LINUX Unplugged - Episode 262: Tribes of Init

Episode Date: August 15, 2018

The FreeBSD community shares the hard lessons learned from systemd, we play some great clips from a recent event. Plus our work-arounds for Dropbox dropping support for anything but vanilla ext4, the return of an old friend, and a ton of community news and updates. Special Guests: Eric Hendricks and Martin Wimpress.

Anybody that's listened to the Unplugged program for a while knows that part of our secret sauce is our virtual lug. And we use an open source program called Mumble for the voice chat. And anybody is welcome to join us. You can just go Google search like Jupiter Colony Mumble Guide. And as long as you got a working mic and you have headphones and a microphone and we can do a sound check, you're welcome to join us and chime in on your opinions about what we talk about. But I think for some of you, the big holdup has been, it's a dirty GUI app.

Who wants a GUI app? I mean, if you're really listening to an unplugged program, you're probably in a terminal somewhere. Right? I mean, you're downloading your podcast using Bash Potter. You're using links to browse the web. You're just doing everything via the command line. And you're hanging tight,

hoping we'll have another cool cururses-based interface for you. Well, good news. Bernard is a terminal-based client for Mumble. So now you can run Mumble on the command line. It's a Go app. It's available up on Git, and of course it's free and open source. You can even participate in the chat.

It's kind of neat looking. So you can install Bernard if you have Go installed, and be up and chatting on the Unplugged program by next week. This is Linux Unplugged, episode 262 for August 14th, 2018. Welcome to Linux Unplugged, your weekly Linux talk show that's fighting the BSD temptation. My name is Chris. My name is Wes. And my name is Brent.

Hello, guys. It's great to be with you. We have a really fun show. So not only do we have our traditional community news, some new fancy Linux features, my favorite, don't call it a distro update. But I listened to a BSD can talk, and I was blown away with what they had to say about the tragedy, quote unquote, of SystemD, and the lesson that the author believes FreeBSD has to take from SystemD, and the trap that they may be leading themselves into right now.

I've got a couple of clips, plus we'll link to the whole talk. I can't wait to get into that. And then, at the end of the show, after 10 years of using Dropbox, every minute, knowing one day it would bite me, this last week, Dropbox announced it was dropping support for encrypted extended 4 partitions and any file system on Linux that's not extended 4. And as some of you know, I'm a big, big user of XFS.

I think it's a fantastic file system. It's what all of my home partitions are formatted in. It's what all of my main data partitions are formatted in. You keep reformatting my hard drive with it. That's right. I learned from Alan, you know. Alan used to come around and put everything on ZFS. Well, this week they said, no, not for you.

We are dropping support November 7th. Now, trust me, I've known this day was coming. I've been planning an open source salvation. I've been running them in parallel, some of them for a couple of years, planning this mass exodus. But after using a tool for 10 years and building a large production pipeline around it and having a whole team of people that use that software, it's going to take a lot of work. So we'll start this week with telling you how you can buy yourself some time with a couple of simple and handy little cheats.

Trick Dropbox into thinking it's chugging right along on Extended 4, and you don't have to touch a thing. Plus, Fedora also has another solution. It's a little outside the box, but we'll get into that. But before we go any further, we've got to bring in that virtual lug. Time-appropriate greetings, Mumble Room. Hello. Hello, Eric.

Hello, Wimpy. It's a small mumble gathering. That happens from time to time. Last week, we had nearly 20 people in there, and this week, it's just a couple. Everyone's out having fun in the summer. It's holidays. Holiday season in Europe. It's good to have you back, Wimpy. How was your holiday? It was wonderful. I had a great time, thank you.

Some of the best pictures I've seen in a while on Twitter. I mean, you got some really great shots on this trip. I was watching from afar and I was quite impressed. Yeah, we were lucky. The weather was amazing the whole time we were away. And there were only a few pictures. You'll notice that they were basically on the way to where we were going and on the way back because we were in a very remote village in France with no internet whatsoever.

So two weeks of complete blackout. It was great. That sounds like an interesting psychological study in a way. Yeah, it took about 24 hours to adjust, but then it was great. Yeah. Did you feel like you were disconnected? Did you feel for the first 24 hours like you were missing out? Did you have some FOMO or what was it at first? It was when we were having discussions about where we were going to do and what we were going to do and I kept picking my phone up and then I was like oh sure I can't actually find any of the information that we need to answer these questions and it was like so we have to

drive into a town and find a tourist information office and pick up some leaflets yeah it's very very old school. It sounds wonderful. Yeah, it was good. It was really good fun. Good for you. Well, glad to have you back, and let's get into some news. Let's start things off with a fresh kernel.

Santa Linus has dropped 4.18 out there. And I just wanted to mention, I don't normally call these out anymore, but the whole way it was written is sort of noteworthy. He says, It was a very calm week, and arguably I could have just released on schedule last week. But we did some minor updates, mostly networking, but with some VFS race fixes, and a couple of driver fixes, and some other minor random things. You know, we just figured it was time.

So some of these new things I could have just delayed to the next merge window, but they were marked for stable anyways. So it just would have caused more backporting. Anyway, with this, the merge window for 4.19 is now obviously open. Can you tell he's wearing slippers while he writes this? And the world's most powerful operating system sees its next significant update. It's funny, too, because there's a bunch of other neat stuff in this kernel.

I guess that's not what he sees. People are super hyped about this one. That was the other reason I put this in there is a lot of people sharing the release of 4.18. I mean, I know there's some Steam controller stuff in there. Yeah, this is the first kernel version for a long time. I've been really waiting for

this one to drop. Oh, okay. Tell me why. It's all rather selfish. This is the one that lands all of the enablement for the GPU in the Hades Canyon NUC, for example. So I've been following along with the release candidates on the Hades Canyon to have that ticking along. But there's a bunch of fixes in there for AMD stuff generally. So the gentle ascendancy of AMD at the moment in the Linux space is marching on, and 4.18 is definitely a landmark kernel for that. Yeah, that's what I grokked most the excitement was about.

But I had wondered about the NUC stuff. That's good to hear that it is in there, because that, to me, is still something that I hope could work out to be a great box in the studio one day. I don't know. You know, those NUCs have a hit and hit, hit sort of hit and miss track record when it comes to recording. Sometimes they have noise in them because they're so small, so tightly built. But we'd love to try. Yeah. I've, um, I've been trying out a new co-working space since I've been back from holiday. And some of the desks there have monitors on arms on the backs of the desks so you can plug your laptop in. But then I was sat there today thinking, hang on a minute, I could just pack my NUC into my backpack.

So tomorrow I'm going to go to this co-working space with my NUC and a wireless keyboard and mouse in my backpack. So, so tomorrow I'm going to go to this coworking space with my, my nook and a wireless keyboard and mouse in my backpack. And I'm going to plug, plug my nook in on the desk and use that as my workstation for the day. Cause it's actually no heavier than, uh, than my laptop. Right. So, um, what about the video session with Popey? I mean, uh, if you're in a coworking space, how does that work now because i know that was a thing you guys did well i've been so uh this because it's been so hot in the uk i've actually been working out of a bar in a local hotel for about the last five weeks

wow because they have air conditioning right exactly? So I've already got used to, you know, sitting on conference calls with Popey and other people at work, you know, in a bar with people, you know, pouring pints of beer over my shoulder and all the rest of it. And I recently discovered this co-working space was opening up nearby. So I've been going there just recently, which is obviously more conducive to video calls and also has 100 megabit Wi-Fi.

Hey, not too shabby. So it's good, yeah. So tomorrow I'm taking the Hades Canyon NUC in, and I'm going to use that as my workstation for the day. That's great. So there was an announcement with the Android 9 release, Android Pie, that kind of went under the radar. Tucked away down in the bottom of the mailing list announcement

was a note that reads like this. I also want to take a moment to introduce myself as the new tech lead and manager for AOSP. My name is Jeff Bailey, and I've been involved in the open source community for more than two decades. Since I joined the Android team a few months ago, I've been learning how we do things and getting an understanding of how we could work better with the community. This is a new lead

for the Android open source project. This is coming from opensource.googleblog.com. As Jeff notes in his introduction, he has a history in the free and open source software world. He's been an avid user, contributor, and maintainer. Apparently, he has been a Ubuntu core developer, as well as a Google open source team member, and he's done work on Debian as well. He also worked for a company

called Savannah, which for many years distributed some GNU software. And now he is in charge of the Android open source project. Seems pretty credentialed. It's interesting to see some change in momentum over there in an ecosystem I perceive as maybe a little stagnant at times. Mm-hmm, mm-hmm.

And I find it kind of noteworthy that this was just unnoticed for about a week until I was digging through this, and I go, wait a minute here. Let me look into this. And here it is. I didn't I was digging through this. And I go, wait a minute here. Let me look into this. And here it is. I didn't see any headlines about this. And we now have a, that's, I wonder what, I wonder what might change.

Because he says specifically, he's looking for ideas from the community. He says he wants to hear from you. If you'd like to tweet him your ideas, at Jeff Bailey, B-A-I-L-E-Y, A-O-S-P on Twitter. At Jeff Bailey, A-O-I-L-E-Y, A-O-S-P on Twitter. At Jeff Bailey, A-O-S-P on Twitter. Or you can email him, Jeff Bailey plus A-O-S-P at Google.com. That's pretty direct.

That is giving you the information right there. That's pretty direct. So anyways, check him out. Maybe Jeff wants to come on the old show here. Come on the old Unplugged program. Yeah, tell us about what changes he wants to see made, what his imperatives are. And what is actually

valuable community feedback at that position? Yeah. People looking for feedback right now over at KDE Neon, the maintainers have begun publishing preview images of Neon based on

Ubuntu 18.04. And this is my favorite don't call it a distribution to distribution right now because you get that classic stability of an Ubuntu LTS that just over the years just gets iterated on top of with a rolling user environment built around the latest Plasma goodness. And it's been running pretty well for me on a couple

systems. On my production systems, I'm now on Kubuntu. But I've been wondering how long until Neon was going to make this transition, because this is the first major transition like this Neon has had to make since they existed as a project. For their entire existence so far, they've been based around 1604. And this is the first time the project is dealing with a transition to a new base. Wimpy, do you have any thoughts on that?

I'm sure this will be absolutely fine, because Jonathan Riddell was maintaining Kubuntu for years and years and years. So this is not the first time he's been through this and knows how to handle LTS to LTS upgrades. And they are taking it pretty calculated. They've been in alpha for a while. They've been expanding the testing circle and now they're expanding it once more. And if this works out, they think the end of August

will be when they officially announce the transition to Bionic. Exciting. Fantastic. Yeah. Can't wait to see it. I'm really keen to see

this 1804 release from them. Popey's been at Academy over the weekend, so I'm keen to hear from him tomorrow everything he learned and found out about whilst he was over there. Yeah, I agree. I've been following some of his progress on Twitter too. That's been very

fascinating. William Wold, I think is how you say it. He's a canonical employee who just posted over on the community.ubuntu.com site about some big improvements to Mir. to Mir. Yeah. Including support for the XDG shell protocol, which has just landed in Mir, and it will ship in the next release. This is

bigger than Mir, though, because this represents a stability in the Wayland ecosystem. Yeah, I just used that term. It is actually becoming an ecosystem, and standards are emerging. This is really good for end users, and Mir is plugging right into this. This is from Williams Post.

He writes, The switch of both clients and servers to support this protocol is an indication of the Wayland ecosystem, the fact that it is maturing. Because the protocol is marked as stable, future revisions will not break backwards compatibility. That's huge. And for display servers, implementing the protocol, this means far less code and easier

maintenance is required to support all clients, old and new. In addition to supporting the latest clients, XDG shell stable is a requirement to implement the new protocol, LayerShell. LayerShell is written by the Sway and WL Roots folks, and it gives clients a flexible method for drawing non-Window services,

such as taskbars, notifications, and lock screens are just a few examples that are now possible. We have not yet begun work on the LayerShell support, but plan in the near future. Wow, this is exciting. It is nice to see Mir going, isn't it? I mean, you know, it's all kind of happening in the background, but it's also neat to see, you know see the amount that the people over making Sway

and those other things, it really is becoming an ecosystem with different options out there, and you don't just have to be running one of the big major distros anymore. When the Ubuntu community hub was established and the MIR team started using that to seek input in the directions that they were going to be taking the project. The Sway developers were some of the first people to come and join the discussion there. So there's been good collaboration between the MIR team and the Sway team and also the Weyland developers. Yeah, and it's important to emphasize that, you know, mere is uh an active project and it has a lot

of momentum behind it uh william's a a recent uh member a canonical very talented developer and i think you'll be hearing some more interesting news from him in in the weeks and months to come yeah i noticed that i would be interested to chat with william in the future it seems uh yeah it seems like he joined a few months ago. Yeah, he joined around the start of the year. I met him earlier this year. And he came from the Arch community. He's the second guy I've met recently who's been recruited into Canonical from the Arch Linux community.

Poaching some of the best. I got a question for Wes and for Brent and for you too, Wimpy. So I'll start with you, Wes. Say you're going to sit down at a computer. How do you know if that computer is fast enough for your work? Well, this is on Linux Unplugged, so is it compiling the Linux kernel? No, I think for you personally, like your workload, like how do you assess a computer's potential?

Say somebody gives you a laptop. Like this happened to Brent when Brent was out here during LinuxFest. I just gave him the Librem 15. I'm like, here, do your photo work, Brent. And how do you know? Good luck, right? Yeah. How did you know if that computer, I mean, other than it's a fairly modern computer,

how would you know just looking at it if that computer was capable enough of being fast enough for your workload? Yeah, I think for me in that specific case, it certainly helped that I was surrounded by a bunch of people that I knew wouldn't put up with slow machines. So it turned out that for my workflow, it was fine, especially that I didn't have a laptop while I was in the U.S. So any laptop is going to do fine. But to answer the question a bit more broadly, I think for me, there's a set of work loads that I do typically on any computer, you know, multiple times a week. And so if someone's going to give me a laptop,

I'm going to do a set of workflows that I'm so familiar with. I know just sort of intuitively how long they should take or how they should perform or how snappy they should be. And, you know, that workload is going to be a little bit different for everybody. So I'd be interested in hearing everyone else's thoughts or methods. But that for me is tried and tested. That makes sense. It's a true way of doing it.

So Wimpy, you and I have experienced this in another context as well, and that is getting review systems in for the Ubuntu podcast or for this show, and you have to sort of assess this machine's capability. So do you have a couple of go-to methods when you sit down at a Linux box to try to assess, okay, right, how fast is this system? When we do our reviews on the Ubuntu podcast, we usually revert to um benchmarks in in popular games as the means to determine how fast it is most of the feedback we got from our listeners when we were

figuring out how to review hardware the things that they were most interested in was um what's the battery endurance like how hot does it get how noisy is it what's the quality of the screen how good is the keyboard um so much more practical things about you know the portability and and utility of the device and performance doesn't come up very often as a thing that they're interested in so we have a few games that we run through as an indication as to what the performance is like right personally when i assess the performance of a device, it really comes down to compiling and building software. So when I'm doing work stuff, it's building snaps. And they're usually large snaps because I work on some complicated stuff

that I don't get to talk about here. But also, when I'm compiling and building packages for the Mate desktop, I know how long those things take, and those are the things that I would say are the most CPU-intensive things that I do. Yeah, okay. I think that's a good way to put it. If you have a workflow that you're really familiar with, like the pace of it, because it's just something you've done a lot, and then you sit down on a machine and you go through it, you do get a good sense.

I saw the chat room was throwing things out, like how much RAM does it have, the CPU model. Those are like what I used to use. Oh, does it have 32 gigs of RAM? How many Electron apps can you open at one time? Yeah, exactly. That's my benchmark. Yeah, exactly.

Well, and that's part of the RAM thing. But it feels like modern systems are a little more complicated than that because especially with laptops, you've got thermal throttling to consider. You've got your throttling to consider you've got your base clock plus turbo depending on if it's something that pings one core versus all the four cores uh so there's right and like there's also responsiveness versus throughput and what's

more important for how you're going to use that system i bring all this up because this week blender has reintroduced and i say reintroduced because this, if you have used Blender in the past, was something that was available in the past, but they removed it. It's back now in a big way. It's a benchmark, the Blender benchmark. They've built the Blender benchmark platform with a maximum focus on transparency and privacy, they say. They use only free and open source software, GNU GPL software.

Everything that is all testing content is public domain. So it's all licensed public domain. And the test results are being shared as anonymized public data. So this is really exciting. And I wanted to use this as another opportunity to point people over towards what I think is Michael Larbel's really good work at openbenchmarking.org. The open benchmark tool, which is available in your distros repository, is amazing. You download this package and you install it. And then when you run it on the command line, there's GUIs available as well. It comes up and says, what do you want to test?

And it's everything. It's everything from it will download a game in the background, run it, benchmark it. It's everything from it will download a game in the background, run it, benchmark it. It's everything from it will download different versions of the Linux kernel and with different compiling options and benchmark them and generate the same charts. It generates all those charts that you see at Pharonix are all created using the open benchmark tool. The other thing it does is it uploads the benchmark results to a community comparison site where there's like 23 million different benchmarks that have been posted that you can compare your system against. Now only like 1.6 million of them are public. You can choose to

keep them private and just compare against your own systems. It's so fun to just run a system through this and really give it like just a beating. And like he's stayed on top of getting things in here. Like Linux 4.18 kernel benchmarks are already in there. So you could, even if you don't want to run 4.18, you don't want to switch it out on your distro, you could use OpenBenchmark to download, build it, and do some benchmarking. Like that's pretty neat. There's also Blender tests in there. There is entire workflows from photo workflows to development workflows to gaming.

And it's a great way to just download this on your box, run your system through a set of paces, and get real numbers out of it. I like that you have things you can actually compare to and, you know, a certain amount of standardization there. So you can know that other people are running this, you can run it on the same thing, and you don't have to be an expert in trying to set it up to have a good suite of tools there. Yeah, and you know, it's a nice thing because it's constantly updated, so they're constantly changing it. And you don't necessarily know

every test to run. Say you want to stress test the I.O. subsystem. Like, you want to just bash on the disk subsystem and the network subsystem and you really want to stress those two things, you could put something together or you could use scripts they've already created that will run it through its paces.

And it's just like that at every level, including encryption benchmarks, packaging. It's just great. It's just great. And if you're ever wondering how to really kind of get an assessment out of a system, take a look at it and run it against that. I'm almost wondering if I could look into it. Maybe I should try to figure this out. If there's a set of numbers we could generate it so we could compare different laptops over the years as we

review them to see how they perform. So we could have like a chart of the fastest laptops. It seems like we could definitely do that. Yeah. That'd be pretty good. Kind of like a leader chart, you know, that we could have a chart of the fastest laptops. It seems like we could definitely do that. Yeah. That'd be pretty good. Kind of like a leader chart that we could use that we would always run the same standard open benchmark on. The results would be public, so people could audit them if they wanted. Maybe a link back to the episode so you could hear the review of the unit.

They could benchmark their own systems against it. That might be pretty fun. It's like doing laps around a track with cars. Jupiter benchmarks. Anyways, moving on. I thought that was pretty neat and I wanted to give them a plug because it's a great way to see what your system is capable of or get a sense of a new box, which I'll be doing maybe in the near future.

A couple of mentions of some events that are coming up. First up is the Libre Application Summit, poorly named LAS. I'm just going to let that hang for a second. Why do I got to take that? Why do I got to take LAS? Why do I got to do that to me? That's an affront. I mean, you gave it up, man.

You walked away. Actually, they took it before the show was even over. Oh, okay. Never mind. That's just. You know, the bell's getting popular. A lot of dings today. But the nice thing is they just posted their early event calendar, so you can

get an idea of what's going on at the Libre Application Summit. And I think Eric, the IT guy, is going to make it down there. You think it's a lock, Eric? Yeah, it is a definite thing. I've got my ticket. I'm volunteering with the GNOME engagement team and even going to shack up with one of the System76 guys who's got an apartment not too far from there. Nice. You can give a high five to Daniel Foray, too. He'll be there doing a talk about the Elementary App Center. I'm hoping to catch both his and Cassidy's talks. Oh, yeah, Cassidy's going to be there, too.

Oh, super exciting. Nice. All right, so you'll be our Libre Application Summit correspondent. Report back with what you discover, all right? I would choose to accept that mission. Good man. It's not an impossible mission. In right? I would choose to accept that mission. Good man. It's not an impossible mission. In fact, it's the best kind of mission. So, I think

this gentleman's name we're going to talk about now is Beno? Beno? B-E-N-O. Beno. Yeah, okay. Beno had a talk at BSDCAN 2018 that was a few weeks ago. It was back in June, and it was titled The Tragedy of System D.

And he says, System D, to put it mildly, is controversial. Depending on who you ask, it's either a complete violation of the eunuch's philosophy, a bloated pile of bugs, a complete violation of elegant simplicity, or all of the above. And he starts a talk that really got my attention. I'm going to play, it's a 30-minute talk. I'll play about three minutes of it. But I will link it in the show notes where you can download the slides

and you can watch the entire talk. But it started really, it caught my attention. I decided to share it with the class here because he brings up something that is not even related to Linux. It's not related to systemd. I think it's something we all do as tribal animals. We do something to sort of show our allegiance to a

particular tribe or group that brings down other people. And I think we're probably all guilty of it. All right. So I would just like to initially thank everyone on BSDCAN for inviting me to speak in front of all of you lovely people. And in full understanding and acknowledgement of that, thanks. I would like to present a live reading of Bedazzled by Blockchain by Michael... Sorry, no. George Neville Neal could not be here, and it wouldn't feel right doing this without him, which is a tragedy. Link. So I came up with the name of this talk around the same time that I came up with the first sort of rough idea of it. And it seemed to fit really well, especially when I went to Wikipedia and saw that tragedy is a form of drama based on human suffering that invokes an accompanying catharsis or pleasure

in audiences. And I figured there was plenty of drama and human suffering to go around in System D. And so I figured that it was a good way to go. Another thing that really came into it when I was developing this talk is this piece by Oren Shaw called Contempt Culture. It's a really interesting piece and it's kind of confronting sometimes because in a lot of the communities that we're in, part of the way that we show that we're in the community is by heaping shit on the other communities. And that kind of sucks. You know, your Python developers heap crap on the Perl developers and the Perl developers heap crap on the PHP developers. This piece is very focused on language stuff, but we kind of do that too.

And so I highly recommend that everyone reads this. And the other thing that I think goes into this discussion is the notion of change. Change threatens a bunch of things, the most obvious being familiarity. And familiarity feels really comfortable because it's the thing that we're used to, but occasionally it's really good to reassess what the familiar and decide, you know,

whether we need to change it or not. I felt his comment there about how we sometimes as a group will throw shade on another group to sort of show that we're in, we're in with you, like I think the way you think. I think that's a really common behavior. And what he's talking about, obviously, is how the BSD camp has been really disparaging towards SystemD.

That's really what he's addressing in this talk. In fact, he starts to, over the next 15 minutes, he starts to sort of set the audience up. And mind you, this is a room full of free BSD developers. The graybeards of the graybeards. In fact, in one of these clips, you'll hear Alan Jude in the background if you listen really carefully. And there's some coughing and some hacking because some of them are a little bit older. And he's trying to convince them that they're already on this path. He starts to go through the history of Unix

and shows how the very early init tools have now morphed from just getting the system up to mounting file systems, starting network interfaces, and doing much, much more. And after about this point, he starts to mention that these complex tools are beginning to create the need for a system layer in FreeBSD. So you have the user layer, the system layer, and the kernel layer.

SystemD essentially introduced a system layer to Linux. User space, system space, kernel space. And he says we may need something similar in FreeBSD. In fact, we're already kind of on our way. But the thing is that dynamic stuff is often better managed through user space than it is through kernel code. And I think that Windows natively understood this from the beginning. Mac OS definitely understands that with SystemD,

you actually understand that there's a layer in there in the middle called the system layer, which I'm calling the system layer. And so system D implements that system layer for Linux by bringing in a bunch of code that manages various system functions like network connectivity, time, device management, and all that kind of things that just don't necessarily belong in standard user space, but also don't necessarily belong in the kernel. And the thing is that on top of that, it doesn't really do it in a way that everyone finds familiar, which I think is part of the source of the angst to it. But it does give you an explanation as to why it sees the need to bring in things like network manager code,

as to why it sees the need to bring in things like network manager code, UDEV, time, resolver libraries, because all of those things are systemic services that are getting provided to user space. So how did that work out? And then he goes through and shows the history of SystemD adoption. He dives in a little bit to Lenart's history with, it turns out, Lenart Pottering, the guy that created SystemD and was its primary advocate, was at odds with management. Management

wasn't all that big on it. They had made their choice. They were perfectly happy with how it was functioning in RHEL and didn't really see the point in systemd. And he said, no, no, trust me, let me show you. And despite pushback from his own management, he continued to work on it, which was an interesting part of this clip. And then he starts to get into some of the most common complaints about systemd. Like one was just mentioned in the chat room, it runs as PID1. And he has a response to that. But the number one complaint systemd gets, especially from the FreeBSD camp,

because remember the audience here, is it's just not damn Unix-like. Moving through the rest of this section, I kind of want to look at some of the complaints that people have about SystemD and just sort of address them a bit. Number one, there's a lot of suggestions that it violates the Unix philosophy, which I usually take to mean that you want to write software that does one thing and does it well and then connect it to other things. SystemD as a project contains a lot of things. SystemD as a daemon is a thing that reacts to events and starts things and does it very well. And so you could claim that it does not actually violate the Unix philosophy. You could claim that there's a bit of violation

and that it's bringing all of this extra functionality into the project, but I think for BSD projects to criticize another project for bringing a bunch of tangentially linked software into one repository to manage it collectively, that's a bit rich. I think that might be one of my favorite parts right there. It is a bit rich. And he says there is also the issue of Lenart's personality.

It is something that seems to have triggered a bit of the reaction. I think that's fair. I mean, it's definitely the communication aspect and how you go about that. It's a big part of growing and trying to get adoption for a project like SystemD. Yeah, that's it. So, open source communities are fun places. Ask me how I know. Is it July 4th yet? So Leonard Puttering looks like an interesting guy in a bunch of ways. Lots of people don't like him and I can kind of see why.

He tends to wade into a place with a bunch of ideas that he holds very strongly and then somehow actually managed to get them implemented and delivered and in your face. And if you don't like them, it's very hard to get him to change his mind. The thing is that I've just described a whole lot of people in open source. Yeah. One thing I think you have to admire about him is that he does actually get this stuff done. I mean, for someone to show up to his manager and say,

hi, I want to rewrite init, and I think we should do it like this, and for them to go, no. And for him to go, no, no, no, just wait. And to actually go do it, and to have it be so successful that it's spread outside his own project into a whole bunch of other major Linux projects. I think that actually is quite impressive. The other thing that I saw is that there are a bunch of cases where he's dealing with bugs where he does actually remain

incredibly polite given the amount of stuff that's being thrown at him. There's a famous bug where it looked like SystemD was mis-parsing a username that started with a digit and using the digit, which was zero, as a user ID. And it turned out that it was actually not being read as a... It was being interpreted as an invalid username, and there's a long discussion about the validity of usernames. And while it does get quite rule-lawyery at points, he never goes into name calling.

He's never calling people idiots. And that, when you're facing that kind of stuff, is quite admirable given the amount of things like death threats he's got over this, which is not cool. I'm wondering, Wimpy, if you have any reaction to his comments about the typical open source personality there and about how sometimes that causes some of the reaction to their proposals this is a really well thought well-reasoned presentation um i'm gonna i'm gonna go and find the full the full thing and have a listen to this later i will have in the show notes for sure yeah yeah what what a great speaker on that point yes i mean there are always stereotypes in communities right and definitely those people exist and anyone who's a leader um in any sphere including open source can in some respects be put on a pedestal and held up uh and and celebrated by sections of the community.

But if you have strong beliefs and you want to set a project in a particular direction and you have conviction behind that, there will be another group of people who don't subscribe to your philosophy, who will resist what you're trying to do, and that will manifest itself in lots of different ways. But I don't know necessarily that that's open source. I think that's people. Behavior.

I think the difference is, yes, it's human nature. And I think one of the things is because open source and free software, all of this stuff plays out in the public domain, you know, the public sphere, then we see all of this interaction happen and all of the fallout and all of the, you know, the drama that goes along with it. So, you know, we're just a microcosm of humanity in that respect. Yeah. Yeah. It's fascinating that it just, it infects everything, everything we touch. We just mess it up.

We have more access and we're maybe more passionate about it too. And we just like to talk. It's to me also a data point in where this conversation is now in the FreeBSD community. Because one of the things that I enjoy is if you can go back through the history of this show. In fact, I would really just be extremely flattered if somebody took the time to make a YouTube playlist

or sent me the links of all of the episodes from our past where we covered System D from essentially its initial idea to the end of the Debian drama and then Ubuntu's adoption. I would be fascinated to know what episode series that is because I would love to be able to make a playlist,

like a one-off feed. That was every episode in 2015. Yes. I mean, it was just, but my thought along the entire time was, well, what the hell are the FreeBSD guys going to do? Because I could see what SystemD was trying to solve. And so now this to me is a data point

of where they're at in this conversation. And it's taking them a while to get here. But he ends the talk on a couple of really strong points. And there's even more in the full talk. I'm not giving a full spoiler here. There's more good stuff in the talk. But he ends on a warning to FreeBSD users. Don't mock system D.

And don't recruit around systemd. Mocking systemd is the wrong attitude to take. It's not the kind of sad thing that Linux people have to deal with and how sorry we are for them that they have systemd. We should be looking at going, why did they choose to do that? And what I really find problematic is using system D as a recruiting tool for BSD. I don't think that that should be done at all. Because when you think about the kind of people that we would bring across,

if we were to say, come to BSD, we don't change. Or come to BSD, we don't have system D, oh, but we've just come up with a really good idea for something that works like it. I don't think that those are the kind of people that we, you know, we don't want to bring that attitude into our thing. And again, what I said we should be doing instead is asking, why did they see this as necessary? And the thing is that if you look at it like that, if you don't see it as, oh, those poor Linux people with their terrible init system and go, hang on, they just did that and I don't understand it. Why did they do that?

What do they see in it? That means that we can then look at that and start to find things that we can get out of it. The other thing to be careful of is the next generation of people, the people that come after us, they don't think the way that we think necessarily. A lot of the people who are now coming into IT and software engineering,

they're so much more used to APIs and not library APIs, remote procedure call APIs. They grok things like containers in ways that may seem quite unfamiliar to us. That doesn't mean that containers are bad, and it certainly doesn't mean that containers are something that should be mocked. I heard someone describe Kubernetes as the POSix of the cloud, and while that may sound ridiculous at first, I think it's a scary thought when you think about the fact that we can't run Kubernetes. I think that's true. Yeah, I hadn't thought about that, obviously, because it's not

my area, but that is a big thing. That whole point, don't mock systemd, don't recruit based on that, because if you bring in a whole user base that resists change then it's going to hurt the project. That's really insightful and true for all projects. The Linux community need to take a leaf out of this presentation.

The message that's delivered here about critical thinking about why a project or why a person or why a distribution has made the choices and decisions that they've chosen, rather than just shitting on it for the sake of shitting on it because you're not in the camp of that distribution, is a very, very positive message to take away from this. Within the Linux community, if you're not running Arch or you're not running, you know, Slackware,

or you're not running KDE Neon, you know, there's this group of people that will criticize your choices. And this really needs to stop because if we're linux users it really doesn't matter what distribution we've we've bought into we all agree on the fact that linux is for our purposes the best operating system available and we should not be throwing shit at one another over details you you know, because we look like a bunch of asshats from the outside. And it's utterly destructive because the amount of energy that goes into all of this negative, you know, product is pointless.

It would be far better to have, you know, if you don't agree with what a project's doing, as this guy's just said, ask yourself, well, why did they do that? What was the reasons that drove them to do this thing? Maybe try and understand where they're coming from, because there might be something in that. Maybe you don't believe in that particular technology, but as a technology direction, you can see some value in it. Yep. I completely agree. Perfectly said. I would love for this to be held up as an example of critical thinking in open source and how not to sling shit at one another because we really need to cut that nonsense out.

And we can all be better, right? If we can productively learn, they'll take some ideas that they like, add some new ideas, and then we can learn from that in the future. And it's a high-level perspective, and I think that's the perspective that gets the work done. And I loved it. I thought it was a great talk, so I wanted to share it with you guys, and we'll have it linked in the show notes,

linuxunplugged.com slash 262. And I'll tell you this, after listening to that, I had been on the fence. I've decided I'm going to go to MeetBSD in California, October 19th and the 20th. I'm going to be down there. It's going to be at Intel's campus. Oh.

How cool is that? Heck, yeah, yeah. You should consider coming with me. I bet you we could make that drive in one day. Yeah, I think we're up to that challenge. We could run down there. We could do a tech snap. Get some BSD on.

Get our BSD on, hang out with Alan Jude, record a tech snap, and then we could make it back up here. It's at the Santa Clara Intel campus, and it's going to be on October 19th and 20th. I'm going to go. Let's do it. I think, you know, after hearing that, I'm like, that is exactly the message. That is exactly the message we all need to be taking in, and it's that atmosphere, which is going to be conducive to open source development.

Plus, it'll be good to see Alan. Exactly. Yeah. Yeah. That was pretty great. And you can find links to all of that in the show notes. Now, I also want to talk about Dropbox this week and kind of just a cool temporary workaround,

even if you're not a Dropbox user, something that's just a neat trick you could do to hide stuff on your computer and all kinds of other things. So let's take a moment and thank Ting for making this episode possible. Linux.ting.com is where you go. When you go there, you'll get $25 off Ting service

if you bring a device or $25 off a phone. Ting is smarter than unlimited. If you use less, you pay less. The average Ting bill is just $23 per month per phone. It's just a fair price for however much you talk, text, and data you use. That's it. $6 for the line, which is great if you want to have a couple of family lines or a backup line,

or if you're a small business and you need 25 lines, it's great. And they have nationwide coverage, CDMA and GSM, no contracts, no quote-unquote agreements, nothing weird like that. And you're in control the entire time. You can always see your usage. You can change different individual aspects of your service, like just turn stuff off if you want. You can set usage alerts as well. And Ting's not in it to try to push an app store or a music streaming service. So they don't really have an agenda of how you use Ting. So they don't have to push down like a Ting experience with a Ting image with Ting

branding on your phone. And they don't really care if you just never want to use a text message, ever. If you do everything over WhatsApp or Telegram. You mean they'll just like, they'll send a text if I want and otherwise they don't care? They don't even bill you for it. How does that make sense? Wow. I know, I guess that's how it works in some other places in the world. We are so backwards.

Not here in the States. I wanted to point out, if you need a phone right now, and you don't have $1,000 to spend on a note, why not look at the GS5 Plus? Or, I'm sorry, the G5S Plus. It's the Moto G5S Plus for $254.

You would own the phone outright when you go to linux.ting.com, because it's normally $280. But when you go to our linux.ting.com URL, you get it for $254, unlocked, multi-network, 2 gigahertz octa-core Qualcomm Snapdragon processor, 5.5-inch IPS display, 3,000 milliamp battery, and it also has an SD slot, so you could put up to 128 gigs of microSD into it for $254. That's a capable phone right there.

No contract, nothing. And it's just a clean Android experience. And you just pay for what you use. Start by going to linux.ting.com. Oh! You know what I wish I did? I wish I had like a good, like an old classic Kyra clip. I feel like some classic Kyra, but I don't think I have any classic Kyra loaded.

I don't. That was just poor planning. I do have, look at this though. I have classic Brent loaded in here. That's interesting. That's good. I've got that set up. So if we ever need some classic Brent, I can always pull that in. That's nice. That's good. What else do I have in here? Of course, I've got Richard, but I digress. He's taking a nap right now.

Now, let's take a moment and talk about Linux Academy because today is a huge day for Linux Academy. I'll get there in a moment, but start going to Linux, go to linuxacademy.com slash unplugged and sign up for a free seven-day trial. It's a platform to learn everything about Linux and all of the stuff that runs on top of Linux that you want on your resume.

And it's a great time to sign up because they've been adding a ton of content, really working like crazy on this stuff that runs on top of Linux that you want on your resume. And it's a great time to sign up because they've been adding a ton of content. Really working like crazy on this stuff. And there's things they have in the works that I'm not even supposed to tell you about. I know. I can't help it. So I'll tell you about something that's a little public. There's no getting around it.

Today, Forbes released their fastest growing startups. And it's like, you know, at a five, so they narrow this list down to 5,000 companies in the United States that are just growing like crazy. And Linux Academy is on the list. Whoa. They're on Forbes' 5,000. And out of 5,000, Linux Academy is ranked 78.

And I don't know what the period of growth is here because they don't tell you the scale. So you just have to guess. All right, Wes, you ready? And Brent, if you have a guess too, you are welcome to chime in because I know I teased you in the pre-show. Can you guess the growth percentage for whatever it is?

I don't know, a year since they started. I don't know because Forbes doesn't say. So we're just, it's who cares, but it's just for fun. Can you guess the growth percentage of Linux Academy as listed on the Forbes top 5,000 fastest growing companies? 200%. 200% is pretty good. That's pretty good.

That is pretty, pretty bold. That is pretty bold. Does anybody in the mumble room, Brent, do you have a guess? What's your guess on Linux Academy's growth number? Well, what's the, they're on the list of total how many? This is Forbes' 5,000 fastest growing companies in America. And they're number 78 on the list. What do you think their growth percentage has been?

For whatever the arbitrary amount of time is. I'm going to go, they're number 78, so I'm going to go 178. Oh, okay. So Wes is 200-something, 178. Wimpy, do you have a guess? Eric, do you have a guess what Linux Academy's number is? It's 250%. 250? All right. This is like prices right here. What do you think, Wimpy? 300%. Those are all really good, bold guesses. Probably not even, I wouldn't have gone that high.

I wouldn't have gone that high. The official Forbes Inc. 500 fastest growing companies in America growth percentage is 4,653.64%. Whoa. What? And I actually think that's based on a low guess of what the revenue is, but that's pretty amazing. That is pretty amazing. 4,000%. And they're teaching about Linux. There's a reason why they're growing like crazy because let's be honest, one of the hardest things about the IT industry

is the fact that it's changing faster than just about any other industry in the world. It's sort of the deal with the devil we all accepted when we got into IT. We all knew that was the case. And Linux Academy is kind of a tool to help keep you current. It's really due diligence when you're in IT is keeping your skill set fresh. And I think that's why they're growing like crazy, is when you look at the explosion of cloud-based services like Azure and OpenStack and AWS and

people that are rolling in their on-premises infrastructure and how everything is based around Linux, it makes sense. They are the go-to platform. So go there, linuxacademy.com slash unplugged, sign up for a free seven-day trial and the upshot of that is they look good on the resume now. You know, that's like a recognizable brand name

now. You went to Linux Academy. Oh, I've heard of them. Yeah, they're on Forbes' list. Like, that's good for Linux Academy. So congratulations to them for making the list too. That's pretty great. And I want to take a moment and tell you how you can get a $100 credit at DigitalOcean. So go to do.co slash unplugged. Wes goes there every single day. He signs up with a new email address every single time. Actually, he doesn't. He uses the same email address, so he never gets the credit. I mean, I already had it open. It was one of my tabs. You actually did. No joke. It has his dashboard up on there right now. How could you not? It's so great.

Digital Ocean's great for spinning up systems that you want to try something out for a couple of minutes or something that you want to put in production for the next few years. I literally have droplets that have been running for four plus years. It's great. It's my data center on demand, and it can be yours too. And you can get a $100 credit that lasts 60 days when you go to do.co slash unplugged. Super nice, fast systems, all running Linux. KVM is the virtualizer. They have enterprise-grade SSDs. They have cloud-based firewalls so you can block traffic that never even hits your rig. Easy to use DNS management so you can set up your domains to point to your droplets,

40 gigabyte e-hypervisor connections, and data centers all over the world. And then you top that off with a great dashboard and an easy-to-use API. And you're like, okay, good enough, good to know. Yeah, here's a couple of the other things. You can get a great system

for three cents an hour. Four gigs of RAM, two CPUs, 80 gigabyte SSDs, which is an enterprise-grade SSD. It's crazy fast. Three terabytes of transfer, three cents an hour. 4 gigs of RAM, 2 CPUs, 80 gigabyte SSDs, which is an enterprise-grade SSD. It's crazy fast. 3 terabytes of transfer. 3 cents an hour. That's great. And then you're like, okay, Chris, that's enough. I got it. I'm like, well, yeah, but hold on. There's one more thing.

They also have the best documentation in the industry. On top of one-click deployments of things like GitLab and WordPress and Ghost and many more, that you just hit one button and you get, if you had been setting up a WordPress stack as a career, your entire career, when you're like, damn, they set it up exactly like I would. And I say that as somebody who has been one of the biggest advocates of

setting it up yourself so you know how it works, you know, going through Arch, going through Gen2 so you know how it works, building a system so you understand how it works. I now do these one-click deployments because I've gone through and deconstructed how do they set it up? What repos are they using? What versions do they have installed? What are the defaults in the config files? And it's all the stuff I just would have done.

They really know what they're doing over there. I mean, once you've learned how to do it, then they've got it for you, and you can trust that they know what they're doing too, so it's reliable. You can just deploy a base system and learn too. That's what's great about DigitalOcean, especially when you can get a $100 credit by going to do.co.unplugged. And a big thank you to DigitalOcean for sponsoring the Unplugged program.

do.co.unplugged. All right, so we covered this on Linux Action News 66 this week. Dropbox is stopping syncing on any file system other than unencrypted extended 4 on November 7th. After 10 years of using Dropbox, they've finally done something that's going to force me to abandon Dropbox. Chris, I just switched to BcashFS. This is terrible news. Yeah, or ButterFS or ZFS. It is a real shame.

And the explanation from Dropbox is a clown show. They say a supportive file system is required as Dropbox relies on extended attributes to identify files in the Dropbox folder and keep them in sync. We will keep supporting only the most common file systems that support extended attributes so we can ensure stability and a consistent experience. Well, that's the biggest chicken shit reason that they could possibly give because the file systems they're not going to be supporting have fantastic support for extended attributes.

It's the most ridiculous on-its-face answer they could possibly give. And there's been some speculation that it actually comes down to how they try to hide your local encryption key and the fact that they can't do it the same way they do it on Extended 4 on XFS and things like that. So there's various speculation to what the real reason is, but that's the chicken shit answer

that they've given us, and it has me upset. And I want to take a moment not to defend Dropbox or encourage anyone to use the service because I am switching off. That's just a foregone conclusion. I have to. First of all, I'm paying for the business version, which is $1,000 a year.

Yeah, okay. Well, if you're paying, I mean, then that does change the value proposition. Hell if I'm going to spend $1,000 a year on something that doesn't work the way I want. Here's the reality, though, is I have two terabytes in Dropbox. I've been using it for 10 years. Everybody who's ever worked for JB in a media production capacity has all of their work in Dropbox. Every clip for Unfilter is stored in Dropbox. One of the great features about Dropbox is they pay for the disk. They manage the disk. I buy a subscription plan, and they add the disk.

And it works out to be cheaper than it would be to buy the same amount of disk on a VPS. And it even works out to be cheaper if I were to buy all of the disk and set it up. It still works out to be a cheaper price if you consider that I also still need syncing. I would like revision copies so I could do restorations. I would like a web UI so that way I can view the files when I'm remote. All of these things, the cost ends up being essentially... Right, you need like a business class reliable interface to your files. Yeah, all of a sudden $1,000 is not that much. It's really not when you're considering,

okay, so if I'm using two terabytes today, I probably want at least a minimum of five terabytes of space. I'm going to want that to be in a RAID array. I'm going to want that to be in a box. I'm going to want that to be connected and I need to set that up. It's not that far away from $1,000, even if I use an existing PC I have. Yeah, and you only have so much time in a week to be setting up more and more free NAS boxes. And then fundamentally, there is a huge advantage to having a local client. So Google Drive and things like that are not particularly appealing because when you have a software daemon running on your system syncing folders, what you get in essence is a network file system that's always available offline. And I can save a file to one location and it can be tilde slash Dropbox slash music and it is the

same file path on every system I sit down on. And I can use config files that way. I can use music files that way. We can use music files that way. We can do media production that way. I can export to a folder after we're done recording the show in Dropbox, and it syncs up to a cloud system that Joe can log in through a web browser and download it. And it's super, super convenient

because it creates an omnipresent network file system that is fully available offline that can sync the changes up once you connect. And it does it in a way that is simple because you can install a client, you log in with a username and password, and it begins immediately syncing. You need to forget about it. The additional benefit to Dropbox

is it supports something called selective sync. So I mentioned I have like two terabytes in Dropbox. That's a lot. There's no system where I would sync down more than 50 gigs of that. Because so much of it is historical stuff. Old assets for shows that we don't have anymore, but we still want to keep around. Clips that I don't use anymore. Interviews we've played on the show. I don't need that on a current system. And you probably have many machines that just can't store that much anyway. Yeah, exactly. They're on SSDs these days. Dropbox has

a sophisticated selective sync system. It's simple, but it works really well. And so I can use the storage. I can use the five terabytes I have allocated in Dropbox to hold stuff that I don't have on my local laptops. It is also extremely useful when I'm traveling to be able to pull up like an audio file or even a video and play it in the Dropbox streaming player in the web player, because I only maybe need to listen to the first 10 seconds of a file.

I don't need to download the entire file to play that 10 seconds. And if I'm on a mobile connection, that's a massive difference for me. So Dropbox has been a very functional tool that has worked with people that are remote, that are on Windows, that are on Macs, that are maybe somebody that's working to do art for us for a little bit that isn't using Linux, that wouldn't... As long as they have a web browser, they can play in your space and yet have access. Yeah, or there is a client available for their platform. And that's, you know, both have been very beneficial. You know, like, for example, back when we used to edit video on Linux, I mean, I'm sorry,

on the Mac, and when... The dark days. When Rikai was our video editor, and he would all sync, I mean, I'm sorry, on the Mac. The dark days. When Rikai was our video editor, it would sync to a Mac, and he would use it on the Final Cut because that was our production pipeline back then. And it meant that I could be sitting down here in the studio on Linux,

and the editor could be on a Mac, and there was no concern about cross-platform compatibility or anything like that. And not that that is unheard of, but keep in mind, I've been, I've been using Dropbox since like 2008. So that was a little more uncommon. There was no NextCloud or OwnCloud. There, you know, there wasn't a BitTorrent sync back then. So we're talking 10 years ago and there's how many tools do you, how many software tools do you use today that you were using just as heavily 10 years ago?

Yeah, like three. I know. So that's where Dropbox fits in our workflow. It's not irreplaceable, but it is a big consideration because then it's not just a matter of switching out of technology. The next question becomes, how do I get all of the remote people to install something else? How do I get all of them to buy in on this? You know, we're talking nearly half a dozen people that are all over the world.

Right. There's also that social factor of Dropbox, because it's been around so long, has a lot of investment already. People already have clients installed. It's kind of like how Slack ends up. And it's not unusual if you see a public Dropbox URL when somebody's sending you a file. Hey, can you send me that sound clip and I send them a Dropbox file? That's a link that's not unusual. It's a normal thing. So there's that aspect of it. So you have the human aspect of people adopting it.

But you also then have to ask yourself, like, well, if I'm going to move off of Dropbox, is it time now to go through and clean some of this up? Should really some of this be on our free NAS? And some of this doesn't need to be in a sync system anymore? Should we reorganize the folder structure? Because, you know, it turns out we're no longer producing Castablasta.

You know, I'm no longer doing plan B. But yet all of that stuff is still in the Dropbox directory structure in the current working shows stuff. Oh, up at the top level. Yeah. So is it time to go through and refocus that onto what we're currently

doing and the assets we currently need but that becomes a massive project that's a huge project that's going through a decade's legacy of production and deciding what is current and what needs to be resorted and it's just a massive undertaking that makes that november 7th deadline a little uncomfortable you You see where I'm going with this. So the question became, how can we punt that down the road just a little bit? Not forever, because this is something we absolutely have to address. But if we don't get all of these questions answered, because I would hate to move to a new system and use a terabyte more of storage than I need. Like we just wrapped up production of the unfilter show.

Do I now need all of the Unfilter show. Do I now need all of the Unfilter clips in Dropbox? Probably not. Probably not, but where else do I put them? Yep. And maybe they go on the free NAS. But then I need to slice out space on that, right? Because it's like a terabyte worth of clips. And then those are disks that you manage.

Yeah. And so you lose the convenience there. So this is why we wanted to talk about ways to sort of kick the can for a little bit. And the first idea that we kind of came up with behind the scenes here was, well, what about just creating like a sparse file on your file system that was formatted as extended for? What do you think of this idea? Create a sparse image, which would only use as much as it's actually using, right?

That's the nice thing about sparse images. So if you have a 500 megabyte Dropbox, which would only use as much as it's actually using, right? That's the nice thing about sparse images. So if you have a 500 megabyte Dropbox, it would only use 500 megs, but you could let it go up to like 2 gigs or whatever. And then you mount that, and then you set your Dropbox inside there. So it'd have to be mounted to boot up. What do you think of this idea? I mean, you can just put it right in your

old F-stab there. Once you've got it formatted, maybe just stick it somewhere in a convenient place that you know or a secure... Do you really call it fstab? Well, no, but it's more fun. So, yes. That is. Yeah. So, look at the truncate command. So, you could do truncate-s 2 gigs and then you could do something like dropbox.ext4 and then you use makefs.ext4 to format that and then create a directory inside there and you can mount that. It's a simple pseudo-mount Dropbox.ext4 to say, like, colon backslash Dropbox, and you would have a Dropbox folder that would show up

that would actually be a mounted sparse image. And that isn't that difficult to do. You just move your Dropbox to there. You keep on syncing. It's living on your XFS partition. Yeah, and then if you put it in FSTAB, then, yeah, it'll just happen for you, or, you know, SystemD on syncing, it's living on your XFS partition. Yeah, and then if you put it in FSTAB, then yeah, it'll just happen for you, or SystemD can do that too, speaking of.

Lots of options. It's not ideal, right? You just have another layer just to make the Dropbox client happy. Yeah. But here's the real rub, is it's worth keeping in mind that it's also encrypted extended for partitions. Now, if you're doing whole disk encryption, the entire disk is encrypted. That's getting unlocked at boot time. That is sort of

beyond the application layer that's being done at the kernel layer. I don't think that's going to affect Dropbox. They actually haven't been clear about this. But if you're doing like just the encryption of your home directory, that's where you're going to get screwed. And there are several distributions that do that by default, that type of encryption. And that's where, even if you're on extended four, you're getting screwed over. You know, if you're in a lot of enterprises have a policy, have an encrypted home directory. And now it just, you're like in a hard, between a rock and a hard place because your work demands that you have an encrypted home directory, but this tool you use to do work won't work anymore. So this could be a solution for you. Eric or Wimpy, do either of you have a particular idea of a workaround that comes to mind

for keeping Dropbox happy? We even speculated about ways of fooling to think it's talking to Extended 4 regardless of the file system it's on with a few little library hackeries. But I'm curious if either of you can think of a way to keep Dropbox chumming, even if it's, say, on ZFS or XFS. Yeah, I like the idea of sparse files. You could use, you know, a QCAL file, you know, something that's common to the virtualized environments

and format it in a similar way so that it grows, you know, with the requirement of the the storage that you're putting into it um so yeah i like that um i was i was looking so you talked about this at the weekend and i was i went back and looked at sync thing to see how sync thing has been evolving in in recent years and it's interesting a recent release of sync thing so 0.14 but 0.47 actually um supports file system monitoring with within itself now um and doesn't rely on extended attributes so that is a recent change there like um literally at the end of last last

month sort of june uh sort of June this year. So there's always alternatives out there. But yeah, lots of people like Dropbox for all the reasons that you mentioned, most of which is the selective sync capability. Yeah, it's nice. So yes, finding file system mechanisms. And you said about pseudo-mounting. Dropbox, if you're using it on a desktop operating system,

wants to use the Dropbox folder in your home directory. So there's no requirement for you to mount that with sudo. You'd be able to do that with a loopback mount. That is a fair point. I think SyncThing is extremely attractive. I've got a couple of candidates. I've obviously been running NextCloud instance for a long time now. And SyncThing

I've played with, I've dabbled with. The thing I really like about SyncThing is I could set up individual folders and just sync them to certain systems. So it'd be complicated, but it'd be really great. Fedora has a suggestion to just go the other direction. They say just convert the file system.

Use FSTransform. And they give a walkthrough on the Fedora magazine website. They say it sounds like magic, but you can convert from Extended 2, Extended 3, Extended 4, RiserFS, or XFS to another type in almost any combination as long as your kernel has support for that file system, the one you're coming from and the one you're going to. It does seem to work. I did just try it and converted an XFS system to EXT4. Shut up.

I only had one file in there, but it came through unscathed. That is a workaround, but I will stop using XFS when you remove it from my cold dead fingers because I am wedded to XFS. Yeah, it's a great file system. And that's why Red Hat Enterprise Linux sets it by default for the home partition. It is a tried and true system

that's been around since the 90s that it's still getting active development. It was one that I went to not only because it's super fast with small file systems, it keeps itself tidy, but it supported extended attributes before Extended 4 was even released. And so it just drives me crazy that that is Dropbox's explanation.

The extended attributes argument is weak, I think, because although there are differences in how the file systems store extended attributes and the block sizes in the inodes and what have you, that should be transparent through the libraries that they're using. Because Dropbox is implemented using Python ostensibly and LibRsync. Yeah, that's what I thought when I heard that. Or at least, I mean, we just don't have very much insider clarity here, but it would be helpful if they could lay it out a little bit more

and then at least make the argument of like, is this not possible because of some design decisions that you've made? They do have a fancy client, right? So simpler clients might not need this. Or is this just too much work and you don't want to spend the resources on Linux? And those are different scenarios.

I would really like them to be more clear about this. Instead of just one community support rep mentioning something in the form and then mentioning something later. And then we're left to speculate, other than you're going to start getting notifications via your Dropbox client. It seems it really rubs it in our nose, too. Mac gets two supported file systems, and we get one. And we're the one with all the file systems.

I just talk through the kernel. Don't talk directly to the file system. You're doing it wrong. That's my general sense. And so I've learned two lessons from this. And I feel like it's worth sharing with you guys. The first lesson I've learned is, you really, if you're on Linux and you're building something like an essential workflow,

you really have to go with free software whenever you can. Now, in the past, some of our production systems were Macs or Hackintoshes, and so Dropbox seemed perfectly reasonable. If it was 10 years ago and we were starting where we are at today with Linux, I don't think I would ever have used Dropbox. So, you know,

keep that in mind. For your workflow, some of these companies just are never properly incentivized to support Linux users in an appropriate way. They can't even give us

an appropriate response. They don't even have the time of day to give us an explanation of why they're screwing us. That's where we're at. I mean, it's that same thing, right? You get't, you get the convenience, but you don't get the control and free software gives you that back. Yeah. And so that's, that's, it's just a

reinforcement of that lesson. But at the same time, that, that, that scorn is tempered with the fact that I used this for 10 years. While I always knew this was going to be a problem, I got 10 years of practical use out of this. And I've only paid for the business account for like the last three, maybe two. So that's, you know, you have to weigh that. It did burn me in the end, but at the same time,

I managed to not have to manage that problem for 10, I managed not to manage that problem for 10 years. And you might have even had a change if you were doing, you know, one open source project goes out of favor, a new one has more features, you might have to move those files anyway. And I haven't been wasting this time. You know, I've been, you guys have heard me tell the story

about my next cloud instance that runs on a Fedora server on DigitalOcean. I could switch over to that. I bought their, you know, their attachable block storage additional space. So much block storage. I don't have five terabytes, but I've got a few gigs free. So, you know, I have options here, and I've got till November 7th, and I've got other options, thanks to how powerful Linux is,

to even kick this can further down the road if I have to. So I'm not even totally out of options. But it's a good lesson. I feel like there's a mix of a lesson here that is there is a practical aspect of this, but you have to stay vigilant. And I'm just slightly on the other side of vigilance. I'm almost vigilant, but I was really hoping, because truth is, I think we renewed just two months ago.

No. Yeah. And Noah and I have been talking about this for months. My intention was to really get off the sauce this year, you know, looking at the cost. That's why I'm actually familiar with the pricing versus

local, versus going VPS, versus having AltaSpeed host a C file instance. I knew what all of the pricing is because it's something we've been looking into for a while. Now, have you considered tape? We could just fill the garage here in the studio with tape. You joke.

I actually have seriously considered tape. I mean, it might be a good solution for those video assets that you don't need. You can get some serious storage on those tapes for a decent price, especially if you start buying some of those tapes in bulk off Amazon or get a drive off of eBay. And that stuff, those standards, those are the same standards

when I was a... I think of a TechSnap episode that would make. I know. You just put old shows on tape. Trust me. It has seriously... The only problem is

it seems really, really, really short-sighted to take a show that's fundamentally something that's supposed to be distributed on the internet and store the only copies

on tape in a building that could easily burn down because of all the equipment we run 24-7. So that's my only concern with that. But yeah, so there's a few options for you. I may actually just consider FS Transform for the systems here in the studio.

I may just do FS transform. Right, I mean, there probably are systems where, I agree, XFS is a great file system, but there are probably systems where you don't care that much about what file system it is. And so for that, you know, it does seem to work pretty well. I do also think it might be this next year's,

you know, April Fool's prank. Watch out, your file system might just be butter FS before you know it, buddy. Oh, that'll be the day. I think I'll be on might just be ButterFS before you know it, buddy. Oh, that'll be the day. I think I'll be on ZFS on the desktop before then, but we'll have links to sparse bundles, FSTransform, and all of that in the show notes at linuxunplugged.com

slash 262. And yes, go ahead, take a piss, give me a hard time, let me know how silly I am for using Dropbox for all these years. It's worked out so well, but in the end, I knew this day would come. So I'm not, I mean, I can't even say I'm super surprised. I can't say I'm super shocked. I mean, you've talked about it, right? We knew it was going to happen,

or it could. It's just the unfortunate reality that now you have a lot of work to do. I'm looking forward to the next Ubuntu podcast because I assume Mr. Wimpress will return to the show. Yes, it will be my glorious return. We're going to be recording tomorrow night because as I speak right now, I can exclusively reveal that Popey has just landed.

He has returned from Academy. So, yes, due to the unique way that the Ubuntu podcast is produced, we're not recording on a Tuesday evening this week. We're recording on a Wednesday. Ooh, time travel. Very nice. Well, I'm looking forward to it. Always a great show. Go check out the Ubuntu podcast to get more wimpy and Mr. Popey as well as Mark. It'll be good to hear the regular cast. If you have followed any of our PeerTube shenanigans, that's thanks to Eric,

the IT guy. We have a, Did you rename the channel, Eric, the PeerTube instance? Because there has been some other things in the works as we're getting ready to set up some really great developer resources. So you can find Eric there, but I don't know if we have a new URL for it? No, the PeerTube instance is still at...

No, no, I'm talking about the group where you're setting up something new, the new thing you're working on in the group that we should be teasing now. You know what I'm talking about. So, we're setting up something new. The new thing you're working on in the group that we should be teasing now. You know what I'm talking about. So we have a group for PeerTube, a Telegram group, that may be transforming to a more developer-focused group for people that want to work on projects, as well as Eric's working on a GitLab instance

where we're going to consolidate our open-source projects around Jupyter Broadcasting onto that with a great URL. And something that Eric and I have just started talking about but will probably be at that same site is essentially an issue tracker for Jupyter Broadcasting. So I could give you an example right now. The audio stream isn't working. It's something we're in progress, but it's not working.

So that would be an open issue. Hey, I noticed your audio stream isn't working. Or something like that, like if you had a problem with the website or a feed. So it's all things that are in the works, and if you want to get involved in that, it's whatever the Telegram short URL is slash PeerTube. I'm sorry, I think it's like t.me or something. Let me go see. t.me slash PeerTube.

See if that's it. t.me, is that it? Nope, it's not. Whatever the Telegram short URL thingy is. That seemed to work for me. Oh, yeah? JB PeerTube? Yeah, that's it. Whatever the Telegram short URL thingy is. That seemed to work for me. Oh, yeah? JB PeerTube? Yeah, that's it.

So we're going to start some of the planning and work around that. It'll probably get organized in the Telegram group, not to diminish some of the other chat platforms. There's the JB Dev, IRC, and things like that, but it's just another avenue for people to organize. Telegram seems to be hot these days. All the kids. Let Payne join the group.

I need to ban him real quick. Get it out of there. Yep. Now it's of there. Yep. Now it's really kicking back. Mwahaha. Yeah, that's good. I'm glad you got in there. All right.

Well, go get more Wes Payne. He's over at techsnap.systems. You got the Twitters too. Oh, yeah. At Wes Payne. That's good. So succinct. It is.

You are fortunate that you have the unique last name like that, that you can grab your own name. I joined early too. I had something, Chris something, and I lost it. I just lost it. I left Twitter. I'm good with Chris Last though. So I'm at Chris LAS.

You can follow me. The whole network is at Jupiter Signal. You can follow that to find out about show releases and all of that. We do this show live on a Tuesday over at jblive.tv, jupiterbroadcasting.com slash calendar to get that converted to your local time. I usually get the stream kicked off around 1 30 PM Pacific. So if you just want to do that math,

come hang out. Yeah. Hang out in the chat room, hang out in the mumble room, be part of the virtual lug. IRC.geekshed.net is the chat room. Google again, like I mentioned at the top of the show, Jupiter Colony

mumble setup and you'll get the mumble guide get a working mic, get a headphones get in there, chat with us. Tell us what you think about Linux and the stories we're covering this week. And you can also go to linuxunplugged.com slash subscribe for all the ways to get this show every week. Thanks for joining us. See you back here next

Tuesday! Thanks for joining us. See you back here next Tuesday. Get it out of here. Oh, Richard. Oh. Well done, everyone. Thank you very much. Enjoyed that. It was a good, tight group this week. It was.

But that worked out. Our sinking perspectives. We had our opinions. JBtitles.com. JBtitles.com. Let's go boat. Tribes of a Knit is pretty good. System BSD is also pretty good.

Those are both really good. And they're already tied on the boats. JBtitles.com. Don't make us pick. Settle this, guys. You the boats. JB Titles. Don't make us pick. Settle this, guys. You have this. Oh, Dropbox D. Dropbox D.

The Popey has landed. That's also pretty good. Yeah, now that the show's over, everybody go watch the rest of that talk because it was a great talk. It's like, you know what? I got to go to that. If that's the sort of opinions and stuff that they're open to having at a BSD event, I'm going

plus hang out with Alan Jude. That never hurts. Yeah, plus. You know what else? I mean, not to plus and plus here, but I like plus and stuff here. It's a great drive. You could go down the coast, down the Oregon coast. That is a great drive. I know. So I'm like, let's do it. Plus, go hang out at Intel.

Alright, okay. Twist my arm. Twist my freaking arm. And I'm sure there's lots of it. Let's go. Plus, go hang out at Intel. All right. Okay. Twist my arm. Twist my freaking arm. And I'm sure there's lots of fun, exciting stuff happening over in that BSD world. It's been a while since we've checked in. Yeah, right. I'm practicing, so I'll blend in in the group.