LINUX Unplugged - 389: Harder Butter Faster Stronger

Episode Date: January 20, 2021

We showcase a tool that will change your Linux game. Plus our thoughts on the recent Btrfs FUD, a bunch of feedback, and a handy pick.

Hey, Wes, did you hear the good news? No, I don't think so. Wait, unless you mean Nano 5.5, but that can't be it, right? Nano 5.5 Rebecca is out. Clearly one of the most accomplished pieces of software with one of the most monumental releases in free software history. GNU Nano 5.5 has a new minibar, which will blow you away. It has a new prompt color, which will change your life.

And it has a new set mark match option that highlights the result of a successful search that's going to change the way you search a Nano document. And they've also brought a no-wrap toggle. It's Nano, but not like you know it. It's all new, it's brand new, and it's here to dominate Vi.

Vim? Emacs? You know what stood out to me with this release, Chris? What? Absolutely nothing. Well, hello, friends, and welcome into your weekly Linux talk show. My name is Chris. My name is Wes. Hello, Wes.

This episode is brought to you by CloudGuru, the leader in hands-on cloud learning. The only way to lock in a new skill is by doing. That's what it says right here. And you can get your hands cloudy at cloudguru.com. You know it's true because it's written down. This is episode 389, which, if I recall, is also the LDAP port number. So this is episode LDAP, Wes. Perfect. It really is.

And also, you're absolutely wrong about that Nano thing. It cuts deep, and I will never forgive. I mean, it's nice to have stable stuff too, right? I don't need breaking features and all kinds of changes each time. Nano just works, and that's probably what you want in a text editor. Probably, probably. But we actually have a really, really fun tool that we're going to talk about today that it's going to change your Linux game from this episode forever. It's a way to enable your distro hopping habit in a way that may be frankly irresponsible for us to tell you about, but we're going to do it anyways. But before we get there, we're going to get to the community news. And to do that, we got to say time appropriate

greetings to our mumble room. Hello, Virtual Lug. Hello, everybody. It's really nice to see all of you. We've got a good crew in there. They've joined us by going to linuxunplugged.com, clicking our Mumble link, and getting the deets to connect in. And Mumble's free software is high quality, and it's also the lowest latency way to listen to some of our live streams. So welcome to episode LDAP, guys.

Are you ready to get started? Excited. All right, well, let's talk about something that I didn't really know, and I feel kind of embarrassed, but it's changing. So I'll just pretend like I knew all along that Ubuntu had world-readable home directories, and that's changing in Ubuntu 21.04. They're doing away with this existing practice, and they're making new user home directories, I guess,

no longer world readable. What? Wes, how could this possibly be a thing? Yeah, okay. Well, up to now, when you made a new home directory, they were created with 7.5.5 as the permissions. With this proposed change, it would be 7.5.0. And if you don't speak octal,

well, currently 7.5.5, that's read, write, execute for the owner, read read and execute for the group, but also for everybody else. So with 750, now it's read, write, execute for the owner, read and execute for the group, and nada for everybody else. Yeah, so I guess at least going back to 2006, there have been people that have been asking about this, and I just somehow even missed that this was a thing. I was asking about this, and I just somehow even missed that this was a thing. The original logic seems to be that multi-user systems had some level of cooperation amongst the users anyways, and this made it easier, yeah, yes it does, to share and access files between users on the same system. But of course, with Ubuntu on the server, probably not a good idea, is it?

No, probably not, right? I mean, sure, there are some of these local multi-user systems, but these days it's more like you're going to have accounts running various services or you might have untrusted folks that you just don't want easy access to files if one particular user gets compromised. I've always paid close attention to my.ssh directory and those particular folders, but I've never really given much concern about it because for the most part, I'm just the sole user of a system. And when we set up a server, it's pretty much like you and I

and maybe one other person, maybe Alex, and we are actually cooperating in that regard. Yeah, that is true. You know, there are some of those systems. But at the same time, I think it's often, at least for our local systems, often that we all end up with pseudo anyway. So if I need to access things, I can.

I think it just makes it a little simpler as a default, right? You could still go and change your permissions on your home folders as long as you had the right access to do so, of course, and give folks access. But I think maybe the argument before from SunSense was admins knew what they were doing, and obviously, of course, if you want to make a folder that everyone else can't read, easy to do so as a private folder within your home directory. But as it becomes easier and easier to just boot up a random VM in the cloud somewhere on a hosting provider,

there's probably a lot of users who have no idea about that, don't think to check, and are surprised when they find out that everyone can read their private journal entries. Verituna, this sort of feels like it's a legacy of old Unix systems. Yeah, I mean, basically, if you think back to how Unix first started off

with the whole wheel group, the users group, this was basically a leftover from it. I mean, the idea that users would be able to share the same space and share the same home folders was kind of neutral. Yeah, and I guess it's fair to, we should point out that Ubuntu's not the only distribution that does this. I just didn't really think about it very much. I don't really think it's going to affect any existing installs, right, Wes? This isn't something they're retroactively

doing. Yeah, it should be pretty safe. It'd basically be, you know, going forward, if you've upgraded your existing user, they won't go and change existing permissions. It'd just be new users you create. And if you do it right now, that gives three development releases and two interim releases before the next LTS. And so, you know, if there do happen to be any kinks, get those out early. Yep. That was the point I was just going to make is you do, you want to figure these things out now. This was the time to do it. And I think it, I think it makes a lot of sense. So there was a story that we've been waiting to talk about. People have been drive by linking me this with a little bit of a twist of a knife in the side too as they laugh at me because there was recently a pretty significant performance regression in kernel 5.10 and

ButterFS. And today, actually I think it was two days ago. Sunday. Yeah, Sunday as we were recording Linux action news, Linux 5.10.8 came out. Yes, it did. And making that very important is it finally addresses the ButterFS performance regression that was found in the 5.10 series. Yeah, I guess it took eight point releases to get that fixed. So is this fair? Because, you know, we talk about, oh, finally. So it was discovered in Christmas, around Christmas, and now it is

the middle of January. It didn't even go one major kernel release, right? And it was during Christmas and New Year's. So is it really like that big of a deal that it took eight minor releases of the 5.10 kernel release to get this fixed? And they're like, how many distributions even shipped this kernel, really? Let alone how many users that, or at any kind of scale, actually deployed this kernel. Probably very few. And so it really was sort of, in the grand sense,

I think actually tidied up pretty quickly. Yeah, no, I think that's a very good point. It probably didn't affect that many users. And of course, you could always, you know, use a previous kernel version if you really needed to. Of course, this thing is likely to happen sometimes when your file system support is built into your kernel. I guess it's just a problem ZFS users don't have. But the good news is that it appears things will get a lot better for ButterFS in 5.11. But were you clear on what

the regression was, Wes? Because I know that one of the ways they tested it was by extracting some files on, like, say, an SSD, like a high-speed SSD. And they would see somewhere between, like, a 5 to 20 times slower performance than expected. Yeah, it sounds like a while ago they made a change that would sort of clean up how they were deallocating stuff that was no longer in use. And reuse some existing infrastructure that they had in place for flushing inodes or during device replace and snapshot and sort of leveraging existing structures within the code base to clean things up. But

unfortunately, that had a downside of stuff taking 5 to 20 times as long to extract because they ended up flushing a lot more than they really needed to. And to make it more complicated, a different bug fix built on this behavior, and so they couldn't just revert the whole change because that had the risk of introducing some other deadlocks in other places.

So the only strategy was to just roll forward, which is also maybe why this took a little bit longer than it might otherwise have. This also was one of those situations where a lot of people piled onto this story and were like, look, it's still not good enough. And still look at this, they just had a performance regression. As if regressions and little issues here aren't things and as if that's not where distribution step in

and kind of provide a bit of guidance before they deploy it. Like just sort of forgetting the entire way software actually gets deployed in the Linux ecosystem. It was sort of leveraged as an opportunity to make fun of ButterFS again. And I struggle to really understand the logic in this. Like we're all in this together. We all want a file system that's capable of things like compression, encryption, snapshotting, send and receive.

These are just basic volume management. These are things that we want in a file system to be a competitive operating system. And we want them built in. And ButterFS is the window to that world. A friend of the show recently had a conversation with an Oracle executive about the possibility of ever sending a ZFS upstream and relicensing it to a GPL. And he was told flat out, never going to happen. And this was within a week.

Sounds about right to me. You could take my word for it or not, but it happened. And it's just ZFS is in a position where they like it right now, and they're not going to relicense it, at least according to this one person. Now, ButterFS is here, it's functional, and it also solves use cases that ZFS doesn't. And I think this is the only other area I just wanted to touch on, not to get on a whole ButterFS thing here, but, you know, ButterFS is really great on devices that ZFS isn't a particularly good option for.

ButterFS is fantastic on the SD card in my Raspberry Pi 4 or on the USB disk that I use. It's fantastic on a laptop with a single SSD. It works so great as the root file system for our Arch server. It saved our Ars several times because it's bought our FS, it always mounts. So our system boots, even if something went wrong with the kernel module for ZFS, which happens. And it also means we get snapshots after every major upgrade or install on our Archbox. Automatically, because we've tied in the right tools. That's all just built in with ButterFS and just connecting the right dots.

And then we can restore the module, we can get things working again, and we can get our ZFS pool back online. But it means our server comes up and we can SSH into it and we can resolve the issue. ButterFS is just undeniably better at that particular use case for us. If the root file system were ZFS, our system wouldn't boot sometimes. And we solve that and still get compression, still get encryption, and still get snapshots with ButterFS because it's built right into our kernel. And it does a great job. And we're not the only ones that are saying this. Yeah, true. I mean, I think just to your earlier point, the fact that this bug was A, found, is a sign that folks out there are really using it. And B, that it was fixed relatively quickly

is a sign of good active maintainership, that people are using it and that there's resources here that people want to continue improving the file system. And at the end of the day, after this patch set in 5.10, Butterfest is significantly faster than it was in plain 5.9. So, you know, okay, there's dips and valleys and, you know,

it goes up and down, but at the end of the day, we just keep getting a better and better net file system. Right, I completely agree and it means that in future iterations, we're just going to be able to build on what's coming in 5.11. And yeah, I mean, you could poke a little fun that it took eight minor releases,

but I don't know, Neil, I mean, you consider the fact that there's holidays in there, it's a pretty quick turnaround. When it comes to stuff like this, it's really hard. And Joseph Bassick and I talked once about this, and he says what I wouldn't give for having a standardized, comprehensive workload, storage workload test, because then we could catch all the things. And it came up, I think, what, five, six days before Christmas?

Everybody was already jumping you know, jumping out because holidays. And even with that, the turnaround time to the initial fix was still less than 20 days, which is freaking impressive. That is way faster than any third party or commercial software vendor has ever done for me in my entire life. And another point here about dips and valleys and whatever with ButterFS, ButterFS is not the only file system that has ever suffered these kinds of things. XFS, for example, in the 5.9 cycle got a major data corruption bug that was introduced midway through and then had to be fixed after the fact. It happened again with 5.10 and had to be fixed again. Like these are things that happen, major or minor issues. Data corruption stuff,

thankfully, doesn't happen very often with ButterFS anymore in terms of like new developments and code churn. But because the kernel's highly integrated and has a lot of subsystems and there's a lot of people and the development workflow for the Linux kernel is, you know, in my opinion, pretty awful. It's impressive how well the Linux kernel continues to be developed given all of these factors. So most people weren't even going to experience this problem because the workload was very, very specific. You had to be unpacking a large tarball with a lot of little files in one basically IO transaction, which is what led to this particular problem. And I have been bitten by ButterFS bugs way in the past, but as a pretty extensive

ButterFS user now, I have it in several, I have it on laptops, I have it on pies, and I have it on server metal. I wasn't bit in any of those scenarios by this, thankfully. So, again, I don't think it was a wide deployment. But speaking about the kernel development process and what that's like, next week on the show, the executive

editor of LWN.net, Jonathan Corbett, is joining me to talk a little bit about kernel development, the state of maintainers, and also LWN because they have an anniversary coming up. So that should be a good conversation in episode 390. So stay tuned for that. But, yeah, I'm glad to see it fixed. I'm glad to see the performance stuff coming.

And, you know, I just kind of feel like we're all on the same team when it comes to some of these file systems. And, you know, I just kind of feel like we're all on the same team when it comes to some of these file systems. And taking a shot at it when it, you know, when it has a bug is, I don't know, something about it. Just something about it didn't sit right with me. And I think it's where I think the issue is, and this is my last thought on it. I'm sorry, but this is just something I've been reflecting on. You got a ButterFS bug in you here. Better than somewhere else, I suppose. I think what you have is you have people that have different use cases

that are looking at a tool thinking, well, this isn't the right tool for me. There's a reason why our large pool runs ZFS, because in our estimation, that was the better tool for that job. And it has remained that way, despite all the different operating systems we've tried, we've always kept ZFS on that pool, and it has served us well. Yep, it's been great. We wouldn't reload.

We wouldn't, like, if we got a new brand-new disk array, we would do ZFS again on that. That's the use case for it. And I think what you have is you have people who are not in the right use cases that are comparing file systems, and it's an apple-to-oranges kind of comparison in that regard. And it's totally fine if it doesn't work for you, right? There's plenty of file systems I never touch by and large, but they are really helpful

for other folks, like exFAT, say, right? And that's perfectly all right. With Linux, we have options and a huge array of options. I love that. Linode.com slash unplugged. Go there to get a $100 60-day credit towards a new account, and you support the show. Linode is our cloud hosting provider, and Wes and I were having a lot of fun last night. And thank you, Linode, for the snapshot system, the way it works.

It really just gives us peace of mind because we wanted to increase the amount of RAM and upgrade our Synapse server at the same time. The software, the RAM, and we wanted to also do the OS. And so it just was so nice because, first of all, one button to make a snapshot right before you start. But what they do that other cloud providers don't is they give you an opportunity to label it. So you label that snapshot, you hit go. And then the great thing is, is there's a notification area in the dashboard, just kind of tucks it up in the corner, you don't have to see it if you don't want, but if you click it, it actually gives you a live progress bar

of how the snapshot's going with time estimation. So Wes is at home. I'm at the studio. I can tell him, all right, we got four minutes until the snapshot's done, and then we'll do the upgrade. And it just works so smooth. And when we rebooted the system

and realized that we didn't know the root password anymore, I was able to get on the console and get that resolved and reset the root password in seconds from their dashboard and get us back in and get us going. And I just really appreciated the tools that Linode made available to us. On top of other things, like we have moved Linodes between data centers now, and that's really cool. Adding additional memory to our Linode was easy. We went from 4 gigs to 16 gigs in seconds. They make all of this really straightforward.

So you don't have to be an expert. But if you are, you start to see a few things that when you drill down in the UI, they're there for you that you can really appreciate. But what I get out of that is you can tell they have a love and a passion for this technology. But really, that's what got them going. That's why they started in 2003

as the first company in cloud computing. They started three years before AWS because they are Linux users and they get it. They get how to build on the technology. That's what I love about a lot of the vendors in our community. And through the years, they've supported events like LinuxFest Northwest, All Things Open,

the Kubuntu project, a bunch of others. Those are the ones that I have seen personally. And of course, now they're making it possible for Jupyter Broadcasting to go independent and give our content away for free. That is a massive value. And I just think it's such a great combination because it's a great service built by Linux users

supporting a Linux business. And you can get a $100 60-day credit when you go to linode.com slash unplug. Go there, support the you can get a $100 60-day credit when you go to linode.com slash unplug. Go there, support the show, get that $100, and play around with their object storage. Just a quick mention, we are going to be talking a little bit about how you can use object storage for backups with our PIC later. And it works with any S3-compatible object storage, which Linode has. And it's so great because you get fast, reliable storage up in the cloud without having to manage a server in front of it.

There's a lot of uses for that, including backups, static websites, and many other things. So get started at linode.com slash unplugged. Before we get into the show much further, I think this is a good moment for us to take stock and do a little housekeeping. show much further. I think this is a good moment for us to take stock and do a little housekeeping. And I want to start with Minimek because he has some updates for us on an upcoming LUBLUG session. Yeah, thanks, Chris. So now that probably the whole world knows that we will have a talk about accessibility tools in Linux next Sunday on LUBLUG, the only thing left to me is to reveal the name of our special guest we have during the talk.

So I have the pleasure to announce that Daniel Ferre, the founder of the Elementary Project, will join us for a Lopluck session. So what can you expect during that talk? So we will have Justin, a user of our JPMobile community, and he will share his experience and workflow with assist his experience and workflow with assistive tools and Linux and Daniel Ferre will give us an inside view on how the elementary team tries to make life easier

for people with disabilities and maybe Daniel might also give us an overview on features we can expect in the future elementary 6 release oh yeah get some info yeah yeah yeah there are some really cool things to mention there and i hope we will have a really good talk yeah so that'll be january 24th on sunday at noon pacific 3 p.m eastern and all they need is mumble right

just join the lobby just join the lobby i hope you will find the time next sunday just join the lobby and if you want we can do some sound checks if you want to participate actively during have next. If you hear the pre-shows, we're always kind of kicking around new show ideas and whatnot. So get subscribed to the All Jupyter Broadcasting shows. One feed, you get Coder Radio, self-hosted Linux Action News, and Linux Unplugged. It's nice, it's tight, it's tidy, and it's a bunch of great content. That's the All Shows feed. You can find that at jupyterbroadcasting.com. And Mr. Bain, I think that's all the housekeeping we have for this week.

It looks like everything around here is tidy. It sure does. Nice job. So I'm going to say Ventoy. You think I got it? Oh, gosh, I didn't even think to question it. That's just what I've been calling it all along. Oh, I question everything I pronounce on the air. You probably should.

All the time. I have a complex about it at this point. So Ventoy is a new bootable USB solution. It's 100% open source. It's simple to use. And you write it to a USB thumb drive. It creates the Ventoy bootable partition, and it creates a storage partition, which you can kind of format as you like.

And then you drop ISOs, just mount it on your desktop, drop ISOs in this directory on this Ventoy USB stick, and they show up as bootable options when you plug it into a PC and boot off of it. So if you have, say, several ISOs up there, a grub-looking menu comes up, and each ISO is listed right there for you to choose, and it just begins to boot that ISO. So it's like one, you no longer, you write this to your USB stick once. And then from that point forward, you just drop ISOs on it.

Yeah, aren't you tired of constantly having to reflash that same USB drive when you already had Fedora on there, but now you're putting Ubuntu and suddenly you want Fedora again. It's annoying. I actually can't tell you how useful this is for a guy like me who often like takes laptops home for a weekend project for the show. And if I'm not careful, I get in a situation where I need an ISO and it takes forever to download an ISO on my connection. And I will write, right now Wes, I have two different USB thumb drives in my bag.

One is an Arch ISO and one is an Ubuntu ISO or maybe Fedora. I can't remember, like I overwrite them from time to time. And so then I get home and it's like, oh, I got the wrong one. And now I'm just going to wait for two hours while I download an ISO. This is so much simpler because I can just put all of them here at the studio on one thumb drive and just take that thumb drive with me. What's also really neat is it's not just ISOs. Like maybe you boot into Windows sometimes or you want to have something to help around with family or friends. We need to install Windows perhaps to re-kick things. If they've got plugins to boot from VHD files or even .wim images, that's awesome. Yeah, and we've kind of toyed around

with other things like this before, but this is just straight up the simplest. If you're on Windows, they've got like an EXE process to go through to write the stick. But on Linux, you download a bash script. I mean, there's several ways to do it, but the simple way is you download their bash script,

give it a little quick look, and then you run it and just point it at the thumb drive device. Make sure you get that right. Yes, don't do dev SDA. For me, it was dev SDF, because I just have a few drives.

You just kick it off. You just execute the shell script, point and give it the device path, and it's in a few seconds after it asks you, are you really sure you want to wipe this thing? It's done. And it's good to go. You can tell that this has been a pretty battle-tested tool, because they have a lot of nice options. So by default, it'll set things

up with an MBR still, you know, MBR style, so that it supports both legacy BIOS and UEFI right on the same thing. But if you want to just go full future forward, you can use GPT and UEFI only if you want. That's what I did, because I don't have any more legacy biosystems these days, thankfully.

It's nice and consistent. And in the same way, they set up one sort of grub system partition where everything, all the stuff that Ventoy uses internally is. And then you've also got a partition just for all of your image files. And by default, that's exFAT, which I think is designed to be

pretty universally accessible. So if you're trying to boot it on some random Mac or Windows PC, it'll just work, and you have access to actually update and add new images without having to worry about some weird Linux file system on there. But on the system I was using, I had an older kernel. I hadn't set up the exFAT support. So I just reflashed that to EXT4, and you can do that no problem. They've got that in the docs.

So if you're just like me and using it on Linux systems, that might be the route you try. Yeah, because then you get larger ISO support, and this thing supports big ISO files. In fact, I think they said they tested up to 600, more than 600 ISOs on this thing at once. Wow. It must be quite the menu to scroll through. Because it's all really kind of a grub menu. I wonder, though, if you saw, so when you boot it up, you know, like a typical boot screen, you got your options along the bottom and some of

them are kind of nice. Like you can kind of preset some video mode stuff and actually there's some really nice things that it lets you do. But did you see what F1 does? No, I don't think I tried F1. Okay, so this is right up your alley because what F1 does is it copies the contents of ISO into RAM. So then you're running the live environment out of RAM. Instead of using the USB ISO image, it just copies all of the contents into a RAM disk and then executes it

from scratch. Like, you get the full boot menu, you get grub where you can select, do you want the open source only boot, or do you want the one with proprietary drivers? Like, if your distro asks you that, you get all of that still. But just all of the contents of that ISO image are coming from a RAM disk. Wow, yeah, it's so cool. And there's all kinds of other plugins if you want to customize things

like add your own theme, customize how the menu system works. It's just really well developed. Yeah, I'll have to just keep this now as one of my tools. I used to have forever ago a Zantech, I actually still have it, a Zantech USB 3.0 hard disk with USB-A plug. And inside I put it like a 128 gigabyte, two and a half inch SSD. And what it does is it uses a CD-ROM emulation, the USB CD-ROM emulation. So when I plug it into a device, it actually shows up as a CD-ROM, and it uses the ISOs that are on that 128-gigabyte SSD as the CD-ROM contents.

And that's pretty neat, but it's got this tiny little janky 80s-style LCD screen that I have to scroll through all the ISOs and select. And it's this weird process where you plug it in, and you need power from the bus to select the ISO, but you want to select the ISO before the system fully posts. So you either race into it really fast or you like power it up, select the ISO and then do a soft reboot. Like you have to do this whole weird dance with it, where this is just, you tell the

system to boot off the USB stick, choose your ISO. And then it's like you, it's like you, and then it's like you had a USB stick that only had that distribution's image on it. It's just great. It's a lot of fun and makes it so much easier to distro hop or just have the different distros I might need easily available so I don't have to re-download them like an animal. It is interesting that, you know, some folks in IRC are commenting

they've had problems on older hardware or even some reports here from Rolf on Skylake laptops. So far in my testing, and I tried it on pretty much every system I've got in my house, no problems, but you do be aware that maybe it won't work everywhere. But I like the idea of just having this as default, and okay, I've got extra USB drives

if for some reason it just doesn't work on some machine. Yep, and like Wes said, it does have UEFI support, but it also has ARM64 UEFI support. So if you've got an ARM64... Fancy! I know! I know, Wes!

I know that! And it has a plug-in framework, which it did not play around with. So I guess in theory, you could really kind of customize this thing quite a bit if maybe you're an IT shop. Yeah, you could imagine

having a sort of standard Ventoy setup. You just give to all your techs. They've got it in their bag. They can boot up to whatever, you know an IT shop. Yeah, you could imagine having a sort of standard Ventoy setup. You just give to all your techs. They've got it in their bag. They can boot up to whatever emergency CDs. There's even a Vento live ISO they've got if you don't have it already set up and just need an environment to get it installed.

And the thing that's nice, too, is they do make it kind of straightforward for Windows users to get going with this. That's not something we talk a lot about, but it does matter. So much easier, right? Especially if you're just flirting around playing with Linux. I mean, here I am using ddrescue in the shell like an animal,

but that's just not very accessible. Whereas dragging and dropping an ISO file and then rebooting, that is. Yeah, and I wonder, like, what are the must-have ISOs you'd put on something like this? Like, System Rescue CD seems like a pretty solid ISO to always have on there regardless just in case all of a sudden you need it. Yep, definitely. Or Clonezilla,

similar things. Yeah, yeah. Veritunus is Clonezilla. That's a good one. Yep, that is good. Good point. I think too, like this might actually

push me over the edge to finally get a USB-C thumb drive because I'm kind of getting sick and tired of using adapters and whatnot. Mm-hmm, yes. And maybe I'll get like a kind of a higher and tired of using adapters and whatnot. Mm-hmm. Yes.

And maybe I'll get like a kind of a higher speed one if I can. And 128 gigs would probably be plenty. I never even, when I had my Zalman whatever it was thing, I never exceeded the 128 gigs. Because by the time you're using that much space, the Linux ISOs on there are so old, they're not really relevant anymore. You just kind of cycle them out. So you don't need a ton of space

unless you're a maniac. Tails. Pycrash points out Tails. Cassidy in the chat room says D-Ban. Yeah. Yes. Yes. Yes. Yes. I once had a situation where I couldn't get to D-Ban quick enough. So it was kind of like a

there's like, I was working at this place that was called Dream Dinners. And they specialized in this online ordering process, which was brand new back then. And they specialized in this online ordering process where you would put together your dinners and then you'd go into one of their affiliate kitchens, which were kind of just all over the place, all over the country. And you would go in and you would actually assemble everything into bags and stuff and then take it home and freeze it. Oh. It was kind of nice because they did all the thinking for you and they did all the ingredient selection, but then you could add, like, if you like a little bit extra garlic, you'd add a little bit of extra garlic. But they've got it all chopped up,

prepared or pre-cooked or whatever needed to happen for those ingredients. Yeah. And the instructions you put in there, and of course they've got a professional chef or two in there that answer your questions, and everything's chilled and clean and all that. But of course, what people really wanted was they wanted meals delivered all the way to them, and they were just like one step removed from that. You had to go to their franchises.

But to power all of this, it's kind of funny, this is our LDAP episode, because to power all of this was a Red Hat Enterprise Linux setup with LDAP that powered the user logins, that powered the desktop logins through Samba, that powered the website backend to manage it all. It was all done with a single sign-on system that used LDAP. And we would kind of monitor this system, and I would kind of run it, and I was the server guy,

and I had an IT help desk guy, and it was basically me and him and a developer, and then eventually two developers, and that was the entire team. And this infrastructure ran really great. I was very proud of it, because it was, I guess, what we would now call hybrid, but we didn't call hybrid back then,

but it was a good chunk of on-premises infrastructure, but then supplemented with cloud infrastructure for our front-end web server and database and stuff. And, you know, that was just how I ran it. And this efficiency group was hired to come in and make recommendations. And they came in and they analyzed everything and they said, first thing you got to do is you got to switch to ASP. You got to drop this PHP stuff. This is

some open source garbage that isn't going anywhere. You got to get rid of this. You need ASP, Active Server Pages, and you need to run it on IIS. This back then was probably Server 2003. And you need to switch to Exchange and Active Directory. And you need to fire your IT director,

which wasn't me. But they were a friend of the owner, so instead of firing her, they just transitioned her. And they brought in this new guy who was recommended by the company and had worked for Boeing. And he had just kind of started decimating

the entire infrastructure, and he and I never really quite got along. You know, it was one of those things where I'd get in like at 6.30, 7 a.m. to like make sure everything was working when the business started. And then I'd maybe head out at 4 o'clock. And I was passing him in the stairs. And he looks over to me and goes, you know, I can tell who's dedicated by the cars that are in the parking lot at the end of the day.

Your car usually isn't there. I'll remember that. And that's like one of the first ways I met him. And you know, like screw him, right? Because I was actually getting up, commuting down there and from Smoky Point, you know, so, you know, it was a good drive, getting down there and I was making sure all the systems worked. And my IT guy, Josh, he and I worked together. He was there to make sure everybody got a smooth start in the office. So the office was there to support the rest of the infrastructure. Like we really went the extra step. And so when he wanted to remove all the Linux,

I was just kind of done at that point. That's when I actually decided I was no longer working in corporate America, and I was just going to do independent IT contracting. That's actually what kicked off my IT contracting career. So this was about a dozen years ago. As one of the hiring conditions, when I started that job, I wanted a System76 laptop. And so I used the hell out of that laptop. I ran that entire infrastructure from that System76 laptop, which made me one of their very early customers too. And when I was kind of coming to the realization I was leaving, I went in, talked to the boss about it. I said, you know, I think I'm going to move on, but I want to give you guys a good transition. He's like, oh, okay, well, all right. Well, why don't you go to lunch,

think about it and come back and we'll talk. All right. Well, there's fish and chip place next door. So I was happy to go get lunch. So I'm eating my fish and chips like an idiot when I should have been downloading D-band. But instead I get back from my lunch, I sit down on my laptop and I literally think to myself, you know, I wonder if he's just going to can my ass. I better get ready just in case. And I'm not joking. I launched my web browser when he comes up to my desk and says, I'd like you to come into my office. He fires me on the spot and then doesn't let me ever go back to my desk. He goes and boxes my stuff up and takes my laptop, which was logged in, had my web browser up. I don't know if the screen had locked or not because I didn't see it.

And that was the moment where I was like, I regret not having D-Ban in my bag ready to go because I could have popped D-Ban in and walked away and gone into his office and just let it wipe. Because if you're not familiar, D-Ban audience is the tool to do like an intelligence agency grade wipe of your disk. And I did have, I was talking, I was not happy with him. So I had emails on there and stuff like that that were critical of him.

And I didn't want him having that and keeping that. And some of them were private conversations too in Pigeon. They weren't like stuff I wanted him to see. Yeah, so it was a real shame that I couldn't wipe it. And so, you know, having a tool like Ventoy is kind of powerful because it means that I could always have a thumb drive on me that has D-band on it. Just right in literally your back pocket. Always ready to go.

And so, I mean, these tools like this are kind of important for that regard. And you can boot and nuke, yeah, good old boot and nuke. But man, that guy was such a turd, you know. And then they went out of business, not shortly after that. Oh, I wonder why. Yeah, but it was something, you know? It was one of my better infrastructures. I wouldn't say it was my best, because I think I went on to build my best when I was a contractor.

But it was up there. I was really proud of it. And, you know, that single sign-on stuff through LDAP and Samba worked really good. Windows desktops, Linux machines, website logins, all of it. It was great. It's so nice when things click like that and it just works. It's reliable. You feel like the system's actually doing what you need

and you're serving the business, and that's the whole goal. Now I have literally no idea how I would set any of that up. It's funny, right, that I used to build that and support it, but today I'd probably start picking it back up, but I couldn't tell you right now how I would set a system up like that. Oh, it's fine. It's just a couple of clicks in the cloud.

Oh, dear. Ohdeer.app. Complete website monitoring from multiple locations with broken link and mixed content detection and advanced SSL certificate reporting and a lot more. Visit ohdeer.app and use promo code Linux for a $10 discount on any plan. But you can sign up with no credit card required.

So it's when you go to subscribe, use the promo code Linux. This is really a tool I could have used a lot back in the day when I was running infrastructure because I want to be the first to know when something isn't available. Not the customer or not even my customer at the time, which would, of course, be the end users. Odeer has global uptime checking with servers worldwide that report on problems as soon as they happen. But the nice thing is you can go deep into the site, too.

They can crawl and index your entire website. They can detect broken links, which, of course, can hurt your SEO, but also just don't look good. And they can probably notify you. They have a really intelligent scheme around that. You can also monitor things like your cron jobs or your Windows schedule tasks to make sure that they actually triggered and fired off like they should have, because that also can lead to other problems. And Odeer is always monitoring for performance. You get the

idea of the speed of your website right now and over time. So you can kind of see if there's a trend happening. But what makes Odear very special is their API. It lets you configure everything about the application. Everything you see in your dashboard can be controlled with their easy-to-use RESTful API. It's got kick-ass documentation. And as a bonus, all the changes you make via the API will be visible in real time on your dashboard. Odeer is monitoring,

and it's a solution that embraces automation. So if you're looking at monitoring solutions right now and you're thinking, I'd really like something that we can integrate, that we can automate, that we can deploy as code, you need to check out odeer.app. It's Odeer. It's a comprehensive API. It means third-party integrations are simple. It's ODIR. It's a comprehensive API, means third-party

integrations are simple. It's really easy to get going with command line clients, telegram chatbots, SDKs, Terraform providers, and a lot more. So right now, head to odir.app and start a 10-day no-strings-attached trial. No credit card required.

You can get set in less than minutes. And when you do sign up for any plan, use the promo code Linux for a $10 discount. And if they ask, tell them Linux Unplugged sent you. Ohdear.app. Check them out and use promo code Linux when you do sign up. We got a nice batch of feedback. How about I take the first one and then you take the next one and we just kind of do it like that.

Sounds good to me. Martin writes in about OpenSUSE and FOSS Adventures. He says, Dear Chris and Wes, I'd like to give you some feedback on your OpenSUSE feedback. I'm a longtime JB fan since 2010. Wow. That has been a while. Oh, gee. That's great.

Yeah. And he says, he uses OpenSUSE. I remember when you and Brian were quite positive of OpenSUSE. Yeah, but I'm not, was Brian just positive because it got a job later? I'm not, I don't know. But he says, when you guys talked about Tumbleweed, it was quite negative. But Chris, those are your experiences. And I think you should be able to voice that.

But I've run OpenSUSE for 10 years, and I love YAST. I appreciate how YAST works. It enables my mom to install her HP printer without my help, who is a total Linux novice. It's hard to argue with that, right? That's one less call for tech support. But what I don't get from that is, is YAST great for him anymore? And is that because he's only used YAST, so he has to keep using YAS?

So, Martin, I wouldn't mind, like, a follow-up on that. But he says, your comments about OpenSUSE documentation, about the voice and tone and the community and about the NVIDIA proprietary drivers, they were 100% on the mark. I experienced those pains, too. He goes on to say, that's why I have my blog, FOSS Adventures, to help improve the communication

around OpenSUSE. I try to show others to help improve the communication around OpenSUSE. I try to show others how to solve common problems in OpenSUSE. So that's great. I also love the blog of Cubicle Nate and another open source user. Check out cubiclenate.com. So his website is Foss Adventures, and then Cubicle Nate has a website at cubiclenate.com. He says, keep up the great work. regards martin super solid feedback martin and that's a great way to i mean that's a great example of how to communicate hey i heard i heard what you said about that distro i heard that it doesn't work for you well that's fine here's why it works for me and it's just a real nice like i got real value i know i know exactly

why he asked works for his mom that totally totally makes sense. And then he acknowledges that the documentation and the process around the NVIDIA driver is horrible. And I think that's such a great way to engage with somebody who tried to distro but maybe had a slightly negative feedback, some kind of negative experience in that trying process. And honestly, if this was the experience that I had primarily received, I don't think we would have even mentioned the tone and the feedback. I don't think that would have even come up. Right. But here we're not arguing. We're just sharing experiences and we can all learn from that. Now, I like this next one because it's about Neon and Plasma 5.21 looks like a barn burner of a good release.

Excited already. We just kind of did a rundown in Linux Action News, if you're not familiar with what's coming in 5.21. But yeah, so Nacho writes in. Wes, take it. Hey, Chris and team. I'm just writing to describe my recent experience with KDE Plasma through KDE Neon.

I heard your predictions episode and all the great things KDE is bringing and thought I'd give it a try. I had a 2018 15-inch HP Envy laptop that never fully worked when I installed Ubuntu proper, Ubuntu Mate, Zubuntu, and lastly Deepin. Well, yeah, I haven't really tried those things. It always had some glitches that I would have to put up with when running off of one of those other distributions, but KDE Neon worked great, and I was surprised at how beautiful it all looked stock.

It was a bit rough trying to install the distro in high DPI, but thankfully I'm not new to distro hopping. But once installed, I was pleasantly surprised and can see why so many people champion Plasma as their desktop of choice. Also, shout out to Alan's ThinkPad Corner on Ubuntu Podcast. You can pry it

from my cold, dead hands. I'm right there with you. Thanks again. Long time listener, Nacho. So, Nacho, I'm curious if you had, you mentioned that you tried a few distros there, but what you didn't mention was Kubuntu. And the reason

why I ask is because you say Ubuntu gave you issues you mentioned that Mate, Zubuntu and Deepin all had problems but I'm surprised that Neon works for you because Neon's based on Ubuntu right? It's just Ubuntu with the Plasma

desktop and fancier newer Plasma yeah it's 2004 with newer Plasma so I wonder if Kubuntu, not that you, now that you got something working, it doesn't really matter, but just for troubleshooting,

you know, I'm just curious to know what was going on there. But good to know, and I agree, KDE Neon's going to look even more attractive here in about mid-February when Plasma 521 lands. It's going to, I'm going to have it, I'm going to be updating the day it comes out.

So Matthew writes in. He says, my request is simple. I really miss Chris's next Tuesday in the outro. It always makes me smile and look forward to the next episode. Can we please bring that back? Smiles are in short supply this day. I will try. I kind of stopped doing it because I just figure after a while,

a gimmick like that gets annoying if people listen and listen. But if the people want more Tuesday, Wes. You have no right to deny these folks. Yeah, I'll tell them to see you next Tuesday all day long if that's what they want. Do you want to take listener Chris's email? Yeah, I better because otherwise, if you read it, it's just too confusing based on the name. Hello, Chris, Wes, and the JB family. I wanted to share an upcoming product release from Lenovo, and Chris links us over to the ThinkPad X12 detachable 12-inch Intel-powered tablet. Interesting.

Continues on, what do you think of the prospects of Linux running on Lenovo's answer to the Microsoft Surface Pro? I've been desirous for years of an all-in-one Linux device that serves as both my laptop and tablet when needed. And given that I occasionally run a few VMs and containers that require x86 hardware, something in the ARM world just wouldn't work for me. Thanks for the awesome shows and all that you do. Life wouldn't be the same without my JB Linux fix. And I'm curious to hear if you guys would be interested in something like this.

Hmm, so this guy is a 12.3-inch detachable tablet-style laptop. It does come with a little pen that kind of snaps onto the side of the keyboard. And it looks pretty lightweight. It looks like it's probably significantly lighter than any laptop I've ever had. And it has also a kick-out stand. So if you do take the screen mobile, it does kind of have that surface-style stand that snaps out. Oh, it looks like it might even have Intel Xe graphics. So that's interesting. Up to 16 gigs of RAM, you can get an i7 in there or a 1 terabyte NVMe SSD. So it's not a slouch of a system.

Yeah, that makes sense, right? It makes sense that it would have 11th gen. MVME SSD, so it's not a slouch of a system. Yeah, that makes sense, right? It makes sense that it would have 11th gen. That's nice. It's a 1920 by 1280 IPS anti-reflective, anti-smudge, 400 nits Gorilla Glass screen. Looks like it just has 16 gigs of RAM and a terabyte MVME. It may not even be customizable in that regard.

You know, it's funny. If you would have asked me six months ago, I'd be like, nah, nah. But there might be a place for this in my life. I am starting to realize that, and I apologize that I'm making this somewhat kind of a car analogy, but there's a lot of different laptops for different people. And you got your commuter laptops. You got your truck laptops. You got your sports car laptops.

Is this just you trying to explain why you have so many computers? I'm buried in laptops at the moment. Thankfully, only a couple of them are actually owned by me. But I'll tell you what, having a bunch of different laptops at one time, it's like never something I would have just done, right? I wouldn't just all of a sudden like have a bunch of laptops because what's the point? You don't need that many laptops at one time. But because there have been a bunch in and out for review, I've got ThinkPads, I've got Lenovo's, I've got Asus's, there's a MacBook, like the whole gamut for, because we're doing them on Coder 2. It's like the whole thing, right? And that's really been an interesting perspective

because it's like, well, which one do I want to drive right now? I think I'm going to take this one because I'm doing this kind of job. And I could see if I was like rich, like Jay Leno, instead of having a garage full of cars, I'd have garages full of different kinds of computers and laptops and stuff. And I could see a place for this because I really like having a tablet when I am sitting on the couch next to a family member and we're looking something up or I'm entertaining myself while the kids are watching some video on the TV. You know, like, I like having the flexibility of that. But they're so limited.

Anytime I want to do anything, like if I need to go log into my bank account and figure something out, like anytime I want to do any serious work, I put the tablet down and I go dig out the laptop. But if this, it'd just be a matter of snapping it into a base. Obviously, there'd be UI and desktop environment issues to work out, which I guess in the real, I guess, so I guess, Wes,

you'd have to look at this in the real world. That may be a disincentive. It's getting better. But it's not like there's a perfect UI for this right now. That was kind of my main question is, what would a real Linux experience be with a modern Linux desktop? Like, I'm not sure we've optimized that well for this case. I don't know.

I don't have any device that's really like this. So I've not tried it personally. So at this point, I think picking one up would probably be on the experimental end. But that's the first step to figuring something out that works nicely. And if, you know, Lenovo seems more and more friendly to Linux, maybe it means the hardware will just work and we need to just work on the presentation layer. If Lenovo was smart, I mean, who am I to say?

But it seems like they would be kind of interested in what it's like to run Linux on there because otherwise they're competing against Microsoft's own hardware, which, hmm, you know, like the Surface is pretty well known and pretty popular, and if you're kind of into this form factor, you're probably pretty inclined to go get the Surface. And there's been news recently that there are several performance improvements and other kinds of improvements for the Surface landing upstream in the Linux kernel right now. It seems like it's a popular product. Yeah. Oh, hey. Okay. So we might have more than

a few options pretty soon. Yeah, so we'll see. All right, so I'll take this last one because it's totally on me. Hang writes in, he says, I noticed, I noticed that Google Forms were being used for the tuxes. Therefore, I did not fill it out. I was astonished to see that members of an open source community use proprietary and spyware tools such as Google Forms to conduct surveys. Have you considered using an open source survey software

rather than closed source proprietary ones by a data mining company? What do you think of Cosmo Communicator, Gemini PDA, and Astro Slide Projects? There you go. He says, what do you think? Well, he's right, Heng. I'm sorry that we used

Google forms. It was, uh, totally a proof of concept. Here's the idea. Oh my gosh, this is a cool idea. How do we implement it right now? Because we want to talk about it and get it out there. And how do we implement it in a way where once we've all of a sudden decided to create this extra work for the team, we can consume the information that the audience in several hundred, you know, in the form of several hundred submissions will be palatable and processable by us and that's where when it was like kind of a day before final go no-go decision we were like oh right well we know that if we use google forms we can get it in a spreadsheet that we can share amongst each other, and then we can massage the data, which is very much what we did. And that was really what drove the decision

to use Google Forms. I don't love it. And I think what I want to do is just kind of be a little more intentional as the end of the year begins to approach to just start working on this. We've got emails from a lot of folks that want to help us. And I think that's probably when we're going to rally the troops. Yes, which we're very grateful for, by the way. Super grateful. It's really nice because I want to do it right. And now that I know it's worth the time. Yeah, maybe we called this last year Tuxy's Beta. Yeah, because, you know, I thought it would be just something we'd do it. We would enjoy it. It'd be a way for us to spread a little joy and then it'd be done. But we got contacted by several projects. We had lots of people write in and say, oh, I didn't get to vote.

We had people that wanted to kind of have a robust debate about which winner we had chosen. Like, there was a lot of people that kind of got engaged. And so that's great. That showed us that it was worth the effort. I actually thought it was going to be a flop. It was a misread on my part. And so the Google Forms was a way for us to try it without standing up something, without really having to go in deep on something. But my goals are now to have a dedicated website in the future to let you guys know about a week

earlier when the submissions are open and to not use Google Forms. We'll see if we check all those boxes. All right. Well, if you want to get an email in, go over to linuxunplugged.com slash contact. Wes, why don't we wrap it up with the System76 prediction before we get out of here? Kevin writes in with his System76 prediction. My 2020 prediction, all right, it's a little late, but we're going to allow this one, is that System76 will unveil an in-house designed laptop in 2021. Yeah, he's saying that that's what's going to happen in 2021. I thought this was a fun one since Social Happiness, aka Emma, is in the chat room right now. I thought it'd be fun. I'd love to see it. I think you don't get

to laptop without getting to keyboard, Kevin. So I think your prediction might be a little ambitious because I think what we first need to see is a keyboard that ships. And that is such an opportunity to learn so many lessons about the manufacturing process, but also you want a laptop to have a killer keyboard too, right? So you learn about making that component as well. Plus, you know, System76 has a high bar for quality. So they're not going to push something out

until it's ready, right? So that's just why maybe the timeline's going to be a little longer than we hope, but I would love to see it just whenever it's possible. Right, and also, I would buy the hell out of a System76 keyboard, and I think a lot of people would, because you know they'd do it right, and then that helps fund the laptop. I mean, I'm not going on any information, but it seems like a pretty practical and logical approach that's likely got a shot. So that's, I suspect that's it, Kevin.

So maybe 2022, I don't know. I mean, you never know. Maybe, maybe. But I like it. I like thinking that way because what a moment that would be when someone in our community starts building hardware like that, that, you know, it's like a laptop option. I mean, who knows what that could be. And then you'll have one more laptop to buy, it's like a laptop option. I mean, who knows what that could be? And then you'll have one more laptop to buy, Chris. Just what you needed.

Yeah. All right. Just a quick pick. Really, just a quick pick because we got to get out of here. We don't want it to be a long one, but it's starting to get that way. Wes found a really handy tool to back up your Postgres databases into object storage. Tell us about it, Mr. Payne. It's called PGHord, and yeah, that's basically it. It's a Postgres backup daemon that's focused just on Postgres and has a special support for storing stuff in cloud object stores. I like this because the database is kind of special

for a lot of web applications. It's where all of your important data, customer information, everything that you care. For us, it's often show notes or other kinds of stuff that we just don't want to lose. And it's sure you can, you know, we run everything in Docker and it's pretty easy to back things up that way. But I like the rich support here. They've got automatic periodic base backups. You can integrate with the transaction log. There's optional standalone hot backup support. And of course, it just goes and sticks it right in object storage,

which works really nice for us since we're already leveraging that. There's encryption support and compression, and you can restore directly from a compressed and encrypted file in object storage. You don't have to download it and do any fancy workflows. PGHord just takes care of it. And I know like half of the things we run in containers, there's a Postgres running somewhere. I'm sure that's true for a lot of other folks. So it seemed negligent that we constantly advertise fun software to run without also giving you a tool to help keep it safe.

Absolutely. And this is a good one. And we've, you know, we've gone on and on about object storage recently, but we're just finding other useful ways to take advantage of a system like that. And it's really simple to just host it even on your own box. There's open source projects that make your system essentially look like AWS S3. And you can just store to a system on your network using this, or you could store to systems in the cloud. And because you have things like Linode and Google Cloud, and obviously AWS and Azure and lots of other services that support object storage, DigitalOcean. There's a lot you can choose.

So you can kind of shop. You have a lot of choices there. So it's pghord, and we will have a link to that in the show notes. Special thank you to our members, our core contributors at unplugged.core. You help keep this show independent. Let us be choosy. You give us the leverage to walk away from a deal. You know, if we have an ad deal that just doesn't seem like it's the right fit,

we don't have to be desperate. We have our members, and they give us multiple ways to remain independent, and we're really, really appreciative of that. So we give a couple of perks if you become a member at unpluggedcore.com. You get a limited ad feed, fully produced, sounds great, ready to go in the car, just less ads, a little bit shorter runtime. There's a second feed that's the exact opposite of all that. It's the bootleg feed, the full live stream, every mess up, every minute we're on air. It's super long.

It's like double the show. And there's often a lot of really good conversations that just don't get captured when you hit record because that's human nature. And that is in that feed. So you get two choices. And you can support the show at the same time. UnpluggedCore.com. I would love to have you join us live. We do this show Tuesdays, and we do it at noon Pacific, 3 p.m. Eastern.

See you next week. Same bad time, same bad station. And, of course, that's at J Pacific, 3 p.m. Eastern. See you next week. Same bat time, same bat station. And of course, that's at jblive.tv. You can get Mumble, then participate in the LUP plug and participate in our live lug. And you get a real-time stream. And it's free software. It's a free software stack from top to bottom when you go that route. That's a win-win-win.

Yeah. Links to the stuff we talked about today at linuxunplugged.com slash 389. Our contact page is there as well as all our RSS feeds and all of that good stuff, including Matrix and Mumble server info, et cetera, et cetera, et cetera. But I'll let you figure all that out. I say it enough. I don't need to tell you again. Thanks for joining us.

And we'll see you right back here next Tuesday. jbtitles.com Let's go vote. See what we want to call this thing. We need your votes. So it seemed like we stirred up quite a bit of feedback after the Poe show last week when I mentioned my Fedora woes. I heard from folks on Manjaro and other distributions that we're having the same issue where flat packs

are just non-existent to their Plasma desktop. Not in the launcher, you can't add them to the menu. It just doesn't seem, and maybe we found a common thing amongst all of us. We did a little bug-a-thon off-air, I guess, in a way. Async bug-a-thon over in Matrix, look at that. And it seemed like, I don't know who, maybe Neil or Carl, I don't know which one of you would be best to speak to this,

but it seemed like the common thread was everybody had swapped out their default shell in console, which I don't know how you connect those dots. But I guess whatever you set as your default shell actually is respected by the rest of the Plasma desktop session. And for whatever reason, that broke things. And when you just switch it back to the default, I guess it unbreaks.

I haven't done it yet. I was thinking maybe I'd try it now live on the show. I think some of that too was, you know, some folks had changed it with like changing it system-wide in the shell. I'm not sure how what you normally do when you install Phish versus say you can also configure just console to launch Phish as your shell. I think that was maybe the workaround of like don't change it system-wide, but just if you want to have console or whatever terminal emulator you're using,

launch it for you. Oh, I don't think I do set it system-wide. So maybe that's not it. Let me see. I got a lot of user accounts on this one user laptop here. I did set it in password. I did change it in password. I did.

So you're telling me that's where I have to change it. It's worth a shot anyway. Switch that back to Bash and then... I don't even remember doing that. I must have just done it right after install and just got it over with. You were so sick of not having Phish set up that you just did whatever it took. I want my auto-complete.

Where's... Once you're spoiled, I mean, it's hard to go back. Let me just go edit Etsy password with Nano real quick. A little pro tip for Nano. TAC-W for really wide files, which Etsy password is absolutely not. But you feel like a pro when you do TAC-W. I thought you had switched to using micro.

I did, but with the new release of Nano, I'm back on the Nano sauce. Someday we'll get him a decent text editor. It's when there'll be a terminal-only version of VS Code. That's what it'll be.