use a little bit older version of GNOME, which was good because the problem I'd had with folders in GNOME 3.6 isn't there because they're using GNOME 3.4.

It does use a 5.3 kernel. And then another thing that was just really kind of awkward to me, I've seen other distributions do this, there were two terminal applications. There were two mail applications by default. So there seemed like a lot of stuff they could have cleaned up. Yeah. It's just an old way of doing it huh yeah well and it's you know they just auto populate the the menus and and that's what

you get right so i mean i i like that it's installed by default like having mutt there is cool but if you're going to do that does mutt really need an icon in your application menu if you're just going to open up the terminal and then open MUT that way? So it seemed a little weird in that regard. You know, I would have personally cleaned up a few of the apps and made that selection a little better for the end user. But on the good side, I mean, it did seem pretty snappy. I know I'm the only person in the world that's a fan of Hot hot corners, but they were turned on by default, which I thought was cool. The offline installer was super cool. Yeah, I do like that. It threw me back into a little bit of nostalgia, right? With

Yass, the offline installer, everything's there on, you know, the 4.6 gigabyte ISO. Reading the release notes as files install. Right, right. You know, so, I mean, it was cool. It was cool. And then the overall selection of default apps were pretty solid apps. I mean, they weren't bad, but there were some redundant apps, which is kind of weird. But overall, I mean, I think it's a good distro. Is it something that I'm going to use on the day-to-day? You know, probably not, at least not right now. But I will review it again in the future as, I mean, it was the very first user-friendly distro that I ever used.

I think I pretty much am right in line with your observations there. It feels almost like a different branch where we're trying things out, and some things have stuck that didn't stick in the other branch of Linux, and some things should be adopted that haven't been adopted by the other branch of Linux. Wes, I'm curious your thoughts on this guy. I think, in my opinion, the installer manages to strike that balance of easy to use for the most part, but gives power users a lot of options. And perhaps I think it gets there a little better than Anaconda does for Fedora.

You know, honestly, I was really impressed with the installer. It's been ages since I've installed OpenSUSE. And yeah, you know, at parts, especially during the actual installation, it was a little bit slow for me. Now, full disclosure, I was installing in a virtual machine and then rebooted, you know, into the real OS afterwards. That might have just been my own internet or, you know, whatever mirrors I happened to be using at the time. But the rest of the installation experience, really, I couldn't have been happier with. And I was really impressed at the end, especially, you just got this really detailed rundown of all the options you've configured, and then little clickable links, if you need to go adjust that. One that really stood

out to me was, let's say you want to change, you know, if you're enabling or disabling Spectre mitigations, you just click on that, pops you up to where you can go make adjustments to whatever grub parameters are going to get set for your kernel command line right there in the installer. And I don't see that hardly anywhere else. It really felt like it was at home in the enterprise, but it didn't make you go through those things if you're just a normal user and, you know,

you don't even know what that means. I do agree with you that anything to do with package management feels like your soul dies a little bit. I just wince when I click on the online updates and it does like the four or five prerequisite things that it has to do before it can even execute the very task that I've asked for. Just compared to every other package manager on the entire planet that just immediately gets to work when you tell it to do an update, I find that just soul-crushing. The transactional server option was interesting. So this is an installer mode you can now select, and one of the key features is a read-only root file system. You kind of need that if you're going to go with the whole transactional style, right? You kind of need that if you're going to go with the whole transactional style, right? And we've all had, especially thinking of upgrades over in the Debian world with Apt,

where something breaks, or maybe it's on Arch, and suddenly you just, you can't change your packages, you don't know what to do, and you just wish you could roll that back and try it again later, right? And if you enable this transactional server support, that's what it does. It comes with a read-only file system, and then when you install stuff, it all goes in a brand new snapshot. And only if that installation to the snapshot volume goes well do you actually go reboot into the new system. One thing that's different, too, compared to things like OSTree or some of the CoreOS style,

there's been a lot of different takes on this idea, but in the OpenSUSE land, it's all powered by the same zipper and ButterFS. So you don't have to rebuild your packages to support this new transactional style. You just basically get it for free. There's a good example of a distribution that is starting to leverage the built-in capabilities of ButterFS instead of having to create that code themselves from scratch. And it's just an aspect of maturity of shipping Butterfest for a while. These things are just now commonly available to them, and they can integrate them more and more.

And I think that was a key point I was trying to make earlier with the Fedora adoption is they have to adopt it now in order for features like this to come along down the road. This is pretty neat stuff here. Yeah, the other stuff I thought was really kind of neat is there's a transactional shell feature. So if you'd like, after that installation installs into this new snapshot, you can get a shell, you know, basically a cheroot into that environment. If you want to go test some things out, make sure your package is installed correctly, tweak any of the configuration before you reboot. And I think another sign of maturity, there's a built-in health checker framework. So after you've rebooted into the new server, the health checker runs and verifies the state of the system. So you make sure that all the services you need are running,

make sure that the database you installed starts up okay. And it's just a bunch of shell scripts, so it's easy to extend if you're using this to build your own custom application. You can include your own logic there to make sure that your custom app starts up okay. And if not, the system automatically rolls back to the previous snapshot and reboots. So you're done. That is so what I'm talking about. That is really a really nice setup. And like you said, it just seems like a mature way to do it. It's enterprise grade. Exactly. So check out Leap. It's a pretty solid option. And if you're not an old grump like me who's stuck in his ways, it may be a great opportunity to try something that has a lot of unique features and capabilities to it. I think it's something kind of special in the Linux world, and we don't talk about it a lot because none of us are SUSE users, but I'd love to hear your experiences with it. If you have tried out 15.2, go to linuxunplugged.com slash contact and write in and let me know how it's gone for you.

I think it's an opportunity to kind of crowdsource people's results. You know, it is kind of, it's been something of a blind spot for us, I guess. And there's really a lot of interesting stuff there that's happening, especially around YAST and the Butterfest adoptions. And a lot of good ideas, and I think we could do a little better job of highlighting that. And that's something the audience around YAST and the Butterfest adoptions. A lot of good ideas, and I think we could do a little better job of highlighting that. And that's something the audience could really help us with. Linuxunplugged.com slash contact. Okay, so on December 7th, 2016, it feels like it was just yesterday, Pebble announced they were ceasing operations. Two days later, on December 9th, 2016, the Rebel site went live. And our very own Rebel,

Mr. Drew, has brought his pebble back to life. So Drew, tell me about this process and this experience. It's something I've always wanted to try. Were you still actively using the pebble? Did you have to dig it out of the drawer and it just happened to still work? The very first step was remembering which bucket I had put it in and finding it. The second step was finding the charging cable, which was not right there with it because I just am terrible at sorting old junk that I've tossed in a drawer. And it's got a weird charger too. Yeah, it's got the little pogo pins and all that. It's got a weird charger, too. Yeah, it's got the little pogo pins and all that.

But once I found it, it charged right up. And the process to install is a little involved, but it's not terrible. The instructions on the Rebel website are pretty clear, and they're separated out for if you're on iOS or Android. And me being on Android, the number one thing that you have to do is you have to go and get the old Pebble app, which is not on the Play Store anymore. You've got to download it from APK Mirror. And once you do that and kind of get everything connected, then it's really simple. You just have to go to two websites. And in the back end, the Pebble app will switch from the Pebble web services over to the Rebel web services, and you're off and running. How does that switch happen? You literally just go to auth.rebel.io in a browser and sign in. And then after that,

you go to boot.rebel.io. It'll reopen the Pebble app, and you'll be greeted with the Rebel login screen instead of the Pebble login screen. That's it. Is it using some sort of software update function to do that? That's pretty simple. That's nice. I don't know how it works on the back end because when I was reading the instructions, it's like clearly there has to be something more to this than just clicking two links. This doesn't make sense to me, to my little brain. How does it, you know, make that switch within the app itself?

But it does. And once you're in, you can download old and new watch faces and old and new apps. Pretty simple, pretty slick. And it works just the way that you remember back when it was the official Pebble stuff. That's pretty great. Now, simpler than you were expecting, it sounds like? Oh, yeah. Much, much, much simpler than I expected. I thought I was going to have to be like doing some ADB stuff or something along those lines. No, not at all. Are they hosting these services? Yeah, they are. And the Rebel Web Services are basically a drop-in replacement.

It's all hosted on their end. And they do have a premium tier where you can pay $33 a year or $3 a month, and that'll give you dictation and weather services. But other than that, the main reason to do it is just to support the Rebel project so that they can keep their servers online. So once you got this thing up and going, were you able to do anything cool? I did try to get the old home assistant app up and running now that I'm a home assistant user, but unfortunately the app was not authenticating properly for me. But I found another app called

HTTP push that allows me to put in custom HTTP, you know, get or push requests. And I set up some webhook automations in my home assistant. And so now on my watch, I can launch services, I can toggle lights, I can even set it up so that it will broadcast to my Google Homes and let my wife know that I'm on my way home if I'm out picking something up. Neat. Very modern. And all the old health features and everything still work.

You know, it still even ties into Google Fit, which I really like. And I think my number one feature, the thing that I missed most of all, is the smart alarm, where it will track my sleep and figure out when I'm not in REM sleep. And that's when it starts vibrating on my wrist to wake me up. That makes all the difference. Oh, I missed it so much. It was so good. Yeah.

You know, I just did this yesterday and I used that smart alarm feature last night and I woke up feeling better than I had in quite a while. Is the integration pretty good with your phone OS? Like music controls or podcast controls work okay and stuff? Oh yeah, totally. I gave my Pebble to my dad and he still uses it. And I don't know if he did this or not. He may have.

He's just technical enough. And it sounds like the process wouldn't be too difficult. Does it integrate in with calendar services and stuff like that as well? Absolutely. And they show up on your timeline just like they used to. And the Rebel people recently set it up so that you can push other things to a separate timeline and it'll come down to your watch. So you can do things like set up, you know, if this and that integrations to have like RSS feeds and things pop up in your timeline, which you couldn't do before because that was tied in with the old Pebble service.

But now they've stripped it out and polished it up and made it fresh again. So you're going to keep using it? Oh, yeah, absolutely. Yeah? Really? Yeah, totally. Oh, cool. How's the battery? Is it still hanging in there? Yeah, it's doing great. It's as good as I remember it being, you know, like eight days of battery life on the model that I have. And it's, you know,

it's only been a day, but it's sitting at like 90 to 100%. So. Oh man, I wish something would have come from this line of technology because it just makes so much sense to use an e-ink display on a watch. And when you say eight days of battery life, my mouth drops open because I've had several newer watches, digital watches, and they get a day, maybe two days if I'm lucky. I'm curious, Drew, if you're now going to look around at the old Round and the Steel and the other series, because you should be able to get those for a ridiculous amount of money nowadays. I mean, they should be pretty cheap. I loaded Rebel a long time ago, probably in 2016 or something like that on my old Pebble, but I've since put it away. And it was okay then, but it sounds like it's worth a revisit now.

Yeah, I remember they were really cheap right when the service had just, you know, been announced that it was going away and they were getting rid of like the old stock and the used market is not as cheap as you would think. I was looking at some time rounds to replace my wife's, her old one, she loved it, but her battery exploded and haven't been able to, you know, get that going again. So now I either have to get a new battery and do some soldering or try and find one. But they're like 120 bucks on eBay. Oh, wow. Well, they're not making any more of them, so I could see why. You hold onto it as long as you

can, Drew. I mean, it already kind of is a collector's item, isn't it? Yeah, totally. Well, thanks for taking one for the team and totally making me miss my old Pebble watch. That sounds pretty cool. And it sounds like a hearty recommendation if anybody has one rolling around in a drawer out there. Yeah, absolutely. And if anybody has an old time round that they want to sell to me for cheap, I will take it. There you go. At Drew of Doom on the Twitters. you go. At Drew of Doom on the Twitters. That's right. Slide into those DMs.

All right. Let's do a pick really quick. Wes Payne, you found a Markdown editor for the 21st century. And you know me, whenever there's a Markdown editor, I want to talk about it. This is a local native application, right? I think. Well, it is Electron. Okay, okay. But they've got Fedora and Debian packages and app images. So, you know, almost native. They're making a good concerted effort to at least publish it to the Linux desktop. I'll say this. It looks really great. It looks like a solid app.

And one of the things it has that seems kind of appealing, like a differentiator to the other Markdown editors, is it has a pretty sophisticated research mode. They have a video that explains it, and in just a couple of seconds I grabbed a clip that talks just about the research mode. If you're a professional author or researcher, Settler has all the tools you need to write efficiently. Simply group everything together using the project feature and arrange all your chapters neatly. A mighty citation engine sits at the heart of the app, letting you cite your sources with ease. Pull in your literature and choose from over 9000 citation styles. You can even use templates to directly export your paper using the format desired by the journal.

The other thing that's kind of neat is it supports exporting to LaTeX as well as other formats, and it supports importing a bunch of formats. It has code highlighting for different languages besides Markdown as well. And of course, best of all, it's free and open source. No, no. Best of all, it's got dark mode support. But seriously, you know, there's a ton of really great Markdown apps out there, but a lot of them kind of just stick to doing Markdown.

And that's good, you know, but there's a whole world beyond Markdown. And a lot of times, you know, you might use Markdown just for some casual notes or a simple document. Settler really lets you take it to the next level and make serious publications, do research, integrate with citations, integrate with LaTeX. It has Pandoc integration as well. So if you need to pull in sources from Word documents or render out to them, it's all built right in. It really seems like, too, they've taken a focus on helping people collect research, either journalists or everybody, like anybody that's trying to just pull together a bunch of

stuff and then cite sources with a reference manager sitting on top of it. It seems like that component, along with the UI and the import export options, do kind of make it stand apart. And I have now taken it just I write everything in Markdown. I just text editor. If I'm even just taking notes for something, I write it in Markdown. If I'm making a to do list, I do it in Markdown. So this kind of stuff I find to be pretty appealing. I, for the most part, just stick

with plain text editors, but I absolutely could see busting out Zettler for a research project or trying to pull something together. Say we're going to do a history of an open source project special. Absolutely could see using it for that kind of stuff. And then it keeps it all organized for you in a really easy toto-browse way. And you can combine everything into a project so it's all kind of in one spot, even if there's multiple files. And, of course, on the back end, it's all just plain text. So, you know, if you want to stop using it one day, no bigs. I like that, too.

Pretty good find there, Wes. Pretty, pretty good find. Where can people find links to that and everything else we talked about today? Well, in our show notes, of course, linuxunplugged.com slash 361. And that does bring us to the end of this week's Linux Unplugged. The show is at Linux Unplugged on Twitter, where you can get news announcements about the show, live stream announcements, releases at Linux Unplugged. The whole network, that's at Jupyter Signal.

But if you don't use the Twitters, I don't blame you. We'd still like to hear from you, though. Go to linuxunplugged.com slash contact. I'd like to start featuring more of your feedback in every episode. Right around this point in the show, I'd like to be reading your emails. And the only way that's going to happen is if you send them in. linuxunplugged.com slash contact. I'm at Chris LAS.

He's at Wes Payne. Thank you so much for joining us on this week's episode of the Unplugged program. We appreciate you, and we'll see you right back here next Tuesday! Thank you, everybody, for joining us. It's time for me to head out, hit the road. I'll be on the road early tomorrow, so that way I can get a solid, I don't know, 100 miles in before work starts. And then I'll work. And then as soon as work wraps up, I'm jumping on the road, and I'll probably be driving until nearly midnight.

But that's not so bad because it's so dang hot down here that that kind of works. You drive and then you park somewhere with air conditioning. You work for the bit and then the evening around 8 p.m. it's it becomes nice out again. And then you just drive and drive and drive. So so you've switched to being a night driver. Well, just while I'm in Texas, it's because it's hot, Neil.