LINUX Unplugged - 404: You've Got Mail

Episode Date: May 5, 2021

It's episode III, Return of the Email. Everyone says never host your own email, so we're doin it. We just have one last job to complete. Special Guest: Danielle Foré.

So Wes and I have been kind of geeking out and following the development of getting Linux support for the M1 platform, like completely upstream support. And, you know, Hector Martin and the Ashi Linux team, they're working on that. And as part of this interest and just kind of document it so we can report on the shows, Wes was watching one of the live streams last night that Hector Martin does of development. Fascinating insights. development. Fascinating insights. And listen to how they're actually developing the GPU driver, which they've just upstreamed some of the M1 GPU driver support to Mesa with 75% OpenGL ES compliance, which is wild at this stage. But how are they doing this? Since they don't have a complete Linux operating system to run and develop on, they have an amazing hack. Yeah, the M1 GPU driver, that's the user space part, by the way.

So it's running on macOS at this point, but it's really awesome that it's passing a ton of the tests already. So that's why I really want to work on this hypervisor, because once this works, it will give me the information I need to write the kernel side of this. And since the user space side is going so well, basically, once the kernel side of this. And since the user space side is going so well, basically once the kernel side has basic GPU context and memory management, we can plug that in and then suddenly everything works on Linux too. So it's really cool that we're getting so much progress.

I had a private conversation with Hector and he told me that he was optimistic that we may have a usable desktop by December. And I thought he was crazy when he said that. I thought, no way, no way. But now that I hear that, he might be on to something. Sounds like you better start budgeting, buddy. I don't think so. I mean, you got to, if you think about it, Wes, if you're going to spend that kind of money, it's got to be, if you're going to do it for a brand new computer, it has to be really operational. I think at this stage, it's going to be, you can do data recovery, you can do forensics,

you could do some sort of like rescue and get basic Linux functionality. I think that's going to be where we get to in 2021. But that's going to be really valuable. Hello, good looking and welcome back to your weekly Linux talk show. My name is Chris. My name is Wes. Hello, Wes. You know, you didn't have to wear the postal outfit for today's episode.

That wasn't required. You turned the studio into a post office. I just thought I'd dress to fit. This episode is brought to you by the all-new Cloud Guru, the leader in learning for Linux, the cloud, and other modern tech skills. Hundreds of courses, thousands of hands-on labs. Get certified, get hired, get learning at acloudguru.com. Well, coming up on episode 404, it's webmail has been found.

It's the finale of the email server series this week. It's a three-part series, and we have one last important task we must finish. We must get webmail up and running. After a back and forth, if we were going to do it, we decided, yes, it's a feature we need. And really, to make it complete, we know we've got to have a little bit of a production test. So we're going to give out credentials live on the show to, I won't say who, I'm going to reveal, but to a test user. Hint, hint, he's in the mumble room, and he's going to try logging in for the first time and send an email and let us know if it works right here live.

Then after all this is done, we're going to kind of just from time to time follow up and do check-ins on the email server. It's not going to be like a regular thing anymore, kind of like how we do the Arch server from time to time. But of course, we've got community news, we've got pics, we've got some feedback. So before I get into any of that, I'd like to say time-appropriate greetings to that mumble room. Hello, Virtual Lug. Hello, everybody. Boy, it's nice to have, look at that, 25 of you in there. Good to have you here today. So let's start with a little community news because one of those members in that virtual lug

is Mr. Daniel Foray from the Elementary OS Project. And we have news of a beta for Elementary OS 6. Dan's joining us to tell us a little bit about it. Hello, Dan. Welcome back. Hey, how's it going? Oh, pretty good. Good to see a beta land. This is kind of an exciting milestone in the road towards elementary OS 6. So what are we expecting

here? This doesn't look like it's a final product at this stage, but it's looking like it's getting pretty close to something you could use day to day. Yeah, I tell you what, it's been a rough year for obvious reasons, but I'm really excited to get this beta out. So what we're really looking at here is we're making this big transition over to a Flatpak-based app center, right? And we have all these great new APIs, both from Upstream and from things we've been working on. And we're in a place right now where we want to really stabilize that developer experience and get people in and building their apps as Flatpaks with these new APIs and getting feedback about their experience so we can get ready to do the real thing. All right. So you mentioned the year.

It's been 370-ish days since 2004 came out, which will be the base. And I'm wondering, what are your thoughts reflecting on that amount of time? Is it the project's goal to generally ship sooner than that, or is it really just a matter of you ship when it's ready, and if it takes 370 ish days to get to a beta that's acceptable and what the user should expect yeah i mean it's a little bit of this a little bit of that like obviously it would have been nice to to ship a lot sooner uh we had planned you know a sprint uh early last year uh that we were supposed to be able to go and get

all this uh flat pack app Center stuff done at that time. And then, you know, the world blew up. So this is definitely a lot more delayed than we really wanted. But that is kind of the beauty of deciding to release when we're ready is we we've never really had these big commitments to dates. And it was more like, you know, this is the work that needs to get done. And as long as it takes to get the work done, like that's the experience we commitments to dates. And it was more like, you know, this is the work that needs to get done. And as long as it takes to get the work done, like that's the experience we want to share. And you know what?

I think the users appreciate that. So there is now ways though, for people to get involved and help accelerate development and help make elementary OS 6 great. I saw you have information posted where you go through the beta, you show some of the new features.

By the way, I really like that new, I think you guys call it accent colors feature that kind of just lets you spice up the UI a little bit. That looks really slick. How does that work, Dan? Does it highlight the default buttons and different colors that you choose? Is that essentially what it's doing? Yeah, so

under the hood, what it actually is is we're now shipping like eight different variations of our style sheet that are kind of built at compile time. And they all use a different base accent color. So when you're getting things like the suggested action buttons or like selections and radios or in menus or lists or things like that, like any time that you see, like normally you just see like a blue accent, you know, by default. But now you can kind of choose, you know, what your favorite color is. Or we actually even just merged the ability to have it automatically select a color based on your wallpaper. There must be a performance benefit to you guys building those themes ahead of time. And then when as a user, I select them, they're built and ready to go.

That seems that just seems great. Is that something that's normal or is that different? Something that elementary OS does differently? You know, I'm really not sure how other platforms are handling it. I think Deepin maybe has a color selection and I think maybe another desktop environment. I'm really not sure about how other ones are working, but this is like the one we felt that would be the easiest to maintain and allow us to add like as many colors as we want and make sure that when people go to switch them, they're just they're switched and they're done.

I just like to know these kinds of things because I think honestly, one of the great things about the elementary OS experience is I love a lot of GTK applications. But in the past, you know, I've run into issues with GNOME Shell itself. A lot of that's been fixed up. But, you know, elementary OS has always been there as an alternative that lets you play in that user space, kind of like Budgie does, and with some real nice choices you guys have made. So to that end, I'm curious just to get

your high-level thoughts on GNOME 40 adopting a horizontal layout that reminds me a lot of the way Pantheon does things. Yeah, I think it's really interesting. I think that a lot of people have this perception that developers from different desktop environments are in fierce competition and don't ever talk to each other or anything like that. But it's really not true at all. We were at the design sprint where they were working on the design for the new login and lock screen. And a lot of the

discussions that we had with the GNOME design team actually informed the design for our stuff. So like we're in contact with designers at GNOME and we go to Guadalcan and things like that. And so there's a lot of cross pollination of ideas. And it doesn't totally surprise me that we would decide to do a lot of things very similarly. Right. And I got to say, as an end user, the experience for me just means it's more consistent when I'm switching between desktop environments. It's just like one less thing for me to have to do a translation in my head. So it kind of works out for an end user.

So I like it. Well, there's a lot in here, like stuff with App Center and Flatpaks, but I'm curious if there's any other areas that you'd just like to particularly touch on before we give people any more info on how to get involved. Yeah, I think, you know, the biggest ones here are just that we were trying to roll out that whole new look and

feel with the new style sheet. And we really need feedback from app developers to make sure, you know, we didn't break their styles. So any kind of weird idiosyncrasies we'd love to have reported there. I mean, we're really just in this stage of like, here's like all the big stuff and like, tell us what we broke. Yeah. Like that new installer needs some hands on. That's probably a really easy way for people to, to give it a test. I like that you have this page. So I'm going to recommend everybody that wants to help out, check it out. Elementary.io slash get dash involved. Every project should do this. The team just outlay outlays a lot of ways you can just get involved, be it from funding

to design to desktop and support. It's super clever and it's really well done. Thanks a lot. Well, Dan, thank you. Thank you and thank you guys for putting out all the good stuff. And I look forward to giving it a try. And then, of course, the release. When that does happen, you come back and chat with us about it? Yeah, absolutely. I think we're probably going to kick off another beta, hopefully relatively shortly. We're getting a lot of great feedback and a lot of bug fixes in, so we really want to keep that feedback loop churning.

Thank you, sir. Thanks for updating us. Good to have you here. Thank you. Mr. Daniel Foray, everybody, from Elementary OS. We will have links to everything we just chatted about in the show notes, of course, so you can go right there. Now, a project that we just love on this show, we use it to help produce this here podcast is Audacity. And you may have heard this week that Audacity has joined the Muse Group,

which is a collection of brands that includes another popular open source music app called Muse Score. A lot of people are familiar with that. And I think that the Muse group is, I don't know a lot about it yet, but we've invited several people involved onto the show to come tell us about it. And I did reach out to the Audacity development team, or at least a couple of the developers, and they don't have a lot to add to this story at this moment. But the way it seems to be coming through is that MuseScore will be involved with the management and day-to-day operation of Audacity, much like how sometimes a foundation will. And then there will be a team of developers that will remain that have been there for a while that will continue to develop under that umbrella now. I essentially believe what's happened here is the equivalent of copyrights and trademarks have been purchased and acquired and transferred to Muse Group. And one of the key goals, it sounds like from the

outset here, is to improve the UI of Audacity. It's like kind of job number one right now, which Wes, you know, it's like the developer that I spoke with literally said to me, I'm also both scared and excited. Yeah, I think that's right. I mean, it's really been a long journey for Audacity. It began development way back in 1999. Yeah, that's right. Last millennium at Carnegie Mellon University and has come a long way since there. I mean, they just recently shipped version 3.0, which added a lot of stuff already. It's not really clear exactly all what's going on here. Muse Group itself was formed just like the company was formed a few days prior to the Audacity announcement. Although obviously MuseScore and some of the other stuff

involved had been going on a lot longer than that. We don't know a lot about the financials or the exact structure. But if how this works out is they've got a new steward who wants to help support the development of this free and open source audio tool, I think I support that. And because it is open source, I mean, if they ruin it, well, at least we can fork it. That's true.

And I'm not particularly familiar with MuseScore, but from just kind of doing some research for this story, it seems to be really well liked. It seems like people are very much a fan of that software and have a lot of goodwill towards this acquisition because of how they have managed MuseScore. Hopefully that's a good sign. This is one of those stories that does actually kind of terrify me just a little bit because it is a tool that is so precious to so many of us in any kind of media production on Linux. Media production, I mean, it's used a bunch in academic research. It's just sort of like it happens with the best open source tools.

It just becomes sort of part of the landscape because it doesn't cost anything. You can always count on it. It's always there for you. And it would be a shame for that to change. Yeah, I will continue to leave our doors open to anyone involved that would like to come chat with us. I've sent out a couple of invitations. And maybe we'll do a follow-up next week. Kind of a head-scratcher.

One thing that was a little strange to me is that the announcement started on YouTube in a YouTube video by the design lead from the Muse group, who's a well-known individual. And so maybe that makes sense. But it wasn't like a press release from the Audacity project or like the Muse group website, if there is one. And that video, I mean, it had a lot to like in the video, definitely. But it was aimed at the existing audience, you know, his audience on YouTube. And so it's sort of like, hey, I'm in charge of Audacity now, isn't that, you know, here's what I'm thinking, which is great. But yeah, there hasn't been a lot of communication geared to the wider open source community or users who aren't familiar with Muse Group,

but are familiar with Audacity. Put very well, and we will have a link in the show notes. There is now a post on the Audacity website, which we will also link in the show notes. But quite literally all it does is it quotes the YouTube video. They just put a block quote of essentially a transcript from the YouTube video. And then the only thing they add to it is that we're scared and excited, and they hope we are too. That's the entirety of the statement from the Audacity Project. And I think it's one of those situations where things are still getting dialed in. And whenever you have a deal like this, it involves copyright and intellectual property and money, that there's just certain things that people can't say right now. And so this news has kind of come out before maybe everything is fully formed based on my read of my conversation and this post.

It's like they just can't say everything fully yet. But the headlines are getting written fast. So that's why I wanted to contact them because the headlines clearly say the Muse group now owns Audacity and is in control of it and will be, quote unquote-unquote, partnering with the open-source developers, which is just not very clear. It doesn't really say who is quite in charge then because you can't make an open-source community do anything. You can't, just like you can't, the old, you can't herd cats. It's like you can have goals and you can have outlines, and if you get everybody on board, they'll go for it.

But if you just come at them from on high with a mandate, not all of them are going to do anything. They're not going to jump necessarily. Right. How many resources will be expended by the project in terms of development time? How much of that will be sort of shepherding and helping out existing development,

you know, volunteer resources or paid resources? We don't know. I guess we'll just have to pay some close attention to the repo and see what happens. Linode.com slash unplugged. Go there to get a $100 60-day credit. Yep, $100 towards a new account. And go there to really support the show, too.

Let them know you heard about it here. Linode is our cloud hosting provider. It's what we've built our new email server on. It's Linode, of course. It's what we do our back-end processing and hosting and our automation system and NextCloud and object storage for config backups. It's how we do a lot of the things that make it possible for a small team to operate at the scale that we operate at. That's why Linode is essentially like part of our team, because they're a fundamental aspect of how we get production done. And unlike the big clouds, unlike big cloud, you know, like AWS and the Googs and Azure,

Linode gives you access to tools they would never dream of. And man, oh man, is Linode's UI infinitely simpler and much more straightforward. And yet somehow you'll be amazed at the speed, performance, and complexity of things you could actually build if you wanted to. And you get 11 data centers to choose from. Every service level is backed by the best customer support in the business. And that matters. That really matters.

Instead of just being a number, you're going to talk to an individual by phone or ticket, and it's going to make all the difference when you're in a tough spot. And it's not just like that one great thing about Linode. It's like all the things about Linode that make them our go-to choice for anything we're building. It's not even really a debate because when you become kind of a Linode connoisseur, there's just things you can't do on the other providers. And they really, truly love Linux. I mean, you can see it in the projects they support and the conferences that they've supported. But when they are building anything,

they start from a perspective of how can we make Linux do this for us, even if it doesn't fully do it yet. I've shared the story before about when they became their own ISP. And back then when they did that, it was ridiculous to be using Linux at the layer that they were planning to use, at the networking layer. And now it's become an entire industry, you know, because they see this stuff like we do. Like we're sitting here shucking and jiving about Pipewire back in 2018 while it's shipping in 2021 because we're passionate about this stuff. We follow it for a long time, just like Linode does.

And that's what I love about them. Because if you're a longtime Linux user, that stuff shines through the product. I think it's really great. So go see what you can do with Linode. Go put that $100 to work for you and see what I've been talking about at linode.com slash unplugged. Build something or maybe even learn something. There's a lot of great places to host, I'll admit. But nobody does it like Linode.

Go see why we choose Linode every single time. Linode.com slash unplugged. You know, Mr. Payne, we have a really major, major housekeeping today. I normally say we just have a spot of housekeeping, But I enthusiastically embraced the housekeeping this week. Well, we woke up and it was a mess in here. Clearly, we didn't clean up after ourselves last week. Oh, it wasn't you and me. It was Minimac and that rowdy Luplug. They had a hell of a get-together this Sunday.

Minimac, do you want to give us like a little update on how the Mycroft get-together went? Well, it went really well. We had an excellent pre-talk. And I want to thank all the members of the Mycroft community that found the time for our really laplace session. And most of them will be

present next week, too. Yeah, that's the main show, right? That's the big event. Indeed. So in the pre-talk, we tried to nail down the structure of the talk that actually takes place next week. But we already had some inside information and everything was really cool. And we played around with our Minecraft bot we have now in the Mumble channel. And that one will be available, I guess, for the whole week.

So if you want to test that one, you can whenever you want. Please be aware that it can be a little bit tricky with the with the wake word with the hey my craft word so maybe you have to adjust your microphone settings a little bit i have some excellent news chris gessling you know them the guy from down under the community manager yep he has announced his presence too for, for the talk. So I guess we're going to have to beam him some coffee. That is great. So it will be an excellent talk. And I really hope that you find the time to join us.

I will put a reminder on all the JB social media channels. And all I can say is see you next Sunday. It's really neat to see the Mycroft team engaging. There's people that showed up just for the pre-talk this Sunday, so you know it's going to be a great turnout this Sunday. And yeah, you heard him right. They now have a Mycroft bot in the Mumble room that you can speak to via Mumble.

It's so cool. It's so cool. Yeah, so check it out this Sunday. Also, I want to put a call out there for your questions. We want to do a Q&A episode coming up later this month because I'm going to be doing some traveling. And so we want to just take that opportunity to just answer questions. We haven't done that ever on the show. So ask us anything. We'll answer as many as we can. Go to linuxunplugged.com slash contact for the Q&A episode coming up real soon.

And also I'll mention, we have a couple very limited retro last items in the garage right now, including the very last few coins. We still have some swag bags that honestly, the swag bags are super popular. Those are up right now at jupitergarage.com. It's the very last of the last coins. So grab one if you would like them. What's this Tech Talk Today portable speaker? I see. That seems like the perfect way to listen to a Linux Unplugged. If you happen to have a device that still has a headphone jack, it's just a little,

it just runs right off the headphone jack. I might have a battery in there. I can't remember. It works, though. I just plugged it in. Yeah, if you were a Tech Talk Today fan, I looked at that little speaker and I thought, you know, I kind of miss doing that show. I don't need to do any more shows. Definitely don't need to do any more shows, but I kind of miss it. So that's all at jupitergarage.com. So this is the finale of our email series. We today finally will finish it all up. We've been building our own self-hosted email server in 2021.

Yeah, I know. I know. We're doing it because everybody says you should never hosthosted email server in 2021. Yeah, I know. I know. We're doing it because everybody says you should never host your own email server. So we're going to do it so you don't have to. Spite email server. It's definitely the best possible reason. It is totally a spite email.

And we have two important tasks to complete today. We must deploy the webmail. And of course, we must get our first production user online. But I have to first acknowledge that we have gotten a lot of feedback about this topic. Wow. We weren't even sure if you guys would love this idea or not. But it turns out, a lot of you crazy bastards out there are also hosting your own email. But additionally, a lot of people are thinking about it for some reason. Maybe it's Google fatigue, or it's probably a dozen reasons.

But this has been remarkable. I spent, oh, I don't know, three hours yesterday just reading email about self-hosted email hosting. I kept sending Wes messages like, I've made it back to this date. I've made it this far. I was giving him like real-time messages. It was a mountain of feedback. Yeah, wow. I did not expect that.

I mean, a few, sure. But apparently email is a spicy topic. Yeah, and Hank wrote in with a pretty good question that I think we should cover on the show. So I wanted to respond to this one on air. He says, what's your plan to keep your email server secure? You tout the fact that your update woes are minimized. I think you've just shifted the need to update the OS, though, to needing to update the containers.

In this regard, I just prefer to keep things within the packaging system. On Debian and its derivatives, I just install unattended upgrades package, which defaults to install security updates daily. Security updates are tested and very unlikely to break anything. I suppose the same is true for a containerized email server, but I'm unaware of any automated security updates. By the way, unattended upgrades is a good little pro tip on Udebian users out there and derivatives, if that's something you might be able to get away with on your box. But he makes a good point, Wes, because I think we kind of maybe tried to connect security

with containerized application, and that is something we should never do on this show. Yeah, right. I mean, of course, you still need to do updates if there's a vulnerability or a bug that needs attention. Well, you got to get those somehow. Containers do give us some nice tools for sort of separating those concerns.

It allows us to sort of choose when we want our updates in our containers versus on our host system. And so if we did want, you know, automated updates on the host system, we certainly could, although we still need to attend to what's going on in the container. What's nice about having this all in Docker and open source is we have access to everything. So if we want to rebuild things, if we want to make tweaks, we have that functionality.

And we try to choose projects that have regular updates and releases to their images anyway. Beyond that, there are some nice tools around there as well if you do want to set up something like an automatic update. But you will have to do that yourself.

True. Real-time update from Neil in the chat room. He says that if you're on Fedora, a DNF-automatic will achieve the same result as unattended-upgrades. Nice tip. Yeah, thanks, Neil.

Yeah, I think the message I was trying to make last episode is that when I looked at, say, Mail-in-a-Box, which I still think is an awesome project, and if you're good with the traditional way

of deploying software where it just uses packages and it just kind of goes all over your system, that's a great option. I was surprised by my reaction, guys. That's, I think, the thing I want you to take away. When I used Mail-in-a-Box and I watched that install script run, instead of having this sensation of, oh, man, it's so great, this is just taking care of everything for me,

I had this sinking concern that i had no idea what was really happening to my system and i just watched package after package after config file after config file get written to the hard drive and i started thinking this is fundamentally altering this box in a way that i don't know if i'm going to feel confident in just upgrading when there's a new update that comes out. It also gets more complicated when, you know, like we were doing for today's show, start hacking in things like webmail, and suddenly you've got multiple conflicting things installed. And maybe you don't do that.

Maybe you've got, you know, a VPS or a VM for everything. But I think it's pretty common to just sort of add those things on when it's convenient. And then you've got mixed requirements that maybe one update needs an update, but that might break another part of the system. I mean, there's several advantages, right, Byte? Yeah, so when you restart a system because of an update, usually you get a message like a restart is necessary. And your entire system is down for a long time.

But if you use containers, a container is usually really fastly rebooted. And if it's only by the service itself, yeah, it's really fast. Yeah. I think that when you look at running your core applications that you want to run for, you know, five years minimum, and you want to maybe even survive an entire OS upgrade, having that isolation is such a killer feature for me. Other things like snapshots or easily quickly rebooting or switching to a different image

or something like those, there's a lot of nice advantages. Migrating to a new host server. Oh my gosh, yes. That's been very handy for us. We did a migration from DigitalOcean to Linode and it was absolutely much simpler by having nearly all of our applications containerized. It made it just significantly simpler. And so mobility and

maintainability of the host OS are paramount for me. And then there was other benefits, like instead of going with Ubuntu 18.04, I could now deploy CentOS Stream 8, which is something I've wanted to test in production so I could just talk about on the show anyways. I mean, Hank does have a good point though, right? That you don't get the security for free and it's still definitely something you have to manage and that may influence how you want to run your software. So Neil, I'm curious what your thoughts are, because I know one of the things that's drive you crazy in the past is when people equate running applications in a container to being secure. But there's probably still some use cases here that you probably agree with. Yeah. So like for me, it isn't necessarily about security to choose to use containers

over something else. Although, you know, security can be beneficial. The main aspects I wind up using it for is I actually kind of use it a little differently than most people. I use it mainly to replace virtual machines. So unlike a lot of folks, I actually have an init. I have systemd init running in my containers. I'm using systemd nspawn instead of podman for some of my stuff because a lot of my stuff are multi-service stuff. And it's just not worth it for me to do the extreme splitting required to do it in something like podman or whatnot so i run them as nspawn containers which means they boot up like they would vms and because i'm using butterfest in my file system they're snapshot and i can

actually do butterfest and receive to transfer them everywhere and i also because butterfest coolness i actually just layer them on top of the host operating system root file system. So I take the whole operating system, in one box, but I have the ability to orchestrate them independently. If they need to have conflicting requirements, they're easily, easily handled. Like if I have to have CentOS for an environment and use different module streams, like being able to use the different application stream from RHEL or modules from CentOS and activate them for the language stacks and whatnot. I can still use packaging. I can still use all the tools that I'm familiar with. I can still manage it with config management or Ansible or whatever I'd like. But because most of my applications don't need a different kernel and don't need

emulated hardware for all of them, I can drastically simplify the maintenance and management of my systems. And the big win for me there is, when I sometimes get trapped into running an application that can't necessarily upgrade right away, the operating system can move forward, and most of my things can move forward. And that one thing can stay back a little bit until either I figure out a way to upgrade it, or I replace it with something else. I don't use it as a way to hold back. I use it as a way to be more strategic about how I spend my time managing my infrastructure. Well put. And two thoughts. First thought is, and this is why we covered all this, is a huge part

of security is staying current, especially with software that has open ports to the internet. We've locked ours down and can because we have mail route in front of it, but we could lock it down further, obviously. I mean, there's all options we have there, but some software like SSH and, you know, web servers, they're always going to be exposed directly to the internet. So keeping them current is a huge, huge part of security. But if you listen to what Neil said in there too, is you'll also essentially hear the reason why developers that want a modern workstation need a more modern file system than Extended 4. He just clearly laid out how he can use ButterFS as part of his development workflow to give him different environments that are safe and disposable and do it in a very clever

way. And that's why you need something beyond what just Extended 4 can do. Even if you or I maybe only use it for compression or something like that, he or other developers like him can use it for a lot of crazy features, including just making their day-to-day job simpler. And that's why it's important that we ship really competitive file systems with desktop Linux distributions that want to be used as a workstation. Just my little side note, and thankfully, most Linux distributions will let you format any disk in just about any file system you want.

All right, so we had to decide essentially between two webmail packages as we saw it. There's a lot out there, and one of the kind of nice things about doing it in these three-part stages that we've done here on the show is in between the audiences had time to give us a lot of good ideas,

and we definitely got two suggestions the most via all the different mechanisms people get a hold of us. And that was Rainloop and Roundcube. And I was a little hesitant to Roundcube just because I had used it before and kind of wanted to try something new. So we first took a look at Rainloop. Yeah, you know, we heard a lot of good stuff about Rainloop

and I'd never used it either. So I was pretty curious where I think I'd used round cube, I didn't run it myself, but it was used by someone else that was hosting email for me. And that had been fine. But it's funny, we were just talking about, you know, how containers are kind of an integral part of this setup. And that's where things, at first we were hopeful, but kind of fell down in our trials with Rainloop. There is a Docker Compose and some Docker setup in their repository, but they don't have a lot of docs around it, at least

that I could find in my cursory survey. And it just didn't seem as set up for that. The docs on their main site really were geared around more of a legacy installation setup where you've already got web server infrastructure running and you kind of just bootstrap this PHP app onto there. By comparison, Roundcube has a lot of nice docs there.

Their site seems really well-developed. It seems like they've got a community around them. And they've got this whole plethora of Docker Compose support, multiple Docker Compose files, and very explicit examples of sort of like, well, here's three different options if you want to go the Apache route or the Nginx FPM route, and here's an option with Postgres,

or here's an option with MySQL. So it just felt like, here's a project that's embracing this as an option for running the software that they make, and that was a pretty big signal. And simplicity in setup is also appealing here because it's not going to be our primary job to admin this. And it's kind of a jackass move even to put a mail server in production and then give it the amount of attention that we're going to give it. So we want to build it in a way that when future Wes and Chris come back to this thing, we know how to fix it or how we built it or how we could even rebuild it. And when we looked at Rainloop versus Roundcube, it honestly kind of just came down to Roundcube was a little simpler, had a little bit

clearer documentation, and just was a little bit easier for us to snap into the old configuration and setup that we had. Now, I still would like to do some playing with Rainloop. And there might be some things I've missed. Or if the audience has a compose file handy that they want to share with us, that'd be great, too. I would totally try it. And I think it's still something I want to try. It was just to get this, you know, accomplish our task, start playing with webmail, see how it would integrate. Roundcube was

really the simpler and easier option. And, I mean, it is pretty neat. Like, it's not a bad experience. We'll get more into that. And they've been around for 13 years, like since 2008. That's also pretty impressive. And it seems actively developed today. A former past version of Chris definitely deployed Roundcube and liked it back then. But, you know, we could sit here and tell you about it,

but maybe we should put it to the test. All right, Minimac, get your telegram ready. This is it. I am ready. So Minimac will be our first production user. We want this to be a tool for members of our community. And we thought, let's put it to the test right here. So I just sent you on telegram the URL for the webmail, your credentials. If you could log in, first of all, tell us if it works, if it loads for you, and if you can send off an email. Yeah, you should use your full email as the sort of username. Login.

Yeah. What do you think, Wes? Are you nervous? Fingers crossed. I should be looking at the logs, shouldn't I? I'm a little nervous, actually, because, you know, we really did not test this beforehand. I mean, we tested it, but it's a whole different ballgame. Okay, first success.

I'm in. Good, good, good. That's a good sign. Okay. Okay, first success, I'm in. Good, good, good. That's a good sign. Okay, now, do you want to send me an address where I should my first famous first words send to?

Oh, yeah. Email me, chris at jupiterbroadcasting.com. He gets a lot of email. He probably won't reply. Don't be offended. Wes suggested that I should send you a welcome email, but I wanted you to send the first email. I think this is a good test because Gmail's really picky about who it actually accepts email from.

So if it's going to reject your email, it's going to be Gmail. My heart rate is pounding, actually. My heart is going. I suppose the question is, does Minimax sound like a scammer in his email? He does have some products to sell you, but that's just, you know, that's a coincidence. Okay. Email is out. Oh, boy.

Let's verify and send. Yeah. Okay, I don't see it yet. Come on. Wes, you see? Do you have the logs up? Oh, no. Here, let me check.

Wait, hold on a second. Hold on a second. Oh, there it is. What? Yeah, you got it. It worked. Congratulations, Wes. Oh, there it is. What? Yeah, you got it. It worked. Congratulations, Wes. Yeah, we got it.

See titles, famous first words. All right, I'm going to write back. I'm going to write back. Just very simply, it worked. And I'm going to, dude, tell me if you got that. I mean, I'm very happy. I suppose we should check it round trip. But did you get the reply?

Not yet. Oh, my God. I hate email. Just the other week you were talking about how it's a cool federated system. And now here you are. I know. It should see it. It should come right in.

Success. Ree from Chris Fisher. Famous first words. Excellent. And the content is, in fact, it worked. Ha ha ha ha! Nicely done, Wes. Nicely done, Minimac. I mean, honestly, I'm still surprised. It just doesn't seem like we did it so hastily and we really didn't know what we were doing.

But thankfully, the tools that we have access to right now are good enough that it kind of takes care of that for us. I mean, not to do a shameless plug, but it is legitimately why Alex and I started the self-hosted podcast, because these tools, the stuff that you can run yourself now, it's at a level way beyond anything we dreamed when we first started doing these shows. And I am so grateful for the open source projects that make all of that possible. And so our plan there is we're just going to slowly roll that out a little bit to certain people just over time as we just kind of expand it. And then maybe eventually, I don't know, we might go further from there.

I don't know what we're going to do exactly. We just want to try to get a few people on there to make sure everything's working because, you know, even though we're crazy, we want to be cautious. And then we'll probably just kind of ramp it up over time. So I guess more on that when we do a future email server update. But that essentially right there completes the email server series. And Wes Payne was a hero and a legend in this one because we put a lot of work, specifically Wested, into Rainloop. And we thought we were going to go with that all the way

until just really last night, maybe this morning, and then at the last minute we said, now let's do Roundcube instead, and Wes got it all set up and running just a little bit before the show. Like a real legend. So round of applause, everybody, to Wes Payne for getting that done. Good job, Wes! Now you'll just have to wait and see how it lasts, right? Can we maintain it? How does it do? Do we keep using it? Stay tuned to Linux Unplugged for more.

MailRoute.net slash Linux. Try out MailRoute today and get 10% off the lifetime of your account and get a 30-day free trial, no credit card required. MailRoute is how we are routing mail through to our mail server. So yeah, during the live show, we went through MailRoute. It happened so dang fast. And if you have a business where, I mean, even just a handful of people, you know how critical email is. And you know how bad threats like ransomware are getting more and more.

And admins are looking for ways to manage that risk. And MailRoute is a fantastic way to do that. Admins are looking for ways to manage that risk, and MailRoute is a fantastic way to do that. For 24 years, they have focused on that core competency, just providing cutting-edge email security. And I absolutely loved the tools as an admin while we were setting up our email server the first time around. I logged in and looked at their real-time logs and saw the emails coming and going and knew that, hey, we'd gotten some stuff set up, so I was pretty confident it was going to work. You know, it gave me a pretty good snapshot, and I really liked that tool,

especially when I'm setting things up. But another thing that I know I'm going to take advantage of at some point, and I would really recommend you consider this as well, is you can use MailRoute to queue up messages for up to like 15 days, or like whatever you set it to. So that way you have time for maintenance or if you have an outage or an internet outage of some kind, MailRoute will make sure you don't lose your email. And then when your system comes back online, it just forwards it right to your mail server like it was never down really. That's just such a nice peace of

mind. Additionally, if you work in government and compliance is an issue for you, MailRoute has some of the most important compliances that you could have in that business. So you can feel safe using MailRoute. Of course, they have a lot of information on their website. So go try MailRoute out today and get 10% off the lifetime, lifetime of your account. How great is that that they're doing that for us and for you guys too? So go get a 30-day free trial, no credit card required. Visit MailRoute.net slash Linux.

This is an example of one of those great partnerships where we contacted MailRoute because we were absolutely going to be using MailRoute. And we just said, we think you'd be a great fit. We're doing this mail series where we'd love to have you on board. MailRoute said, all right, let's do it. And they did it for two episodes. And now they're back because they got a lot of great feedback from you guys. We had a bunch of you try it out and really like it, and so they thought, let's stick around for a little bit longer.

So they're here for a little bit longer, and I'd love for you to go try them out because I think they could provide a real service regardless of the scale of your institution. But this could offer a lot of benefit if you're a large institution or a small business. But if you're just a self-hoster and you want the peace of mind, you don't want to worry about the crap, you don't want to worry about the crap and you don't want to worry about the downtime, MailRoute's going to solve that for you. And they've been solving it for 24 years.

So go to mailroute.net slash Linux, get 10% off the lifetime of your account, support the show, and start a free 30-day trial with no credit card required. Mailroute.net slash Linux. Well, speaking of email, MailRoute.net slash Linux. or 34 listeners would like you to cover. I know you choked about Pipewire being too niche, but as a musician and podcaster transitioning off macOS and onto Linux full-time, I would love to hear more about it.

Audio has been the single biggest pain point in my move since it works dramatically differently than in macOS. I only learned about Pipewire shortly before listening to the episode, and it sounds super cool. Please do share more of your thoughts on it. Best, Murray. Murray, you are coming in at such a great time into Linux Media Production.

Oh, how I could just picture that the future for you is going to be so bright because Linux has really gotten to a good point. I kind of seem to kind of get the impression that you're focused more on the audio end of stuff. Pipewire does manage video.

In fact, it started to manage video. But it has really, really gotten to the point where you can produce media on Linux all day long, and it's fantastic. We do things that I simply could never leave Linux. We just had a hypothetical back and forth this other Sunday when we were recording LAN. How would we even recreate what we do now on other platforms? It's remarkable what we get out of one computer running Linux because of Jack Audio.

And Pipewire takes everything I love about the current state of Linux audio and makes it easier and remains compatible with it. And that's why it's going to be kind of a big deal as it rolls out. It started as really a way to just focus on getting video in and out of the Linux desktop, particularly like in a Wayland future. But it's really grown into a low latency audio and video system for playback and capture. Yeah, you know, we did some chatting with the developers and they kind of emphasized that, that once you've got this sort of graph node processing to handle, you know, little bits of either video or audio, it's kind of the same. And why not unify that in one tool? You can think about this too, in terms of development, you can think of Pipewire

as a server and a user space API to deal with multimedia pipelines on the desktop, like mixing available sources of video and accessing sources of audio and bringing those all together. And there is an API the developers can write to as well as they can just continue to use Pulse Audio or Jack Audio if that's what their application is designed for because Pipewire speaks that as well. That's huge. Right, so you get the legacy support,

but you also have a new Pipewire native API if you want to be future forward. Right, with that more efficient, lower latency Pipewire backend server managing it all now. So really, depending on your distribution, things are going to pretty much work just as normal. You won't even really know anything has changed. And it really has been, for me at least, a very smooth transition. It's been a very positive transition, unlike maybe a decade ago where it would have been a little,

or a decade plus, it would have been a little bumpier. Yeah, it should all just work. But one of the neat things about Pipewire supporting all of these options at the same time is in the audio world, it unifies them. Before you had a lot of neat stuff you could do with Jack. And then of course there was the Pulse world, which was growing more and more functionality with things like Pulse effects. And you can do a lot of clever stuff with Pulse, but it was never really designed for that pro audio, low latency sort of workload. And that just meant you had two or three different Linux audio worlds that didn't always play nicely together, or you had to, you know, really know what you're doing to link them up correctly. But in Pipewire, that just works. Or maybe another way

to put it is stack them on top of each other, right? You would have sound systems stacked on top of sound systems. Yeah, really. As much fun as that sounds like. And so I think, you know, for Murray's case, things should just work. You can just get started. But when you're ready, you can start playing with some of these tools that are designed around Jack workflows, things like Katya or Carla. And then you can start doing some of that fancy orchestrating of moving inputs wherever you need, doing loopbacks, monitoring input on the system,

all that fun kind of pro audio stuff that you might have been using a special Mac app for. Yeah, put really simply, if you're recording a podcast and you want to record a guest and you want to do it all on maybe one laptop or one computer, how do you capture that application audio and bring it into your recorder to record it? You need tools that can link that all together,

and that's what Pipewire will enable in a way that we've had before but was not in a unified framework for exchanging all of this. And none of the previous solutions that we've talked about were really built with our Wayland future in mind and the security requirements that come with that, like having applications sandboxed in flat packs that need ways that are standardized and APIs to communicate with each other.

And Pipewire aims to solve all of those problems, too. Outside of just the multimedia improvements we've talked about, it's really going to be a unified framework for both consumers that won't even know they're using Pipewire, quite legitimately won't even know, and pros like us who are going to rant and rave about it for years now because this is over time and even already beginning to solve pain points that we have struggled with for 13 years, legitimately probably maybe even longer, maybe 15 years. And so it's really exciting for us as content creators who have been using this platform for a very long time to see something that really truly feels competitive with what you see on the Mac side.

Well, we're clearly very excited about Pipewire, the future, especially now that it's deployed in Fedora 34, but it's also a project we've been following for a long time here on the show. So if you want some background and to get an idea of how far things have come, maybe check out Linux Unplugged 272 way back when, October 2018,

when we sat down with lead Pipewire developer Wim Tamens, who told us what he was aiming to do. And just a couple of picks before we get out of here. This week we wanted to share a pastebin alternative. There's a lot out there, so we'd love your feedback, too, for ones that you love that are simple and self-hostable.

But our beloved Slexi.org, which we have used as a JB team for years to pass pastebins back and forth, or when you hear us read an email on the show that we've linked in our show notes, we put it in Slexi.org. When you hear us read an email on the show that we've linked in our show notes, we put it in Slexi.org. And they are shutting down as all hosted services that giveth eventually taketh. Just when you really depend on them and don't think they'll go anywhere. I really had.

I just started thinking, yeah, I guess they are sticking around. So this week, Wes found us a markdown-loving Pastebin service. Although, really, I should give credit to Gamma over in our Matrix chat, the JB Matrix who posted this, and I just thought, boy, that seemed perfect for our needs. Thank you, Gamma. I didn't know that. Look at you, good guy Wes, giving him credit. Alright, now I'm trying to get you to say the name, though, because I'm avoiding the name.

I'm going to say Rentry. Rentry.org. Or Rentry? I don't know if that's better. Hmm. Yeah. And Rentry.co too worksEntry? I don't know if that's better. Yeah. Hmm. Yeah. And Rentree.co too works maybe? I'm not sure, but it's called Rentree, and it's a Markdown lightweight pastebin that is really clean, and you can combine images, inline links, code snippets,

you know, all the Markdown stuff. And we put a link to that in the show notes. And then this is also at the same time the show saying to you, hey, we know of like a handful of these, and we've debated like which one we want to use, which one we want to self-host. But we'd like to point to this link and say, hey, do you know of anything that's kind of clean like this? Because what we need is just a nice, clear reading experience with not a lot of junk because we're reading it live on air, and we don't want like sidebar stuff or big, huge, fat header bars and stuff like that. We just want something nice and clean. And if you know of a pastebin alternative

that's open source and self-hostable, please let us know at linuxunplugged.com slash contact because we'd like to try one out and then standardize one for the network like we kind of have with HedgeDock. And then we have one last extra pick for you this week, and it's a special one, Wes. Do you happen to know why?

Something tells me it must be written in Rust. That's right. Feels good. It's been too long since we've had a Rust pick, and this one is the kind of pick you were never asking for, a replacement for PS written in Rust called Prox. But hear me out. It actually has a pretty great output.

They call it a human-readable format that is colored with automatic theme detection based on your terminal background, keyword search over a multi-column UI, and then additional information that your boy PS doesn't have, like TCP UDP port usage, read write throughput, the Docker container name that that process might be running under, and more memory information, which is nice in the era of system D OOMD. Ooh, you know, this tree view is pretty sweet too.

Yeah, it really does actually look good enough that I'm, I think I'm legitimately switching to this. I initially picked this as an ironic pick because who needs another PS? That's ridiculous. And then I installed it and I absolutely love it. It's even great on our server. I love the Docker support and that is really handy. Whoa. And there's experimental support on both Mac and Windows. Hey, look at them. Everybody gets to play. Oh, I didn't even really think about the fact that there's Rust apps on the Mac.

Of course there are. I knew Microsoft was hot to trot with Rust, but I'm sure Apple's been super helpful getting Rust working on macOS. And Chris, you'll never guess, it's already packaged in Arch. Oh, of course, yeah. Absolutely, I knew that. You know I knew that. Anyways, go check it out.

It's a really nice, really clean, really elegant way to list all of the processes and the resources they're using and all of that. It's called Prox, P-R-O-C-S. And yeah, it's up on GitHub, so it's probably just easier to get the link in the show notes, if I'm being honest with you. Thank you very much to our Unplugged Core contributors at unpluggedcore.com. For like a few more hours, the secret promo that I gave out will be valid, and then I'm going to remove that product from the garage.

So if you haven't taken advantage of it yet, you got hours. You got hours left. But, you know, there's other things we do to thank our members. We also make a couple of feeds available to them, a completely unedited raw version where you hear us screw up and lots of probably, I would say, two solid pre-shows easily, maybe three pre-shows that we just totally blew and did not record except for our members. That's in the feed this week, too. And then also you can get just a limited ad feed if you prefer. It's a little tighter version of the show, same full production, all of that stuff. Those are two different feeds we make available to those of you who help make this show sustainable

at unpluggedcore.com. And thank you everybody who does support the show there. Be sure to go check out jupitergarage.com too for some of that limited time last merch. And those coins are almost gone at jupitergarage.com too for some of that limited time last merch and those coins are almost gone at jupitergarage.com if you do the twitter thing you can follow the show at linux unplugged the website's linuxunplugged.com but we do have a twitter handle which um you know you can follow you can follow i'd follow the network i'll just follow the whole network at jupiter signal i'd probably follow wes wes tweet a picture of some dogs sometime.

Maybe you have. But you know like that epic superhero like A-team shot you took of the pups the other morning? That would make. That's some serious Twitter material right there. Cute dogs. Who can say no? All right. I'm on it.

I can't say no. I can't say no. So go find him. He's at Wes Payne on the Twitter. I'm at Chris LAS. The entire network and all of our fantastic podcasts like the self-hosted podcast and Linux Action News and the Coda Radio program

are all over at jupiterbroadcasting.com. And we'd love to have you join the show live. See you next week. Same bad time, same bad station. Just get your Linux on every Tuesday. Get all Linux-y. And join us live at 12 p.m. Pacific, 3 p.m. Eastern. on every Tuesday. Get all Linux-y.

And join us live at 12 p.m. Pacific, 3 p.m. Eastern. I'm not exactly sure what it means to get all Linux-y. Probably means like boot up a Linux box and join the live stream. That's probably what it means. I think the only way to find out is join us live next week while we get Linux-y. Right, yeah. Maybe bring a towel, though. Yeah, that's not a bad idea

and a beverage. And you can absolutely join us in our Mumble room. Info for that is at linuxunplugged.com slash mumble. That's the way. You get the idea. You get the idea. We have a website. Aren't you proud of us? And we put links of relevant information on there that you're probably wondering. It's a good idea. See you next Tuesday! අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි අපි All right, jbtitles.com, let's go name this thing.

Did we miss anybody or anything in the show that we wanted to catch up on the post-show before we head out of here? Well, Neil let us know that our pick is also packaged in Fedora, so all those folks we convinced to try Fedora 34 don't have to miss out. Nope. Absolutely. Absolutely. I don't know. We kind of played with the whole Episode 404 thing with Coder Radio, so I feel like it's kind of silly if we go all in.

But also I feel silly if we don't acknowledge that today is May 4th, which all of the Star Wars fans are excited about because the 4th is with them. And it is episode 404 on May 4th. There's a lot of fun to be had there. But this show is not a Star Wars show. And we just did the 404 gag with Coder Radio like 11 weeks ago.

So I'm perplexed as to how to title this monster. Well, you could always reference the 504 code. Which is? I don't actually know what it is because I've never seen it before. That is the least helpful ever. Like, I'm sure Wes as a web dev actually knows it off the top

of his head. Method not allowed. Oh my gosh, listen to you. Nicely done.