LINUX Unplugged - 325: DNF or Die
Episode Date: October 30, 2019Fedora 31 strikes the right balance, we get the latest on the Librem 5 situation, and an easy graphics boost for laptops. Plus the best way to share your terminal yet, and more. Special Guests: Alex K...retzschmar, Brent Gervais, and Drew DeVore.
Transcript
Discussion (0)
What do you use for flashing ISOs?
I'm all in on Etcher now.
Yeah, me too.
I'm just using Etcher right now to burn Fedora 31.
And even old Wes Payne here is all in on Etcher.
I usually use Etcher, except for the occasional times if there's something weird.
We did have an issue with the 1910 ISO.
It was Pop, I think.
Oh, okay.
Pop 1910 ISO.
Etcher was just like, nah, no thanks.
I don't know why.
DD Rescue worked fine, though.
Yeah, I kind of feel betrayed that you guys don't DD everything, man.
Come on.
I used to.
I used to be the guy that was always like, just use DD.
It's built in.
Yeah.
I like seeing the transfer rate, the progress bar,
and I like just a little bit extra double check that I'm selecting the right media
because that one time I overwrote the wrong thing.
You know what I mean?
Yeah.
You're going to have to choose, right?
Otherwise, you might have to go run like, run an LS bulk, right?
Like, go run, look for what drive you want.
You might as well choose it from a drop-down.
Yeah, exactly.
See, I've just been piping DD through PV,
which gives you the progress bar and stuff, which is quite nice.
Ah-ha, I knew somebody would say it.
Am I the only one who uses GNOME disks?
Oh, yeah, that's right.
I actually just got into the habit of using Etcher, I think, before who uses Gnome Discs? Oh yeah, that's right. I actually just got into
the habit of using Etcher, I think, before Gnome Discs added that, maybe? I say that,
but I'm not so sure. Gnome Disc is really nice too, because you can also benchmark your disc
right within the app. Right. Yeah, that's the other thing we don't talk about very much.
What are you looking at? You looking at it? I just opened up Gnome Discs, yeah.
How does the imaging thing look? I don't know.
Let's see.
Does it look better than that, sure?
I don't actually know how to do it.
The three dots in the top right of the disk should be.
Yeah, select the drive, go to the three-button menu, and do restore disk image.
Aha, there you go.
It's very Mac OS.
Okay, not bad, not bad.
Choose an image to restore, go browns for it.
It's built in.
Yeah.
So there you go.
If you need something that's going to show you the layout of the disk anyway, I mean, that's nice.
If that's not a strong Westpain endorsement, I don't know what is.
Put that on the box.
Hello friends and welcome in to Linux Unplugged, episode 325.
My name is Chris.
My name is Wes.
Hello Wes.
Hello.
We have so much show today that we've been referring to the show notes as a novel. It's packed full of news and our first take on Fedora 31, which lands today as we record.
Don't worry, there are some romantic subplots, though.
That's right.
There are a few twists and turns.
We'll get you binging in no time.
And to help us get through all of this is our crew.
Hello to Alex, Drew, and Cheesy.
Hello, guys.
Hello, internet.
Hello, panel that's going to help us get through today's show.
And then, of course, our secret sauce, our community, our virtual lug.
Time-appropriate greetings, Mumble Room.
Hello.
Hello.
Hello.
Look at that.
It's nice to have everyone in there.
Really nice.
So I thought we'd start with something we don't do too often,
but maybe should do a little more of now that I think about it,
and that's a little bit of follow-up.
Sirs Lowell on Twitter said,
Hey, Chris, quick question.
What's the device you've been using for your external GPU on your laptop?
Is it a breakaway box?
Is it some sort of PCI pass-through?
How does it connect?
What's the deals?
And I realized I hadn't really talked about my eGPU much.
We did an episode on it, I don't know, ages ago.
I'm pretty sure it was 17 years.
That's right.
Does that sound right to you? I think that was 17 years ago. I'm pretty sure it was 17 years. Sounds right. I think that was 17 years ago.
And I then kind of
changed the way I'm using it and never really
followed up. Yeah, right. I mean, last time
we talked, you were doing the whole
pass-through thing. You had a dedicated gaming
VM that you would set aside your
fancy chip and then pass it
through there. Yeah. That's not what
you're doing now. No, that was a lot of work.
I mean, that was amazing because I could have like...
It sounded like it worked very well.
I had the Ubuntu setup that had all the games installed in one VM,
and I would just move it around.
But truth be told, I wanted something that was a little simpler.
You're lazy.
And a little faster.
And a little easier.
Yeah.
simpler. You're lazy. And a little faster.
And a little easier. Yeah.
So it turns out, it's stupid easy just to run another
application on an external
GPU and still have it render on your
main screen. So you can just have the two
GPUs right there in the host system. You don't have to
virtualize at all. I didn't
really appreciate how simple this would be to pull off.
So the way it works on my system
is I've got my T480.
It has Thunderbolt 3, which is a USB-C style connector.
That goes into what I got is a Mantis Venus.
A Mantis Venus?
Oh.
I don't know.
I'll put a link in the show notes so you can just get the – it's a nicer, bigger eGPU box.
But it also has USB.
It has a slot for an SSD.
Like it's a larger box that I use for other things.
But they have different sizes.
Some that have GPUs pre-built in.
Some where you can put your own GPU in and maybe replace it once a year.
It's a well-established market here.
And really, it's compatible with any PCI Express GPU that fits in the box.
So you could get years of upgrading your GPU out of this thing
and moving it between any system that has Thunderbolt 3.
So the way it works is when you plug in a second GPU to Linux,
it just registers as another PCI device on the system.
Zero is your default GPU, and one is the second GPU.
And you can pass an environment variable to an application
that just says, use the second GPU in the system.
And in the case of Steam, if you were to pass that environment variable and launch Steam,
every sub-process that Steam executes also executes on that second GPU.
So that's exactly what I do.
So you've just done, I mean, you launch Steam and then...
I launch Steam on that second GPU and I play my games.
Proton doesn't matter. It doesn't matter if they're Windows Steam and then play games. I launch Steam on that second GPU and I play my games. Proton doesn't matter.
It doesn't matter if they're Windows games or native Linux games.
And they all just render on that GPU.
But thanks to the great development of all of this lower-level technology,
they still work on my main screen.
I don't have to have a monitor hooked up to that GPU.
That's so cool.
It's so cool.
So I have a nice, portable, high-battery-life laptop.
Until I sit down, I plug in this eGPU, which also provides my—
It's like a magic enhancement box.
It's like a booster.
It's like an upgrade.
You know?
I'm like Bane sucking up the juice.
And the thing is, it's also providing power, so it's one cable.
One cable.
And I get all this.
And so I passed all this over Twitter to Surslul over the last week.
He said he got one, and he's freaking loving it.
Oh, nice.
And what I want to be clear about is I went all crazy with the PCI pass-through and setting up a VM and having a dedicated keyboard, monitor, and mouse.
Like, I thought that was the way I wanted to go.
This is way better.
It's so much simpler just to, say, you know, run on GPU 1, execute the application, and then just, otherwise, I admit,
it just sits there idle. Nothing's using that GPU unless I pass that environment variable,
but that's a trade-off I'm fine with. Use the Intel GPU for my Wayland session, my Gnome Shell,
all of those local applications that I'm using, like my browser, use the Intel GPU for that.
It just works.
Yeah. So if you've been thinking about it, if you've got a box that's got Thunderbolt 3, and you've
looked at these eGPUs, I'm here
to tell you that Linux is actually easier
to use than you expect, because
it gets a little trickier if you've got to load the NVIDIA
binary driver. But you should
have known that going in. There's always that caveat.
But if you're just using stuff that's
built into the kernel, like the AMD driver,
super easy.
But it is totally doable with the NVIDIA too.
I've done it with the NVIDIA graphics card as well.
So whichever your preference, it will work, and it's much, much simpler.
And also, just a little shout-out to Ubuntu Mate,
which makes this even easier because it wraps graphical commands and menu items around.
Of course.
I'm always thinking of stuff like that.
So I'm just thrilled with the setup.
Thrilled with it.
So if you have been on the fence
and want to add a little bit of oomph
to your Linux box
without having to buy a dedicated gaming machine.
Yeah, that's expensive.
And also, it's such a trade-off
because I considered it recently.
It's heavier.
It gets worse battery.
It runs hotter all the time.
It doesn't fit on the airplane tray.
Like there's all these downsides.
Plus it takes up more room in my bag.
I'm already fond of the ThinkPad.
And I only want a game like, I mean, realistically twice a month maybe.
And so that's why this is nice because I can keep buying the cheaper laptops
with the sole requirement they have Thunderbolt and just move this between them.
So anyway, SirsLol, I'm glad to hear it's working out for you. Keep buying the cheaper laptops with the sole requirement they have Thunderbolt and just move this between them.
So anyway, SirsLol, I'm glad to hear it's working out for you.
And I encourage others out there to give it a go.
I wonder if Lenovo will ever consider doing something like that with a dock.
Dude, I have one.
I have a Lenovo GPU dock that has an NVIDIA graphics card in it.
I don't like it as much.
It's a little underwhelming.
Okay.
Yeah.
The bigger boxes, as you would expect, cool better.
Right.
Of course.
The smaller little, like, put it all in one really small Thunderbolt dock, it's a great idea, especially if you don't have a lot of gaming use,
but you just want GPU for some reason.
It's great.
But if you start pushing it really hard, it overheats and it throttles
and then eventually locks up your system and you got a hard reboot.
That's just a bad experience. Yeah, it overheats and it throttles and then eventually locks up your system and then you got a hard reboot. That's just a bad experience.
Yeah, it's a bad experience. So, I do
like the slightly bigger boxes.
I find using AMD graphics to be
simpler, although it's totally doable with NVIDIA
too. I just find it simpler with AMD.
Yeah?
There's your follow-up.
Who did it, Wes?
You've probably pushed me over the line here.
I've got a box that I don't use that much because it's a separate machine with an older graphics card.
Not too old to knit in it.
You've got one sitting in front of you, too.
Yeah.
You could hook it up to that laptop right there.
Perfect.
You know, Steam sales come and go.
You grab a few games.
But you also, there's other things you do that could benefit from GPU acceleration.
Indeed, that's true.
You madman.
Well, last week, Purism's former CTO made news in
an exclusive interview with Pharonix. Interview seemed pretty damning, but we chose to not really
cover that story because there's, you know, there's things like that that come and go when
you're building something like this. But this week, a story did come up we decided to talk about.
This week, the Librem 5 is making news with an update on the Aspen batch.
Listeners, you'll remember that they're shipping the Librem 5 in batches, each batch iteratively getting better.
And the coverage around the initial batch was super positive.
It's out there.
Ars Technica went to the web and wrote, the Librem 5 is becoming an actual real product.
Headline, Librem 5 phones start shipping a fully open GNU slash Linux phone. Well, that was not
actually the case. Now that we are past the deadline for the Aspen batch, it turns out that
only a few Pura's employees ever received a phone or their board member. In fact, maybe only three or
four phones exist in total. So now we're at a point where we need to sort through what Purism
is claiming and what the community is observing. And they are getting further and further apart.
So I needed somebody who's followed this closely. Jay, welcome to the show. And Jay, you've been a little obsessive about this topic, haven't you?
I have.
I had pretty high hopes for this device, and I still do, honestly.
It's hard to find the signal through all the noise that's coming out of Purism.
You and I are both Purism customers.
I was a backer of the Librem 15, and you, until just recently,
were a backer of the Librem 5. Yeah. And I also have a Librem M15. That's very hard to say,
for some reason, but I have one of their laptops now. Yeah. So we're both current Librem laptop
users. And I think both of us are on the same page when I say we would love to see this be successful.
Even still, the Libram 5 would love to see this be successful.
That doesn't necessarily mean there aren't things that raise concerns.
Oh, yeah, definitely.
So I've been following this, but not as closely as you have.
So I wanted to bring you on to kind of go through some of this.
So you walk us through what seems to be the root cause of the most
latest delay of the Librem 5. Well, I've been gifted with some information from informed
people. And the base, according to these people I've been talking to, ex-employees to be clear,
they've pretty much told me that Purism really is having a lot of trouble wrapping their arms around this whole manufacturing thing and the designing of the hardware and whatnot.
I mean, Purism, when it came to the laptops, they outsourced a huge amount of that work.
And they did some really great work in terms of the software with the Intel management engine and core boot and all that.
But most of those laptops, I think, I want to say I
heard somewhere that they're really just an Acer laptop shell and they order it from somebody else
in San Francisco. And those people are doing most of the heavy lifting. But with the phone,
they didn't really have that option. So they had to kind of go to the mat and get into the
nitty gritty where they weren't used to getting into before.
And I think that they don't have a lot of experience with this kind of thing,
especially when it comes to phones and cellular transmissions and things. And I think that's probably a huge amount of the root cause of a lot of the issues.
It's getting harder to argue that point. I have taken a wait and see approach with this thing.
I keep waiting for data points,
but I do have to admit,
every time something is missed
that is somewhat kind of obvious to the community,
or a deadline or a commitment is missed,
or something that's said that doesn't seem quite right comes out,
I kind of do ratchet up the old skepticism another small notch.
And so it's now a few notches in at this point.
And my wait and see approach is still holding,
but the data points are starting to build a narrative that doesn't look good.
And I'll tell you one that really kind of rattled me is
I was giving them a much, much, much more benefit of the doubt kind of approach
a few months ago.
It was about a year ago maybe,
when the quote-unquote final board pictures were starting to surface on Mastodon.
And they were big, and they were like clearly developer boards.
And so it seemed, and to me it's like, well, these are really rough at this point.
But I was getting lots of messages from people on Telegram saying,
hey, look at this, look at that, this is going to be a thermal issue. This is going to be an
interference issue. And they were identifying stuff just by looking at these pictures with
a board on the table. And I kind of just set all that aside, didn't really even bring it up on air
because benefit of the doubt to Purism, they likely know what they're doing. They're testing
this thing. You know, they're working the firmware. They got this thing figured out.
But then the Aspen batch was delayed. Only a
handful of units ever were made, and they have thermal issues, exactly like people said they
would, in exactly the way people said they would. And Purism's solution is even the solution people
said they'd have to come up with, which is going to be super difficult. And now we're starting to
see that again with their approach to solving the antenna issues. Because you have a bunch of individuals
that are in the industry. And so they're following the Librem 5 as a bit of a hobby.
This thing is fascinating to them because they work with these components. You know,
they build the ATM machines and the IoT devices that use these parts.
And so they're fascinated to see a company try to make a phone out of it.
And they're very familiar with how all this stuff works.
So people seem to be seeing some of these issues from thousands of miles away by just looking at the pictures.
Yeah, and don't forget that they've got a number of other issues that are in play here from a hardware standpoint.
So if you look at the latest videos that the Linux Gamer has put out, he's got three videos with Todd Weaver where they, you know,
Todd came to his house, presumably in Maine, and he had two phones with him and he let him play
with it and they tore it down and they put it back together. Well, as part of that video,
he actually admits to a lot of the hardware deficiencies that they somewhat admitted to
in the Supplying the Demand blog post that you referenced. And among those were, you know, the camera doesn't
work right now. Bluetooth doesn't work. There's maybe a thermal throttling issue, but it's not
that big a deal. And we're going to add a heat pipe. And he did actually mention the antennas
there as well. They were going to have to be the next version of the case would have the antennas built into it.
And, yeah, so there's a variety of these issues that just keep coming up.
And then in addition to that, there's also in the public issues tracker, there are current open bugs regarding call audio issues.
So there's an issue with call audio echo.
Like when you speak into the phone, you can hear your own voice being pumped back to you. And then there's an issue with call audio echo. Like when you speak into the phone, you can hear your own voice being pumped back to you.
And then there's an issue with lag.
And of course, some of this will be resolved over time.
Oh, yeah.
I think what it all comes down to is two core issues here.
A lack of transparency around what the current hurdles are.
Another example is, let's be really clear, like a matrix style,
clear on what blobs are required. Does the Wi-Fi require firmware? Does Bluetooth? Does the modem?
What about booting? Do those all require binary blobs? I'd like to see a matrix of what works,
what's free, what's open. Let's be really clear about those limitations. Let's be really clear
about how well this thing's going to work as an actual phone. If I get in a car accident, can I call for help?
I really need to know that.
So there's a lack of transparency that just causes suspicion.
That's core issue number one.
And the second core issue that I am observing is that purism refers to some of these challenges
in terms that seem to simplify them.
Eminently solvable challenges is a direct quote from their recent blog post.
And then electrical engineers and developers of other mobile platforms, they look at the
problem, they go, that is a very, very unique challenge.
To the antenna specifically, a lot of what I'm hearing is, well, that needed to be designed into the case
when you designed your case. Like you had to have that figured out seven steps ago.
Not something that you're just going to imminently solve. So some of these problems
are much more challenging than they're making them appear on the surface. And that again,
goes back to the transparency, but also leads people to doubt their future statements.
And then when these things play out to be true, like the CPU throttling issue, like
the thermal issue, like the antenna issue, and it seems to have caught them by surprise
and delayed the batch, you do wonder if perhaps they don't have a good grasp on the challenge
of this.
Right.
That's kind of what it makes it feel like.
I mean, no, it's okay.
We know that they had to figure a lot of things out, right? That's part of the challenge here. But we're hearing about it in these overblown terms that don't seem to match with reality. And it's taken a long time to become clear about what actually is current. I think we can all accept that there's going to be blobs, but we need to know what they are so we can be real about the product. Well, I think part of the problem here is that, especially as I've heard from some of
the people who were inside the company that I've been chatting with, is that there's a
difference really between purism as a whole and then what the CEO does and says.
I mean, if you start paying attention, you'll notice that the CEO is the one pushing out
a lot of these updates and a lot of these posts.
You'll notice the CEO is the one pushing out a lot of these updates and a lot of these posts.
And as I understand it, there's a culture issue here in terms of when people come to him and they say, look, this may not be entirely accurate.
He is not particularly receptive to that kind of feedback.
Well, Jay, I want to share with you my motivations on why I've dug into this, how I found you and all of that, because I think it's important for the audience to hear this. I really believe that the free software community is filled
with optimists in a good way. I mean, why else would we be fighting for the things that we do?
And we believe in this stuff. And I think that's really great. And I think it's really important
that it stays that way. I think it leads to a very common, well, you got to give them the benefit of the doubt
syndrome.
Yeah, you know, you got to give them the benefit of the doubt.
What they're trying to do is real hard.
And, you know, it's going to slip a little bit, but their intentions are super good.
So you got to give them the benefit of the doubt.
No.
Well, that is the very mindset that scammers can trade upon.
You have to be skeptical because the greater goal, the quote unquote movement,
is more important. And if they ruin this, it will set humanity back. Yeah, I agree. I mean,
I think from my perspective, you know, the entire free and open software movement, it's about freedom, transparency, and empowering users. And it's hard for me to get to a place mentally where I can see how a process that doesn't embody any of those standards and doesn't take on any of those principles can give us a device that's supposed to be the, you know, the ultimate
representation of those principles. Right. We don't need Neo Apple. We need transparency.
Yep. Well, Jay, you have three really thorough blog posts about this. Also, you follow this in
other, you are, you're, I should say, active in other outlets that discuss this.
And you've really done some good observations.
I know you've been speaking with folks at the company.
And I just really appreciate that you've taken a balance for you.
You're not hating on them.
You and I both would love to see this thing be successful because we both appreciate how important it is.
But you're also staying skeptical.
And I think you've done a good job
of telling the
more complete story
than what we could
touch on
in a segment
for this show
so I have a
link to your
blog
jaylittle.com
in the show notes
and just keep an eye
on things
maybe you and I
can chat again
in the future
if something
significant develops
yeah definitely
I'd like that
thanks Jay
thanks
I'll try to put links
to as many of
the things as we talked about quickly mentioned in that chat in the show notes. And, you know,
going back to the point of transparency, which I think is the key takeaway on this,
I think it's such an opportunity for them to set the bar on this, to really lead the way.
lead the way. Purism, they talk about the Libram Five as the beginning of an ethical movement.
Movement is a big term, and I think they might be right. And so I think because that movement is so important, it's worth their time, even when they're busy solving antenna issues and trying to build something new.
It's worth their time to be more transparent.
Set up a nice matrix of what's a blob.
Does the Wi-Fi require a blob? Does the Bluetooth?
What about the modem?
What's the boot environment like?
I'd like to see a chart of that.
Show it all to me right there in a post.
This really lets
me know what the limitations are of this thing. And that, I mean, at the same time helps us,
you know, be able to get behind them because we know just what's going on.
And I just, you know, I was speculating with Jay. I said, you know, I wonder if,
you know, really if the right thing here wouldn't be for Purism to work with the Pine Phone folks and come together on this
thing in some way. I mean, it would mean putting aside creating the next big free product and
instead focusing on the overall movement. But if you brought together the different talent sets,
I mean, you'd really have something here. It's just something I'd love to see addressed because I worry that we're slipping in the wrong direction.
Right.
I mean, the goal is important and we're getting caught up in the specifics right here.
Yeah.
You know, and part of me kind of like, part of me kind of just wants to look around and say, oh, it's way worse than we've been letting you know and we need another million dollars.
Will you fund us?
And let the community just decide and say, yes, we'll step up, we'll do it because it's that important. But we need transparency.
You know, something has to change. A lot of people have a lot of hopes on that device.
We'll keep an eye on it. You know, we tend to really cover it when something major changes that
the real landmark thing here was the end of the Aspen batch.
Yeah. I mean, now you just got to stay tuned and see how the additional batches go.
Yeah.
Time will tell.
Time will—time tells all stories.
That's been my recent thing, Wes.
I've been thinking about that.
It's like a lot of times you get bits and pieces, but if you stick with it and you're patient enough, eventually time does tell all the stories.
Well, we got a super quick housekeeping.
It's already clean in here.
We should probably mention that we're going to be digging into software
and setting up things in Fedora and all that kind of stuff.
You should definitely check out the Choose Linux podcast.
Do it.
Drew here is one of the three co-hosts of that there show,
and he can probably tell you on some authority that software, Linux,
and installations will be a topic on a future show.
Huh?
Yeah.
Coming out Thursday, we're going to talk about Chaos Linux, which is sort of a KDE showcase.
And on top of that, we're going to talk about Flatpaks and Snaps and all of it.
Flatpaks and Snaps and app images, oh my.
And what are the different experiences of them like, really?
Especially, you know, from different perspectives.
I also want to give a shout out to Ubicon Europe.
Wrapped up and now their YouTube channel is loaded up with some of their presentations.
I think most of them.
So we'll have a link to that in the show notes.
Or you can just go Google search Ubicon Europe.
Yeah, got some good stuff.
Like applied security for containers.
Public and open source. There's really a full range there, actually. Yeah, I got some good stuff, like applied security for containers. Public and open source.
There's really a full range there, actually.
Yeah, totally.
Including stuff like Ubuntu Wellness.
What's that about?
Yeah.
And last but not least, I'll just mention, if you love our shows,
be sure you check out the new Headlines podcast.
This is, I think, this is us at our best now.
Oh, yeah.
So if you like what we do, I really encourage you to go check out the new headline show every weekday. Linux headlines in three minutes or less. Wes did one today. I
did one on Monday. We do kind of a handoff system. But what's even cooler about it is behind the
scenes, we've really staffed it out. We've got people that are dedicated to research. We do
cross-checking. We all are, like, very involved in the process.
Not only is it a full-team effort,
but it's one that we've taken very seriously,
and we've made it available for your Amazon LadyTube.
That's right.
I've got to get it outside the U.S.
I haven't done that yet.
It's complicated.
Their UI is horrible.
It's just horrible, but it is on there.
Linuxheadlines.show, though, if you just want to go check it out.
It's, I think, some of our best work.
Three minutes or less.
Fedora 31 lands today.
It features Linux kernel 5.3, Mason 9.2, GCC 9.21, GNOME 3.34, Perl 5.30,
and probably most importantly,
it doesn't see any Python 2.
No.
Get that out of here.
It's gone now.
Don't need it.
Yeah, it actually,
it does change things a little bit.
I was doing some benchmarking that required Python 2 scripts,
and I felt like an old man.
Right.
Also, no root logon for you anymore
if you're connecting over SSH.
They've been disabled by default.
And then brace yourself, kiddos,
because these next two are the biggies.
They've switched to cgroups
version 2 by default. We'll come back to that.
And they've removed the
i6-86 repositories.
And also, I think they're not
even making bootable images.
That's a big change. But let's talk about that cgroups
thing, because cgroups, Wes, isn't that containers
and isn't everything using C groups version one?
How can you just switch to version two?
Well, it depends on what container runtime you're using,
of course, but for you, I know that's Docker.
And the Docker binary you're going to run right now
expects C groups V1.
So that's probably the biggest,
uh-oh, why isn't this working sort of problem
you're going to run into right off the bat.
You just need to go add a kernel command line
flag. It'll switch that off. You'll still get
Cgroupsv1. They can coexist at the
same time. Then it works.
Yeah, it's still possible to run version
one.
Sometime this year, maybe
next year, we should see Docker
working and other runtimes already support
it. So if you're using Podman, away you go.
Including cgroupsv2 brings rootless container, better rootless container support, which integrates with the new API.
That's just what I was going to ask you.
So that's why.
That's really the big, that's like the big thing you gain from this.
It's kind of a, it's just a, it's a cleanup of cgroups.
I mean, cgroups were a new idea in the kernel community, right?
As we were exploring how to group and contain all these processes.
Turns out, you know, you don't always get
everything right the first time, so
gotta fix it up. It will be
painful, but Fedora's
doing what Fedora does best
and, you know, putting their toe in the water
and saying, it's gonna break some things, but
we're making the switch. In fact, they almost
kind of say exactly... That's the rationale on the wiki.
Yeah, that's why they're doing it.
Like, somebody's got to do it.
No one's doing it.
And this, like, was,
Z Group's version 2 was finalized years ago, so.
Yeah, exactly.
I mean, I think it's been non-experimental since 2015.
And I think that is true to some extent, right?
I mean, it does change when you're a legitimate Linux desktop user, right?
And you can go file bugs on other projects saying,
your project is broken for me now.
Of course, they have a bunch of ARM support, but Raspberry Pi 4, I don't think landed officially.
Darn. Not that I saw. But that's okay. That's all right. That's fine. We did have some fun since we had just recently done some benchmarks on stock Ubuntu 19.10. We ran some benchmarks on stock
Fedora 31, just did a little comparing. I mean, once you've got the Pharonix test, we can install them.
It's crazy.
Nothing really all that interesting to report.
Ubuntu is faster, I'd say, in the overall takeaway, but there are some areas where Fedora pulled ahead.
Typical benchmark.
If you're interested at all, though, we'll have a link in the show notes.
Not all the benchmarks fully ran on Fedora either, which is kind of a comment about running software on Fedora versus Ubuntu in a meta way.
It's not so bad.
I mean, it's really not.
But like, okay, you're running some random script you found on the internet or from the readme of a GitHub page.
They almost always have Ubuntu packages.
You don't always get Fedora.
Well, okay.
So let's talk about setting up Fedora 31.
Fedora.
Well, okay.
So let's talk about setting up Fedora 31.
I think a couple of years ago, maybe even a year ago, I don't know.
I don't know.
It's ish.
Let's just say ish.
It was good advice to say wait a few days, maybe a week or two,
before you upgrade to the next Fedora release.
You go back a little bit further in time,
the advice was wait maybe a month or two before you upgrade to the next Fedora release. That's legitimately
how it went. Boy, has
that changed. Now, I'm not
advocating anybody do like Chris do
or Wes or
all of us on the show.
But I
upgraded to Fedora 31 yesterday, which
was the day before release.
Because they got, you know, a little birdie told me that the ISO is all finalized.
Everything's good to go.
We're pretty much ready.
Yeah.
So I was having some problems.
And I thought, well, I've been running Pop! OS, liked it.
But I had a little mouse leggy issue, just a little bit of mouse stutter.
And it's like, eh.
Yeah, we know you have a low tolerance for that.
I have a low tolerance, I know.
So I thought, let's put Fedora 31 on here.
I know you have a low tolerance for that. I have a low tolerance, I know.
So I thought, let's put Fedora 31 on here.
I also, because I like chaos apparently,
decided let's upgrade my Lenovo ThinkPad T480,
which has been running like a freaking champ with Fedora 30
on the road, in day-to-day use, every single day, every single night.
This is kind of your golden boy.
You've been nothing but happy with this install.
So I thought, why not scroll it up and let's upgrade to Fedora 31 the day before the release.
What could go wrong?
Turns out, nothing?
Yeah, nothing.
Nothing.
I had to uninstall Crossover Office, unfortunately.
And I had to uninstall, I think, Lavish, was it?
Two packages. Yeah. Not much, Lavish, was it? Two packages.
Yeah.
Not much, though.
It was kind of surprising.
And the whole process sort of rekindled for us, I think,
like what we really like about Fedora.
I went digging around in some of the scripts running behind the scenes
to do the system upgrade and how that integrates with DNF.
And Fedora's just so clean and modern.
It's a fresh, well-put-together Linux machine.
Oh, the peak experience for me,
I did the little like dash dash release equals 31
to change over to the new version.
And the RPM Fusion folks thought of this ahead of time.
And so it just automatically follows whatever you
set that release version to.
So the next time DNF did an update,
I was just automatically getting the RPM Fusion
31 repos. And to their credit,
those repos were already up and ready to go.
Yeah. Same with like the
couple of Copper repositories
I was using already had their 31
Copper repositories up.
It's so easy because then the rest of it is all covered with Flatpaks.
Right.
That doesn't really change.
You just got to make sure you still have Flatpak.
They don't care.
Yeah.
And so I took a heavily modified dual desktop, you know, because I had XSE on there too,
and GNOME Shell.
Tons of like crazy jack stuff we've done,
my eGPU experiments,
my PCI passer experiments,
all the things I've been doing to this machine since Fedora 30 came out,
it all just went up to 31 flawlessly.
Even my wire guard,
bolted on wire guard support,
upgraded.
It was a nice experience from the end user.
I mean, just watching you go through it,
you kind of sat down,
knowing that this might take a little while. And that transaction test that it
does sort of gives you a little bit of peace of mind. Yes, right. It's not, you're not just sort
of going off into the wild here. There's some checks and boundaries. Now, all that said,
I have had some problems, but not on my laptop.
So I want to make this clear.
This is unique, I think, to hardware setups.
And I've been trying to track down some of these problems I've had,
so much so that I've looped in Wendell and Michael Larabelle trying to help me troubleshoot.
You're desperate.
I really am, because the mouse leg issue that I was having on Pop! OS on my
main workstation got
significantly worse when I
went to Fedora 31.
It got bad. I had mouse lag
that sometimes would last 10 seconds
like where a mouse would just lock up for 10 seconds.
Oh, just hard lock. There's other times where it's just
you're using the system and the mouse
would just lag across the screen
so it would jump halfway across your screen.
Have you considered just switching to i3?
You son of a bitch.
So I hard bailed on Wayland because I did a little digging.
Actually, yes, I am thinking about it.
I was actually thinking about going back to Plasma.
I did a little digging and this is a problem that is getting worked at right now.
There was a recent post by canonical employee Daniel Van Gogh.
I don't know how you say his last name.
It's not Van Gogh.
Great guy, super sharp developer.
And he writes in this post that we have linked in the show notes, quote,
The GNOME Shell and Mutter are single-threaded GLIB event loop apps,
so any pause could cause them to miss the next frame and exhibit stuttering.
Some common pauses are for disk I.O. or GPU I.O.,
but the miscalculating when to render the next frame has also been an issue in GNOME Shell.
Here's the fun thing.
This, way worse on Wayland because of issue in Gnome Shell. Here's the fun thing. This, way worse
on Wayland because of the architecture of Gnome Shell.
And
Chrome's a real son of a bitch
because it's using GPU acceleration
so it's causing GPU IO.
And it's caching to your disk
so it's causing disk IO
so it is the perfect storm
of events that cause
Gnome Shell to lock up.
So I thought, I'll go to Firefox.
Worse in Firefox.
Worse.
Significantly worse.
Wow.
Unusable.
Like, can't select menu items at times.
Not all the time.
About three or four or five times an hour.
So it's just enough that you sort of put your guard down,
and then all of a sudden, boom,
you're hit with this horrible mouse leg.
It's better on X.
It's like 95% better on X.
Workable.
Every now and then I get some leggy legs.
But it's not that bad.
It's just it really stinks because, first of all,
the graphics card is an AMD RX 570.
It's got 64 gigabytes of RAM. It's got is an AMD RX 570.
It's got 64 gigabytes of RAM.
It's got six physical cores, 12 threads.
The main root OS is on MV and E drive. My home is on a fast Intel SSD.
My games, my Steam games, are on a separate Intel SSD.
And all of my scratch work is on a RAID 0 of spinning rust.
All of this is on separate controllers.
Like, I have built this system to optimize disk I.O. and performance
when I'm doing lots of transactions or, you know, encoding or whatever it is.
Like, that's what this system is designed for.
And I'm getting a significantly smoother, better experience on my T480
because it's single screen Intel graphics.
And I really think that my screen setup continues to influence this
because I've got three 2K monitors, two of which are vertical, one is horizontal.
It just had to be different, didn't you?
And reading through the bug reports about these stuttery mouse issues,
first of all, it could be caused by a million different combinations of things.
But the other thing is,
it does get exponentially worse
when you have extra monitors
because of how the whole stack works.
Of course, yeah.
Oh, man.
It has been really frustrating.
Additionally, I got a new sound problem
that I've never experienced before.
Oh, no, really? Yes. This is both a good and bad. frustrating additionally i got a new sound problem that i've never experienced oh no really yes
this is both a good and bad the bad is say i watch a youtube video in a browser
video ends and i choose not to listen anything else five ten seconds go by, maybe 15 seconds, and I hear this audible click in my speakers.
If you've ever had an optical audio connection and it goes to sleep after some inactivity, it sounds just like this, only this is a standard, regular old audio cable.
Once it goes to sleep, my speakers buzz.
And I have a pretty big subwoofer so it's
which is crazy distracting
so between the mouse leg
and the buzz
I'm going nuts
however
the positive side is
unlike Pop and Ubuntu
which were constantly defaulting
even after I change it
to my monitor's built-in speaker
which sucks
this is actually remembering my speaker selection.
Excellent.
Yeah.
You just got to always be playing background music, and then you'll be fine.
The funny thing about Fedora 31 is its positioning
for just this total change-out of the audio system
and migration to Pipewire in 32 and 33,
and then eventually completed in 34.
Like, they've got this long arc plan for Pipewire,
and 31 is all about getting things in position for it.
But there's nothing really significantly different there,
so I'm not sure why I'm having these problems.
However, that stuff said, I really am happy with my install.
It really is quite nice.
The underlying Fedora system remains my favorite.
It's reminiscent of Arch in its cleanness and its tightness.
Yeah, it's modern.
It doesn't feel bogged down by a bunch of craft or legacy.
It's very logically laid out.
However, it feels more like a complete product.
And it stays very up to date,
but with these sort of thresholds.
It has almost the feel of a rolling release distribution
because you get new kernels,
you get new core things during a release,
but it doesn't have the roller
coaster experience that you get with
the really cutting edge distros.
No, I mean, you can feel
comfortable using it on your workstation
as long as you know, you know,
fairly often you'll have to do an upgrade.
I think Fedora
for me fills
the role that Manjaro fills
for a lot of people in the Arch land.
They want that clean, tight, professionally built OS with really fresh software.
But it's, you know, maybe having a testing layer is a good thing, which Rawhide provides for Fedora.
So, man, there's a real, I think there's a real common experience between a Manjaro user's preferences and experience and a Fedora user's preferences and experience.
And I just happened to like the Fedora take on it in part because it dawned on me while doing this review that I have been trying out different Red Hat desktop installs for 21 years.
This marks my 21th year of trying these. And so I may just be a little more inclined
to like the RedHat version more than the Manjaro version, but I completely kind of get the appeal
of it now. Well, I mean, even if I've been playing around with it, it just feels like a nice
modern workstation, especially if you're like, you know, someone who uses Linux and also does
work in the cloud or development, or I love that Git is just already installed. Oh, I know.
Carl, you were saying they kind of have a term for Fedora
that sort of fits this description.
Yes, sir.
Matt, his last name is escaping me right now, Matt Diem,
he calls it a semi-roll-in release.
That's a good way to put it, Matthew Miller.
In fact, I reached out to Matthew and invited him to come on.
He's at a conference.
I invited him to come on next week so we can,
after we spend another week with Fedora, we can chat with him. So come back
next week to chat with him too. So we all gave it a go. I have really grown to like Fedora as a
workstation. So I wanted to really kind of spend some time to go through it. And I don't know what
I'm going to do. I think I'll just hang on X where the mouse leg is acceptable and wait for the next GNOME release because this is something they're actively working on.
What about your experience?
Oh, no.
It's been very solid because I've only been using it on the ThinkPad.
Right, where it's glorious.
And you've been using Wayland?
I've been using GNOME and Wayland, and I've not yet installed Chrome either, so I've been using Firefox.
Honestly, it's just been clean and minimal and kind of perfect.
I mean, you and I, yesterday we did a little bit of tweaking, playing around with GNOME themes.
But other than that, it's just ready to go.
And I really love that about a nice modern workstation.
It was easy to get all the software I needed to do work, to prepare for the show.
So I'm kind of back up and running already.
Yeah, I would say now,
Gnome on Wayland takes me about 15 minutes.
That's it, yep.
And I could probably spend all day tweaking Plasma.
I kind of enjoy it sometimes,
but sometimes I don't want to bother with that.
I want to talk about the installer and all of that,
but Cheesy, I think this is a good chance for you to jump in because you kind of have the freshest eyes on the Fedora
installation experience. I have grown more and more familiar with Anaconda, and I know that can
be a little strange for somebody that doesn't install Fedora systems all the time. Yeah, so,
I mean, full disclosure, I'm not an avid Fedora user. Overall, the experience has been great,
but yeah, so just going through the installer, one of the things that I noticed immediately
was whenever you go to format your disk, it asks you to reclaim space, which just seems
kind of weird instead of formatting.
And I'm kind of used to that Ubuntu installer with the kind of pictogram and a short description
of what you're about to do.
So that was a little weird for me.
So I could see a new user coming into that
and maybe getting to that point and being a bit confused.
Yeah, the disk partitioner stuff is a bit strange to begin with
because you've got automatic mode, which is still a little strange,
custom, which is sort of like the custom partitioner in Ubuntu's installer, and then you
have Beveled or whatever it is, the more
advanced one, which is a whole other
disk partitioner. It's my preferred one,
but it's hard to wrap your head
around, even if you've done it multiple times.
And here's what I can tell you. The first couple of times,
because they changed this a few years ago, the first couple
of times I used it, I was
really very unhappy with it.
The more I use it, and the more I install other distros, the more I appreciate it, because this is really the only one that lets me
properly take advantage of my LVM setup, and really kind of clearly lays out all the disks,
and lets me kind of do the more sophisticated partitioning that I want to do properly. It's
the only one that will do it.
And what I've noticed is the more I use it,
the less offensive I find it.
And I think what's happening is they have reached that phase themselves,
the developers.
To them, they install Fedora,
and it seems perfectly usable because they're totally used to it.
But if you come in fresh,
it's a weird way to lay out your disk.
Even the language it
uses constantly hitting the done button double done buttons and reclaim buttons and yeah the
whole thing is strange well and then of you know you whenever you do that initial setup you don't
set up your initial user until after you reboot which again was a little bit strange and then
whenever you boot in the first thing you get is a welcome to Fedora screen
that you have to click next on
before you can begin your user and setup.
This is kind of the new way to do it,
this post install setup,
because one of the things that System76 says,
and I think Elementary OS said the same thing,
is it's a great way to then take advantage
of using like encryption that an OEM can ship a system,
and that last phase can be just handled by the user,
and then they can set up their own encryption password and stuff.
So there's some advantages to it.
And that's exactly how I felt about it.
I felt that it was more of, if you're coming from Ubuntu or something like that,
it was an OEM-type setup, which, you know, no harm, no foul.
But then, yeah, whenever I was greeted with the welcome
to Fedora screen before setting up my user and you just have to click next, it seems like that's
one step that could be alleviated just to get you right into setting up your user.
So that was a little weird. Searching for some of the software was a little weird. So initially we're trying to set up a telegram. Um, they had the
telegram web version in there, which is just not good. Um, especially if you're older like me and
you need that scaling to like 125%, right. To kind of get it where you need it. Yes. 125% club.
Exactly, man. So, uh, but you know, that didn't take much to get into flat pack get the you know flat pack
repost set up uh pull down telegram desktop without any problems had my scaling back um so that was
that was pretty dead simple to get going i didn't notice though that whenever installing apps
sometimes they would seem to hang for a bit um so like two or three minutes installing without giving you any progress.
Yeah. Which was kind of weird. I think it's like pulling down different flat pack layers
and dependencies and they don't have a good progress signal for that. And that's exactly
what I was thinking is that it was pulling down the layers in the background and just not
outputting that onto the progress bar for me. Yeah. But initially, you know, I got all that worked out, got everything set up.
Overall, pretty stock GNOME experience.
I do like that they had the hot corners set up out of the box
because that's one of the things that I do with default GNOME.
I do like the Pharonix test suite was already there in the shop.
So, boom, just, you know, I didn't have to go to the Pharonix side
and pull that down.
It's a little, it's one version behind.
Love to see that updated.
Just saying, just a note out to the universe.
But the fingerprint reader works
just like, just as well as it did on Ubuntu.
So I think that's just baked into GNOME 3.34.
So that works great on this little ThinkPad.
Overall, I mean, it really seems like a solid distro.
I would like to see maybe the Flatpak repo added in too,
just to give that little more extensibility
to have those apps that I want.
But I do understand that that could be a security concern.
So that might be one of the reasons they don't push that on you. Maybe a middle ground there could be
a nice way to just add a checkbox for the FlatHub repo, because that is a, it is a third party
community initiative. And, you know, I don't think Fedora wants to vouch for the quality or
safety of the applications on there, but, you know, if they could come to a pretty
easy way to get the NVIDIA graphics working, I think we could figure out a way to get FlatHub.
Yeah. And I ran into the same issue with Quassel. Whenever I was downloading Quassel,
I could get the full Quassel, but the FlatHub had the Quassel client, which is all I needed
for that laptop. So there were a little few things that were different that took some kind
of getting used to, but I mean, overall, it's been a great three days, three or four days using it.
I would think it would behoove them to make it easier to integrate Flathub into GNOME software
simply because Ubuntu 19.10 ships with Snaps integrated,
and you get a much larger range of software.
It's nice to have in there, yeah.
And just like Flathub has a lot of great apps.
And it is now a little bit different.
Depending on which distro you're using, flat packs,
or you're using the distro packages,
there's a lot of little decisions you've got to make.
Plug and plug again for that ChooseLytic.
Speaking of which, Drew, have you played with Toolbox?
Because that's something that the Fedora project
is really putting forward on this release.
I have. I've played with Toolbox and Silverblue, which is where you really see Toolbox as part of the base operating system.
You can get it on Workstation now, but it kind of grew out of their work with Podman and wanting to make Podman easier to use in Silverblue. And it's a really cool project.
Basically, it's a way to spin up an OCI container on the fly or keep them around and do work in
that container off of essentially a base image that you can then drop in extra applications or things that you maybe don't
want to put in permanently into your own base system. Yeah, maybe you don't want to muck it
up. And here's some of the neat things that it does is it uses your existing username or
permissions. So you're just boom, you're up, and then it gives you a quick way to jump into that
containers command line, you get access to your own home directory, and you get all your command line tools, including DNF. Containers have come so far. It's
just amazing. Yeah, speaking of Silverblue, Silverblue doesn't have DNF in the base system.
So if you want to look up packages using the command line, that's how you have to do it,
is you have to drop into a toolbox, pull the package name, and then drop back out of the
toolbox and install it via rpm-ostree. So you were our man on the scene over there in Silverblue
town, but it sounds like after some initial success, you wound up back on Workstation.
I did. Yeah, I had some issues with it. I think I got a bad update at one point.
Wait, wait, wait, wait. I thought the whole point of Silverblue was to prevent anything going wrong with bad updates. What happened?
Well, yeah, so I'm not exactly sure what happened to my system. But the idea is that you have an
immutable base. And you can kind of shift up and shift down like I went to Fedora 31, you know, early, early beta,
and I was able to roll back to Fedora 30 just within my boot menu. But something happened.
And anything that I tried to boot just resulted in a black screen. I couldn't switch to another
TTY. I couldn't get anything done. and I had to get rolling. So I just
paved it and went to the workstation 31 beta. So I think it's still a little too experimental for me.
Now I haven't checked it out with Fedora 31 stable, but I have been keeping track of what they're looking forward to in 32.
And I think that they're going to really start, you know, taking down these little bugs and getting Podman and Toolbox where it needs to be and have everything really tightened up.
It's looking good for that.
So I'm hoping all of the feature sets, you know,
make it into 32. And in another six months, I might be giving it another go.
I could see it. I could see how technologies like Toolbox will evolve over time.
I'm going to, after this release, because, you know, they kind of made a big stink about it,
I was like, well, I'll check it out. I'm going to keep using it because it allows me to have
not only a couple of different user environments
I can muck up,
but it's super easy to get started.
And that kind of technology
wasn't really even around when Silverblue started.
I mean, the tech was, right?
But the implementation of it
wasn't around when Silverblue started.
That's the hard stuff here, right?
It's like making it actually useful to you.
Right.
And Flatpaks were nascent.
And things like projects like Flathub hadn't even begun yet.
And some of the advantages that Gnome Shell has with you going over to Wayland are only
now becoming realized.
All of these things are sort of happening around Silverblue, which seems like it's going
to make it a better candidate once they've all kind of crystallized a bit.
So I could see how in six months to a year, it could be in a much better state. Yeah. And for me, part of the
problem that I have with it is there is absolutely no I-686, no multi-lib, none of it. Right.
Which is bad for audio. It is because I use a lot of things through Wine to kind of help out in my workflow, and I can't do that.
So my options are run all my audio in a VM or possibly try to pass things through the toolbox, but that gets a little messy.
And yeah, it's just not the best now.
I'm hoping that the future of Pipewire helps to make things a little better for me down the line.
Or other projects that may be in the works.
You know, there's other ways to solve that particular problem, which Wes Payne's been taking a crack at.
I still think things are looking pretty good for 32.
Some of the story of 31, to be quite honest honest with you is just about the performance improvements
which we talked about
with Ubuntu 19.10
right?
Yeah, indeed.
So it's
if Ubuntu 19.10
hadn't happened
the majority of this
would be about
the performance improvements
to Gnome Shell.
They're all there.
They're great.
It is great.
I'm really enjoying it.
So I guess it's worth
an honorable mention
and it's really kind of nice to see it all materialize.
So what do you say?
Will you give it another try in a little while and report back, Drew?
Yeah, absolutely.
I think I'm going to give it some time, let them really get 31 up to speed, and look forward to 32 and see how they're doing for that release.
Yeah.
And in the meantime, Workstation will do the job, right?
Absolutely. Yeah, I've been really happy with 31 Workstation. And I haven't had any mouse lag issues,
even though I'm using a very similar setup to what you have on your desktop
with dual monitors and, you know, Ryzen card and AMD CPU SSDs. Yeah, but no mouse lag for me.
I think it could even come down to like what USB port or bus my mouse is on. Like it's just so
many variables. It's three monitors to be clear. It's not two, but it's still kind of the same
thing. And I wonder, I wonder if it's not just some unique combination
of hardware that will never be solved
on my system.
Because, and the reason why
I know it's unique to that box
is because it works so brilliantly
on all the other systems.
Well, at least that means it's solvable.
We just got to get you a new rig.
Yeah, that's it.
That's it.
I just got to get a brand new computer,
a really nice, fast computer.
I did a brief rabbit hole deep dive
into schedulers with this Fedora release because they switched to the BFQ scheduler by default.
And now I have learned that Linux applications can actually choose from a list of schedulers
that the kernel can make available, which blows my mind that this is even a thing.
But just as of note, Fedora 31 switches to BFQ,
which has some people excited.
So Alex, last but not least,
I know that you've been giving the new fresh release a go
as we've been recording.
Do you have any takeaways before we wrap up on the new release?
I have about three or four different LVFS firmware updates to do.
Oh yeah, by the way, that is one of the funny things when you install Fedora or something that does LVFS firmware updates to do. Oh yeah, by the way,
that is one of the funny things
when you install Fedora
or something that does LVFS
and you haven't run it before,
you've been on a desktop environment
that doesn't have it.
It's like, oh, I've been neglecting this.
Are you on a laptop?
Yeah, it's my ThinkPad T4 ATS.
Oh yeah, of course.
Yeah.
Yep, I had the same three updates, I think.
I just have a hard time with GNOME.
This is my issue with Fedora at the moment.
Like, as a server OS, I use it every day for sort of hacking around and stuff.
But I don't know.
I just, GNOME.
Ah, I just can't.
No, it's so clean and smooth and gets out of your way.
Simple.
Until you have mouse lag, and then you want to burn it to the ground.
I think I might go back to Pl plasma for a little bit on that machine.
I want one environment on all my machines so that way I can muscle memory it.
And then I want to be that individual that's got the same environment for a decade,
and so they just are one with the UI of their computer.
Because I know a guy who's been using XFCE,
well, besides Joe,
but Joe would probably be also this individual.
He's been using XFCE on this old computer for a decade.
Same computer, same interface.
He just upgrades Debian every time that a release comes out.
And he is like a wizard.
I mean, he moves around and his, I mean, and I move pretty quick, Wes.
Indeed you do.
And this guy is like a wizard, I tell you what. And in part, it's because in his fingertips, he knows where the elements of the UI are.
Have you considered switching to i3?
All right, let's get out of this.
Let's get out of this. Let's get out of this
and let's talk about
our pick this week.
I really like this pick.
As Wes always says,
it turns out if you try hard enough,
you can.
I wanted a great way
to get remote terminal access.
And so what I was doing
for some of these benchmarks
that ran for hours and hours
and I wanted to get out of here
is we'd started up
in a TMUX session
and then you'd have to like SSH into one.
Well, first you set up WireGuard
and then you'd SSH into the server
and then from there you could hop into the box
or maybe you'd go directly from the box
depending on your WireGuard config
and then you'd join the TMUX session
and then it's fine.
It's fine if you got nothing to do.
But it was a few steps.
If you got nothing to do
and all you want to do is play with your computer, it's fine.
This isn't for you.
But for those of us that want to get in, kick some terminal command, and get out,
I introduce to you my friend's teammate.
Instant terminal sharing.
Like nothing we've talked about before, friends,
and they both have a cloud version and a self-hosted version.
Teammate is a fork of Tmux, and they can exist on the same system together, no problemo.
And it's a really easy way.
You start a session in Teammate, and it immediately generates an SSH address with a unique hash
in it.
Somebody pops that into their terminal, and bam, you're sharing a screen.
It doesn't matter if you're behind NAT.
It supports some actually pretty dodgy network conditions.
And the tech behind it is super solid,
all really well laid out on the website.
And this is, I mean,
and then when you self-host it too,
it's nice and secure.
You don't have to worry about them proxying the connection or anything like that.
Right, you set it all up.
You can reuse existing SSH keys if you've got some set up or generate new ones,
and it'll accept them as environmental variables.
Of course, there's a Docker container out there already,
and teammate was in the Fedora repo, so that was easy to get.
Yeah, it is really easy to get installed.
And so this is a super solid way to start something up that you want to finish when you get somewhere else.
You're probably using T-Mux anyway.
Yeah, and if you're not, you really should look into it because this is so, so handy when you know a job's going to be running for a while,
but eventually the dumb, stupid computer's going to stop, and it's going to wait for the monkey to push a button
and it won't finish what it needs to do until the monkey pushes the button
and there's no reason you should stick around just to push the button
when you can SSH to this unique address that it will generate
and you'll be on the box and on that session immediately.
It also supports read-only mode if you just want to demonstrate.
It has some of the feature sets of MOSH as well, which is a pick we've
covered before, which helps restore
connections when you move or
if you're on a mobile connection and
your IP address changes.
You can run it as a service or you can just run it
as a command.
It's pretty cool. We set it up as a server
side service too. Yeah, and it wasn't
too bad to get going.
You've got to go futz with it a little bit,
but that's always how it goes.
It does.
Or you can just,
if you need something
super quick,
you can just install
teammate,
and by default,
it will use
their teammate.io server.
Yeah, they've got
a hosted server.
I mean, I think it's
actually already
sort of geo-load balanced
out there.
So there's probably one,
hopefully not too far
from you.
Yeah, all the code's
free, open source.
It's all available,
and they have packet, you can even get it installed on open source. It's all available. And they have packet,
you can even get it installed on macOS.
It's in Brew.
How about that?
Yeah.
And I think they even have a guide here
on how to get it working on Windows.
Wow.
I tell you what, that's just pretty great.
And of course they got packages for Ubuntu
and it's in the Ubuntu repo as well.
So check it out, teammate.io,
or we will have a link in the show notes.
It's just one of those picks that we have just found so super helpful,
like the magic wormhole, which is a really easy way to send files around.
Not every pick actually sneaks into the tool belt, but this one, I think.
Oh, yeah.
Oh, yeah.
I'm using it.
I'm using it to get into machines in the RV, too,
because what's great about it is those are behind a double NAT,
and I can still use that to get into those sessions.
So I can kick my Shinobi box in the gear, that stupid Shinobi.
Anyways, a link to the picks is at linuxunplugged.com slash 325,
where you'll also find our contact page, our subscribe page, references to everything we've covered today,
all that kind of stuff.
Anything else we need to cover, Mr. Payne? Any other duties we need to today, all that kind of stuff.
Anything else we need to cover, Mr. Payne?
Any other duties we need to perform before we get out of here?
If you want to find out a little bit more about benchmarking, stay tuned.
That's what Jim and I are talking about in this week's episode of TechSnap.
Oh.
Yeah, so dive in.
We talk about FIO, some of his favorite ways of doing that, and what you need to know about IOPS and throughput.
I love the way he explains stuff, so that is going to be great.
And he knows that stuff, too.
He really does, so that'll be good.
Look for that later this week.
TechSnap.Systems for that.
And go follow the network at Jupyter Signal.
The show is at Linux Unplugged, I believe.
Sounds right.
Yeah.
I'm at ChrisLAS.
Go check out Drew at Choose Linux.
And, of course, Alex and I over at Self Hosted.
Selfhosted.show.
It's good stuff.
See you back here next Tuesday. All right, let's go title the show, jbtitles.com.
And then we're going to do a little internet helps, if you guys are ready, okay?
What?
We're going to do a little internet helps.
Well, we need to help out the Linux Foundation.
They're having some communications troubles, and I think we have just the crew to solve it, Cheesy.
Okay, let's do it.
Yeah.
Yeah.
So, anyways, go over to jbtitles.com.
I think something with Fedora
in there would be nice
or a hat full of hassles.
That's kind of fun.
I bet Matthew will be really excited
to come on the show
after that one.
But I do love it.
Purism Throttled
of Mice and Zen.
Where is Aspen? Yeah. really yeah really all right so jb titles
jb titles now we all gotta go boat all right so go over there now it's voting time all right so
let's do a little translation and we're gonna have to start at the headline because even the
headline's pretty thick uh brent i'm gonna ask you to define this first one, okay?
So this one's on you.
Prepare yourself, sir.
All right, I'm ready.
New Linux Foundation effort to focus on data confidence fabrics
to scale digital transformation initiatives.
What does that mean, Brent?
It sounds like a new apparel line that they're coming up with.
Here, I'll put a link in the show notes, too,
so you can read the text yourself.
New Linux Foundation effort to focus on data confidence fabrics to scale digital transformation initiatives.
Here's how I read that.
Oh, my God.
The Linux Foundation has a new initiative that wants to make people more confident in a whole layer of data products.
Don't worry, though.
They did mention AI.
That's in the second paragraph.
Yeah, okay.
So here's the sub-headline.
Let's see, who should we pick on for this one?
Who do you want to pick on?
Should we pick on Alex?
Let's pick on Alex.
Wait, did I win or lose?
No, no, you didn't win.
All right.
Alex, this is the sub-headline.
I'd like you to translate this.
We've only made it from the headline now.
In partnership with leading industry players,
Linux Foundation to create Project Avalarium
to facilitate intrinsic trust in data and applications
spanning heterogeneous systems of systems.
Blah, blah, blah.
Blah, blah, blah, blah, blah.
No, no.
You've got to say what you really think it means. You've got to say what you really think it means.
You've got to say what you really think it means.
I think it was written by someone with a marketing degree
that doesn't understand technology.
No, they're trying to convey a message there.
There is something they're trying to say.
So, all right, Alex.
Cheesy, do you got any guesses on what they're trying to say?
Leading industry players.
Right.
Avalarium facilitate
intrinsic
trust in
data
so
industry players
are
like Dell
IBM
right
and they
are
trying to
facilitate
intrinsic trust
with data
from
various different
applications
on various
different systems
I think that's pretty close I think that's pretty dense systems of systems From various different applications on various different systems?
I think that's pretty close.
I think that's pretty dense. Systems of systems.
Yeah, system of systems.
All right, well, I'm sure they'll explain it in the opening paragraph.
It goes,
The Linux Foundation, a nonprofit organization dedicated to accelerating the growth of Linux
and collaborative development among sustainable open source ecosystems.
Which they are.
Pretty good so far.
Today announced the intent to form the Project Avalarium.
Project Avalarium will focus on building concepts
of data confidence fabrics, DCFs,
to facilitate measurable trust.
Okay, measurable trust. Right.
Now I'm starting to get lost here. Measurable trust
means that somehow
people will be able to
gauge how trustworthy it is.
Right? So that's pretty easy.
Measurable trust and confidence in data and applications spanning heterogeneous systems.
So, okay.
So data in the applications across multiple types of systems.
The product will be seeded by code from Dell Technologies with support from industry leaders including ARM, IBM, and others.
I like the second paragraph, though.
and others.
I like the second paragraph, though.
A DCF, or trust fabric,
is a framework comprised of a variety of technologies that help insert trust into the data path,
in turn facilitating the orchestration
of trusted AI models
and the delivery of data from devices to applications
with measurable confidence.
I'm going to need a nap.
That's all I know.
Can I attempt to sum this up in one word?
Yeah.
Blockchain.