0:01

Well, we're just thrilled to congratulate

0:04

and welcome a new family to

0:06

the Fedora project. It is

0:08

the Fedora Atomic desktops. I

0:10

think this is actually a great move. A lot of times we'll

0:12

talk about silver blue or silver blue-like

0:15

spins from Fedora. There

0:17

hasn't been kind of a catch-all term for these. And

0:20

immutability isn't quite the right

0:22

phrase either. They write

0:24

on the fedora-magazine.org. Some may note that this

0:26

is more of a reintroduction. Project

0:29

Atomic started 10 years ago with the development

0:31

of Atomic Host. As the

0:33

team stated back then, quote, the Atomic Host comprises

0:35

of a set of packages for an operating system

0:37

pulled together with RPM OS tree to

0:39

create a file system tree that can be deployed, updated

0:42

as an Atomic unit. So

0:44

it's more really about how it gets updated and

0:46

those updates get implemented than the immutability

0:49

of the system. And

0:51

so you have a class of distributions that

0:53

kind of fall under that. And so they

0:55

wanted to adjust their branding so

0:58

that when more spins come along, it all makes

1:00

more sense. Like maybe you might have a cosmic

1:02

spin, right? So you'll

1:04

have, for example, instead

1:06

of Fedora-Sorisa or whatever it

1:09

was, it's going to be Fedora-Sway

1:11

Atomic. Or instead of Fedora-Onyx, it's

1:13

going to be Fedora-Budgie Atomic. Yeah.

1:15

Okay. Now that I've seen it, it does

1:17

feel like we probably did need a name for this kind of thing. I

1:20

mean, already you kind of wouldn't know that silver blue

1:22

and Keno white were necessarily trying to accomplish similar things.

1:24

Right. Now they are going to

1:27

keep the silver blue and Keno-night branding just because

1:29

people are familiar with those. Fair

1:32

enough. Yeah. But the other projects,

1:34

Sway and Budgie agreed since they were early

1:36

enough that they should probably just adopt this

1:38

new branding. And new spins, like

1:41

seriously, if we see a cosmic

1:43

version, it'll be Fedora-Cosmic Atomic. And

1:46

then you'll know that you can grab that

1:48

and have Atomic updates to your system. It

1:51

also has the nice side benefit that it feels like

1:53

you're playing Fallout. Hello

2:06

friends and welcome back to your weekly

2:08

Linux talk show. My name is Chris.

2:11

My name is Wes. My name is

2:13

Brent. And my name is Alex. Hello

2:15

gentlemen. Yep, we've got a special crew

2:18

this week to dig into a brand

2:20

new episode idea we've come up with,

2:22

Will It Mix? The idea being take

2:25

something that might be complicated traditionally to

2:27

deploy or at least complex and

2:29

see if we can't make it reproducible and

2:32

simple with Nix or at least

2:35

see if the Nix way might be better or not.

2:37

We'll compare and contrast the two, then we'll have some

2:39

great boosts and picks and more. So

2:41

before we go any further, let's welcome our

2:44

Mumble Room time appropriate greetings virtual lug. Hello.

2:48

Hello, thanks for joining us. I hope you brought your

2:50

next class. Yeah,

2:52

yeah, I hope you brought your next class.

2:54

So this episode is brought to you by

2:56

Tailscale. So go say good morning to Tailscale

2:58

by going to tailscale.com/ Linux

3:01

unplugged. You know, I

3:03

am done with putting my services on the public

3:05

internet. Seriously, like when we deploy stuff

3:07

now, we're just deploying it on our tail nets. Not

3:10

only does it let me experiment with more types

3:12

of software that might not be ready to go

3:14

on the public internet, but it just makes a

3:16

lot of security trade offs a lot

3:18

less of a trade off. It is the

3:20

easiest way to connect your devices to each

3:22

other wherever they might be secure remote and

3:25

it is super fast. You can get Tailscale

3:27

going up and running on your devices in

3:29

just minutes and it's protected by a while

3:31

ago. That's right. The noise protocol. It's

3:34

pretty great. Go support the show and

3:36

try it free on 100 devices, but

3:38

going to tailscale.com/Linux unplugged. Okay,

3:41

so before we get into this new will it

3:43

Nick segment, I got some questions

3:45

for the audience. I want to put it out

3:48

there. I have been trying to come

3:50

up with some kind of tiered storage

3:52

solution that ages

3:54

data that you don't use very often

3:57

to a longer term storage and

3:59

maybe with a. threshold where you set six

4:01

months or a year. And then in

4:03

the background, it's sneaky, moves

4:05

that data to something cheap. Maybe it's cloud storage,

4:07

maybe it's another computer on your network. So just

4:09

to make sure I'm understanding, right, you're not looking

4:12

for like a fixed tiered system where you kind

4:14

of like, you know, you already have pools that

4:16

you're managing, you want something that is smartly rearranging

4:18

stuff, making more, more room on the

4:20

local storage when you kind of stop using stuff.

4:22

Yeah, you got it. But with one really big

4:24

caveat, the applications and the

4:26

OS might not even be aware it's happening.

4:29

So say Jellyfin were to scan the

4:31

file system, it would see maybe an

4:33

18 gig MKV file, even

4:36

though in the background, that's

4:38

been sneaky move to some kind of

4:40

slower, cheaper storage. And

4:42

it kind of the really rough

4:44

equivalent would be like the Dropbox smart sync

4:46

feature, where it tricks your file system into thinking that

4:48

the file is there. And then when you go to

4:50

access it, it actually pulls the file down. Oh, one moment.

4:53

Yeah. And then you just wait while it pulls the file

4:55

down. And I think the reason why

4:57

I'm looking for something like this is because

4:59

if you look at media collections and photo collections

5:01

and all these kinds of

5:03

things, like a lot of these things you access infrequently,

5:05

some of it you access frequently, and

5:08

other things very infrequently, and would be nice to

5:10

just age those out over time,

5:12

but then not have to actually manage that

5:14

and not have to split your

5:17

library up into multiple libraries, say in Jellyfin. Right. You want

5:19

one sort of universal

5:21

view of it. Yeah, if

5:23

possible. And our friends over at 45

5:25

Drives, they're 45 Drives have

5:28

something that hasn't been updated since

5:31

December 7th of 2021. But they have

5:34

something called auto tier. And it is

5:36

a path through fuse file system that

5:38

does intelligently move files

5:40

between storage tiers based on frequency

5:43

of use, file age,

5:45

or fullness. Tearfulness.

5:48

Oh, tearfulness. Okay. So when you start filling

5:50

up the disk, it's called

5:52

auto tier. And it

5:55

seemingly would kind of do this

5:57

job, but there has to be

5:59

another way. One

6:01

wonders if perhaps BcashFS could have a solution for

6:03

this. I know that's something you kind of teased

6:05

Wes, but it's not exactly what I'm looking for.

6:07

I think we got at least try autotie, right?

6:10

This looks neat. Maybe. Maybe.

6:12

I'm putting the request out there. People boosting with

6:14

how you solve this problem. We'll test it with

6:17

Brent's data, don't worry. It's

6:19

fine. I

6:22

don't know, because what I'd really love to

6:24

solve is spend

6:26

more on your, like,

6:28

maybe you have eight terabytes of

6:30

solid-state storage. And that's your most

6:32

frequently accessed stuff, and it's super fast. And then you

6:34

have 20 or 40 or whatever it is of slower

6:38

storage that, you know, if you don't play that movie for

6:40

a year, or you don't access that ISO

6:42

or that drone video you created two years ago, it

6:45

just slowly migrates to

6:47

the slower, cheaper storage. It does seem like it'd be

6:50

nice if you also exposed sort of the stats for

6:52

you on when you did actually your files. Like, maybe

6:54

it'd be a nice way to sort of surface stuff

6:56

that either do want to leave, or you maybe would

6:58

just want to get rid of entirely. Yeah, it's true.

7:01

You could look at this and be like, well, okay,

7:03

actually never opened it, bro. Yeah, that's like when you

7:05

move and leave things in the box. Yeah, exactly. What

7:08

would be the trigger? Would it be a

7:10

last modified time? Or like, because when you

7:12

read a file, it doesn't leave a trace

7:15

that you've done that anyway, does it?

7:17

So something would have to be watching every

7:19

process all the time to say, hey,

7:21

I've read this file recently. Yeah, I

7:23

think that's where autotie uses the fuse

7:26

layer so it can like, it can

7:28

track the rig. Oh, I see. Yeah,

7:31

okay, that's interesting. Maybe IO Notifier also would

7:33

do that. Is there an eBPF hack involved?

7:36

Not yet. I'll tell you what

7:38

comes to mind for me is MergerFS. You

7:40

could very easily have different types of storage

7:43

unified under a single mount point.

7:45

Now, what you wouldn't get with

7:47

that is the automatic moving

7:49

of files based on how dusty

7:52

and crusty they've gotten sat in a corner. But,

7:55

you know, I have a ZFS mirror

7:57

that handles all of my important data

7:59

alongside a JBOD array

8:02

of disks that is

8:05

formatted with XFS on each individual drive.

8:08

MergerFS groups those together so I've got a

8:10

single mount point and all I have to

8:12

do is know which directories live on

8:15

which tier. So I mean you could do a prefix

8:17

or whatever you wanted to do to the folder names

8:19

or whatever, however you wanted to organize it. And

8:22

the nice thing about MergerFS is that you could

8:24

add and remove drives, USB drives would count and

8:26

could automatically be picked up and that kind of

8:28

stuff. So in terms of the

8:31

tools that are already available and tested, Merger

8:33

seems to me to be the MVP. That

8:36

crossed my mind and I was gonna ask

8:38

you how would that look say from a

8:40

library standpoint? Because what I'm trying to get

8:42

is like a unified almost jelly-fin leverage do

8:44

you say? Where I wouldn't

8:47

have to separate out like

8:50

folders because you'd still think you'd have to have folders

8:52

for the older slower stuff maybe?

8:55

You can add multiple folders to the

8:57

same library right so you could have

9:00

an old and a new. So I've

9:02

actually done this fairly recently myself with

9:04

a I have a staging movies library

9:06

where all the new stuff that comes

9:08

in that's honestly frankly not worthy of

9:10

keeping around I just call it

9:12

movie staging and then anything that's

9:14

actually any good I will just manually copy

9:16

across that that folder. Huh I

9:19

like this idea I could do that I could that could that's

9:21

fine yeah cuz it's not like

9:23

we're talking about a volume of input that

9:25

is unmanageable for us like we're talking media

9:28

and it could be media we create here at the studio

9:30

or it could be media that you know it's backed up.

9:33

You might even have some easy sort of scripting things

9:35

where you're like okay I know that in general this

9:37

you know anything older than this maybe I've been reading

9:39

it I can always bring it back but just in

9:41

general flush this every so often. I got this idea

9:44

from my good lady wife who reads ferociously

9:46

yes and she has a

9:49

bookshelf full of books waiting to be

9:51

read and she just called it the

9:53

staging bookcase and anything that's worthy of

9:55

keeping goes on to a separate bookcase

9:57

elsewhere and that's like the collection. and

10:00

then anything else that's no good goes to the

10:02

book box. Yeah, that's a great system actually because

10:04

you can't keep everything. That's

10:07

just, you did just bury yourself in

10:09

storage. All right, yeah, if

10:11

you've got a solution too, I like that one. If you've got

10:13

one that's burning a hole in your pocket, please boost it and

10:15

let us know. And then here's a second question before

10:17

we get into it. I really don't have

10:19

a read on this. I'd like to know your opinion on this too,

10:21

Alex and you, Brent. So for

10:23

the last two weeks for Coda Radio,

10:25

I've experimented with seeing if the Quest

10:28

3 can be used in

10:30

place of the Apple Vision Pro for productivity. Everybody's

10:33

all excited, but for $3,200 less, you can get

10:35

a Quest. Yes,

10:38

there's a MetaTie. But if you

10:40

think about this from an abstraction standpoint about

10:42

is the technology there? And

10:45

Wes and I sorted out about

10:47

two weeks ago how to get the

10:50

Linux desktop into this VR environment. And it's not

10:52

specific to the Quest, but the Quest is sort

10:54

of the technology demo that I'm using. There's

10:57

definitely compromises, but it is doable.

11:00

And I'd just like to know if people are interested to

11:02

hear me blab on about that a bit or not. I

11:05

am tempted to say not, but I thought since

11:07

I have been spending the work and the time,

11:11

and if anybody else is out there considering it, I'd be happy

11:13

to do a review. I don't know, Brent, do you have thoughts

11:15

on that as a topic? Well, I think as someone who isn't

11:18

quite as, I don't know, bleeding

11:20

edge as you are on this

11:22

kind of stuff, I always appreciate

11:24

your perspective on these kind of

11:26

technologies. Especially more affordably. So I don't

11:28

know, I feel like you should do a

11:30

state of VR on

11:33

Linux every six months or something. I'm talking desktops,

11:35

bringing your desktop in and using it in VR.

11:37

What do you think, Alex? Topic

11:40

worth exploring more? So what I'm more interested

11:42

in hearing about is how

11:44

open the Quest is versus Division Pro. Because

11:46

obviously with the Apple thing, it's a walled

11:48

garden and you can only use it in

11:50

the predetermined ways that they've told you you

11:52

can use it. With the

11:55

Quest, I have no idea. Is it a

11:57

lockdown ecosystem? Can you put... cards

12:00

in it can you connect HDMI inputs like

12:03

what can I do with it because there's like the

12:05

screen mirroring stuff in the vision Pro that's super cool

12:07

like if I'm if I'm traveling and

12:09

I want a hundred inch display in my

12:11

hotel room boom the vision pros doing that

12:13

for me yeah can the quest

12:16

do the same from a Linux box I would

12:18

I don't know I would love to know okay

12:20

well I have answers for all this spoiler alert

12:22

though you can sideload apps and I have been

12:24

experimenting with that and that is a big differentiator

12:26

between the vision pro and the quest

12:28

so I'd love to know what the

12:30

audience thinks too if you'd like to hear more I'm not gonna

12:33

do unless I hear a few other people say let's talk about

12:35

it because I think other people listening going oh god VR please

12:39

although when the world's

12:41

one of the world's richest tech companies gets involved you got

12:43

to think something might be there yeah I mean I don't

12:45

know it's one of those areas where

12:47

it'd be nice if Linux didn't totally fall behind even

12:49

if it's not for you you don't use it all

12:51

the time having a foothold in would

12:54

be I will say there's only a couple of

12:56

limited ways to wirelessly bring

12:58

the Linux desktop into a

13:00

VR environment we can have virtual screens

13:02

it is a little tricky put

13:06

it out there before I actually spend the time put them all

13:09

together I suppose all

13:16

right ladies and gentlemen it is time for

13:18

will it nix next cloud edition and

13:20

I will make a disclaimer as we

13:22

realized as we were putting the show together

13:24

for some of you after you thinking wait doesn't Brent work

13:27

at next cloud yes I decided

13:30

to do this and I didn't even think about the

13:32

fact that Brian works at next cloud this

13:35

really is more about the fact that

13:37

I have complained about and

13:39

so as most of the internet I'll put some receipts in the

13:41

in the show notes there's just too

13:43

many ways to install next cloud and unfortunately

13:46

while some of them are great it

13:48

is not clear what is really the most optimal

13:50

depending on what you need and it

13:53

to me seemed like some low-hanging fruit for

13:55

this particular kind of segment could we come

13:57

up with like a lup-blessed way

14:00

to deploy Nextcloud. And

14:02

I mean it's kind of directly relevant, right? Because

14:04

the network has one, many of us have a

14:06

personal one. It's a pretty common

14:08

sort of starting point as people are trying

14:10

to de-Google or explore self-hosting. And the

14:13

question we really had is, it's about at

14:15

a high level, there's common ways

14:17

to deploy popular applications on

14:19

Linux, and now there's the Nix

14:21

way. And there's sometimes multiple Nix

14:23

ways. So what is actually

14:25

the signal in that noise? Is it better

14:27

than say Docker Compose or

14:30

something like that? So we're going to try to compare the

14:32

benefits of each method and tell you which one we're going

14:34

with. And also I just want

14:36

to address ahead of time, it's not a will it

14:38

Ansible segment, but Ansible can be used for a lot

14:40

of this stuff too. But because

14:42

it's a Linux-focused podcast and Nix

14:45

is a Linux-first solution, but

14:47

it's great for other platforms too, we're doing will it Nix.

14:51

But we love Ansible and we're glad you love it

14:53

too. And I think

14:55

the next immediate question that will come into the

14:57

show is, guys, why not the all-in-one

14:59

Nix Cloud? They've built it for

15:01

you, you just run all-in-one. That's right. It's

15:04

built, it's ready to go, it's got some

15:06

nice features like built-in Borg

15:08

backup, Nix Cloud Talk

15:10

ready to go with recording if you want,

15:12

optional clam AV, full

15:15

text search, and they say

15:17

high performance Nix Cloud Talk and turn server, which is

15:19

also optional. So it's

15:22

really a compelling solution out of the box.

15:24

And it's kind of, I think, maybe the most

15:26

official install method you're going to get for

15:29

installing Nix Cloud is this all-in-one, and probably a

15:31

solution that works for most people. Have

15:35

any of us tried it? Oh,

15:37

you and I tried it. Yeah, we've tested it out briefly.

15:41

We've never actually run it for any longer than a month. I

15:45

don't think so. I don't know how you're deploying

15:47

Nix Cloud most commonly. I'm guessing, based

15:50

on our time on the

15:52

show together, I'm guessing it's Docker Compose.

15:55

Maybe the upstream image now? I'm not sure, though. Yeah,

15:58

no, I actually... compile it

16:00

from source code and run it on Gen 2.

16:02

Right? Just do it. No,

16:05

just kidding. Just that one Gen 2 server. Yes, yes,

16:07

yes. No, I, of course, being the Docker

16:09

guy, run it in a container. And have

16:11

done for many years, actually. I've used the

16:14

image from Docker Hub, the official Next Cloud

16:16

image, not the all-in-one. I

16:18

find the idea of

16:20

an all-in-one container extremely

16:24

concerning. It doesn't give me anywhere

16:27

near enough control over the various pieces

16:29

of data that I come to rely on

16:31

through Next Cloud. So mine is a, I'm

16:33

just looking at the source code in my

16:36

infrastructure repo right now. Next

16:38

Cloud 27, backed by a

16:40

MariaDB container. Okay. Yeah,

16:44

I think that's a really common setup too. Wes,

16:47

why don't you tell us a little bit about

16:49

our setup? I think we've done something very similar,

16:52

but we're using Postgres, I believe. Yeah, yeah, Docker

16:54

Compose setup with Postgres and Redis. I think it's

16:56

runted by traffic at the moment. And then we

16:58

went with the S3 backend storage

17:00

for all of the actual files. So

17:03

it's a pretty minimal setup where it just runs those

17:05

containers, the database is the main persistence

17:07

there, and then all the files are up

17:09

in S3. That S3, is

17:12

that the Linode one? Yes, that's

17:14

the Linode object storage. Yeah, you can do

17:16

anything compatible. What would you

17:18

do without a cloud provider

17:20

providing you an S3 backend though? That's always

17:22

the trouble, right? You want

17:24

to run it locally, same? If we weren't in

17:26

the cloud, we'd probably just have a box that

17:29

has more local storage, right? We were

17:31

looking, the problem we always had classically as a team

17:33

running Next Cloud is we'd run out

17:35

of local storage on our VPSs, and

17:38

then the box would break. And

17:40

so our solution to that was to just

17:43

move the backend to S3 and then just periodically go

17:45

through and clean it up. Yeah,

17:47

but we would definitely have to rethink that

17:49

strategy or would we want to back it with

17:51

object storage locally? I don't know that we could

17:53

just do locally too. I

17:55

mean, that's always been a possibility, but it did work pretty

17:57

nicely. Daniel Melzak in the chat.

17:59

That says that TrueNAS Scale has

18:01

S3 support and just acts like

18:04

S3. Didn't know that.

18:06

It's probably Minno I bet. I mean I

18:08

think I saw a lot of these are doing their S3. Me

18:11

and Minno aren't exactly on speaking

18:13

terms after they changed their database

18:15

schemas and lost me my 12

18:17

terabytes of off-site backup without a part of

18:20

the my way. Yeah that's right. Yeah that

18:22

was painful. I was pretty cheesed off about

18:24

that one. Another popular way to deploy

18:26

Nextcloud that looks really great is

18:29

Nextcloud Pi and don't let the name

18:32

confuse you. I want to put it on my Pi. I don't even have

18:34

a Pi. It works on all devices

18:36

all architectures and again this is a ready

18:38

to go. They also have an image for

18:40

virtual machines or like to have one for the

18:42

rock 64. So you could

18:44

you know you could have played that way. So

18:47

there's another way that I think is really appealing and

18:49

then Brantley I'm curious to know it remind me but

18:51

are you are you using the

18:53

snap for for Nextcloud? Is that right? I

18:55

am yeah and it was a

18:58

choice five years ago when I first

19:00

deployed this thing and it's been

19:03

pretty hands-free ever since. So it's

19:05

been stable. I think

19:07

these days it's feeling very molasses

19:09

like so I probably need to

19:12

put some attention to it but that was what

19:14

I chose back then and it was kind of

19:16

like a hey there's a team behind this project

19:18

doing a bunch of work to make sure that

19:21

updates and things are well vetted and just

19:23

work well when you go to do it.

19:26

Well it happens automatically actually which is also kind

19:28

of nice but I got to say

19:30

yeah it's been a great solution for me

19:32

these days. I feel like maybe I would make

19:35

a different choice and many

19:37

audience members have suggested that I should but

19:40

it's one of those projects that made things much easier

19:43

for me back then. I mean that speaks pretty well

19:45

that you haven't had to deal with too much. I

19:47

don't care. I had to do any sort of

19:50

administration or repair or troubleshooting what that was like

19:52

but otherwise it kind of seems like you're doing

19:54

the closest to the all-in-one experience with the snip.

19:57

Yeah have you had a break at all? Only because

19:59

of. storage running out. But no,

20:01

for no other reason, like between updates between versions

20:03

has never been an issue, which I know for

20:05

some folks occasionally can be.

20:07

Yeah, yeah, definitely can be. Yeah. But

20:10

no, it's an in that way. I'm

20:12

like, I don't know, is there a

20:14

solution better than this? And I am,

20:16

you know, curious to find out from folks

20:18

what their experience has been like with updates

20:20

and stuff. So you mentioned in that little

20:23

spiel that you would do something

20:25

different if you're choosing today. I'm

20:27

curious what that would be. And then

20:29

also why? Well, Alex, every time I

20:32

describe my situation

20:35

with my Nextcloud server to you, and

20:37

the fact that I have like 10 active

20:39

users on it, and the way it's set

20:42

up, you tell me that I'm in

20:44

great need of some attention. So I was just

20:46

thinking like, okay, more users than I do. Perfect

20:49

for like, oh, my solo Nextcloud that I just

20:51

run at home. I know that I've got a

20:53

family of five. And I don't have as many

20:55

users as you do. Yeah, well, I

20:57

think maybe I've grown it. But

21:00

also, I think there are best

21:02

practices that I just didn't know existed

21:04

back then. And so

21:06

in that way, I feel like maybe

21:08

I'm pushing things a little bit and I need to be a bit

21:11

more careful about my setup. And, you

21:14

know, off the public internet, sounds like a

21:17

great option to me. It does. I think,

21:19

you know, the single biggest thing that anybody

21:21

can do, in my opinion, to improve the

21:23

long term reliability of not just Nextcloud,

21:25

but any self hosted app, is

21:28

to have some way to snapshot the

21:30

app data between upgrades. So

21:32

for example, for all of my

21:34

containers, I run my app data

21:36

as bind mount volumes as directories

21:38

into those containers backed by ZFS

21:40

datasets, which are then replicated across

21:42

the world in various different ways.

21:45

But it allows me to have that confidence

21:47

if something does go tits up, I've got

21:49

six months worth of backups to look because

21:51

sometimes it takes me a while to notice

21:53

it's broken. You know, it's like daily often

21:56

isn't enough for me to actually notice it's

21:58

broken. Great point,

22:00

Alex. It will take me

22:02

a week or two if I'm lucky. That's

22:04

a good point. My setup has

22:06

been really simple at home, and I've been very happy

22:08

with it, but not perfectly happy because

22:10

it has been a little bit of

22:12

a chore to maintain. The part that

22:14

I think opened me up to

22:17

our initial premise for this episode is

22:20

we solved the reverse proxy

22:22

by just using a dramatically

22:24

simple nginx config in

22:27

Nix that I think Wes might have been

22:29

five lines to define

22:33

all of that, and I realized, oh, this is

22:35

very powerful now because Nix will always just be

22:38

configured with the settings. Nix will always

22:40

just deploy nginx with

22:42

these configurations, and my reverse proxy is just locked

22:44

in now. I started realizing

22:46

the power of that, but continued to run

22:49

Docker as a container. Historical

22:51

momentum, I just continued to use the Linux server

22:53

IO container, which has been pretty great other than

22:55

it uses SQLite. I

22:58

mean, to your point, as much as we love to

23:00

try weird, wacky new things, there is certainly a reason

23:02

not to change things if it just works and it's

23:04

part of your infra. Yeah, yeah. It's

23:06

just slow, and I have had a

23:09

couple of database problems that Wes has had to

23:11

help me resolve, and I

23:13

have had a couple of upgrades where I've

23:15

had to go into the OCC command and

23:17

fix things or manually update the apps before

23:19

things will go out of maintenance mode. I've

23:21

had to do hands-on

23:23

for every single major upgrade so

23:25

far for the Nix cloud component. Yeah,

23:28

I would say you've not necessarily had an easy experience.

23:30

You've had a worse time than the one

23:32

we use for transferring show files. Yeah, yeah.

23:35

Well, I definitely probably didn't help

23:37

myself with when I adopted the Linux server

23:39

IO image, you had to update the two

23:42

independently, both the image and then also

23:44

Nix cloud, and now they've changed that

23:46

into more recent releases, which is beautiful, so that's made

23:48

that a little bit easier, but

23:50

that hadn't been my experience so far. And then

23:52

you combine that with having to keep the apps updated

23:55

and then All of

23:57

that. So I Really wanted something that was much

23:59

more maintainable. And also I

24:01

don't need all like all the stuff

24:03

that the own one or the pie.

24:06

Image the one. a little customize ability a little

24:09

bit, but not a lot. And. I don't

24:11

want to become a big headache to maintain. So. That's

24:13

kind of were all the citizens the existing

24:15

ways to the point. next lot of pretty

24:17

great including the community maintained docker image protests

24:19

like. Tons of five, five on

24:21

five hundred million poles or something like that. So

24:23

there are a lot of ways people have attempted

24:25

to solve this. But. I. I'm.

24:28

Feeling pretty good about the way. We. Figured

24:30

out I'm feeling pretty good. And the nice thing

24:33

is. When. We're all done here.

24:35

People can go look at what we posted on

24:37

a good hub and they should be able to

24:39

replicate it immediately on their system with very little

24:42

changes. Warped.

24:45

Dot Dev/linux Das Terminal yeah, the

24:47

Warp Terminal. The one with a I

24:50

built in built on top of rust.

24:52

that sweet terminal that I have watched

24:54

jealously over there and Mac O S.

24:57

Land. It's come

24:59

under linux later this month. You. Can join

25:01

the wait list and sign up for the launch

25:03

party and get some sweet linux wag when you

25:05

go to war.dev. Slash. Linux

25:07

das Tunnel if you're not familiar. I've been

25:10

watching for a while and. For. Wanting

25:12

this but you're not familiar or what gets

25:14

me excited about it is that it is

25:16

a modern command line built on Russell. super

25:18

fast and perform and. It's. Not

25:20

like some Electron thing. It's really great as

25:23

a I built in with a beautiful way

25:25

to discover the commands that you might be

25:27

looking for with a I suggestion and it's

25:29

highly customized. We can create your own problems,

25:31

the can recall really quick. It has collaborative

25:33

features. There's this feature called warp Drive. The

25:36

less you save parameter I command and and

25:38

run them later and share them with your

25:40

teammates is just a great user experience. Kind

25:42

of feels like the terminal. Hasn't.

25:45

Seen a lot of interesting developments for

25:47

a while and this. This. Is

25:49

really been it and I'm really excited that a seat actually

25:51

coming to Linux. I was. As. I

25:53

soldiers at the Mac users got on healing and

25:55

really appreciate the command line over there, Maquis, they

25:57

just use cove I have. But. as

25:59

linux users enjoy the terminal

26:02

and you're gonna enjoy warp terminal warp.dev

26:04

slash Linux dash terminal to sign

26:06

up and support the show again

26:08

it's a little tricky it's warp.dev slash

26:12

Linux dash terminal that's where you sign

26:14

up get the Linux swag pack perhaps

26:16

maybe even get a little bit info

26:18

about when it comes out before everybody

26:20

else and support the show again that's

26:23

warp.dev slash Linux dash terminal

26:29

well like everything in the world

26:32

of open source software there

26:34

are I don't know a dozen probably more than

26:36

that way we could use nix and nix related

26:38

tooling to try to get ourselves a next cloud

26:40

server but you know we didn't necessarily want to

26:43

have to write a whole bunch of nix ourselves

26:45

try to package stuff from scratch ideally

26:47

we could find something that would be approachable to

26:49

someone just starting out with nix if they want

26:51

to try to move over some of their infrastructure

26:53

but you know also workable for someone has been

26:55

using nix for a while and

26:58

maybe something that you could just copy from github and

27:00

use on your system if you were curious I want

27:02

to try it after the show and

27:04

thankfully this

27:07

went better I think then we had any

27:09

right to hope because next cloud is just

27:11

really nicely supported to nix yeah

27:13

there is a lot of really nicely

27:15

supported projects it's kind of like the

27:17

list of top like awesome self-hosting

27:20

awesome open source filled

27:22

with modules on their github that are kind

27:25

of essentially you can almost call them playbooks if

27:27

you will that build you

27:30

a next cloud system and you

27:32

just tweak a few things in there and

27:34

you can read this pretty clearly so

27:37

from a practical standpoint I

27:39

guess the way we did this is we created an

27:41

xcloud nix file we put that on our system and

27:44

we were eating we did what's essentially called

27:46

an include to include that new

27:48

configuration file and then in there

27:51

we put this module and it's a

27:53

building block for the system and it is

27:55

very easy to read even if you're not

27:57

familiar with the nix language Because

28:00

you go through this and it has to build

28:02

properly if you've typod something or

28:05

you haven't configured something correctly It'll

28:07

catch it much like software won't build

28:09

if you haven't written the software correctly

28:12

if you haven't written the config file correctly It won't

28:14

build and so you

28:16

could make a tweak attempt to build. Okay. I gotta go fix this

28:18

on line 20 All right. I'll go tweak that okay build again All

28:20

right I gotta go fix this on line 40 and

28:23

the things like 52 lines long and you have

28:25

a complete next cloud system Yeah, it ends up

28:27

being a kind of nice fun Interactive

28:29

experience once you get past the like oh gosh, it's

28:31

freaking out. What is it yelling at me about? You

28:34

know you you get to test things the system tells

28:36

you where you're wrong And you know there's a there were a

28:38

few different places where we have had questions as we were getting

28:40

going or had to tweak Things to get it just right so

28:43

you knew once you'd gotten past just oh, it's

28:45

it's airing on a new line now Okay, that

28:48

section must be working and it was Some

28:50

of those error messages can be a

28:53

little cryptic just fair warning But

28:56

the best way to kind of work around that

28:58

is if you look for like a line number

29:00

It almost always give you the line number somewhere

29:02

that it failed originally and then it gives you

29:04

a good jumping off point in your config Yeah,

29:06

that is my go-to and what you

29:09

end up with is something that sets up an

29:11

X cloud system It sets up your TLS

29:14

certificate through let's encrypt You

29:16

can define the applications that get to

29:19

get deployed inside next cloud so My

29:22

wife loves the cookbook app so we

29:24

made sure we deployed the cookbook app we

29:26

deployed several apps just a test You

29:28

get to define the exact version of next cloud that

29:31

gets installed Where the configuration

29:33

paths are where like the secrets path

29:35

is and then additionally? You can take

29:37

advantage of other things in there because

29:39

it's very expressive language And you can

29:41

set up things like a Postgres database

29:43

backup, and you can say every night Back

29:46

up the next cloud database and save it to this directory

29:48

and that can all be part of the module so

29:50

when you deploy it all of that

29:52

from Postgres and Redis and next cloud and

29:54

the ports and The TLS certificate

29:57

all of that just gets configured and set up

29:59

for you But it's not a black box.

30:02

It's not like the all-in-one container where

30:04

you just get something set up for you. You don't really understand

30:06

how it works because it's an opaque box. This

30:09

is something where you go through this 50-ish

30:11

lines of syntax and understand

30:13

exactly, line by line, what this thing

30:15

is doing because it's human-readable. And

30:18

it's a map. It's

30:20

a self-documenting map of how your next

30:22

cloud works. And even if you don't

30:24

touch the thing for two years, you can open this back

30:27

up and understand how you're getting the

30:29

TLS certificate, understanding what domain name it's using, how

30:31

it's getting the host name, all of that. I

30:34

think that's very powerful. It strikes me,

30:36

too, that a lot of this is possible because

30:38

it gets to reuse so much of what's already

30:40

in Nix packages. So

30:43

there's two components. There's the next cloud, like

30:45

the actual package itself, the files that you're going

30:47

to need on your file system. And then there's

30:49

the next cloud service, which is provided in Nix

30:51

packages, which ties all this together. And that's what's

30:53

able to let you automatically create the

30:55

database because, well, Postgres is well-supported in Nix packages

30:58

already. There's a service for that. So it doesn't

31:00

have to reinvent the wheel. It can just go,

31:02

oh, yeah, turn on Postgres, turn on Redis, if

31:04

you want. And it just relies

31:06

on what's already built in Nix packages. And

31:08

I think it especially works nicely for next

31:11

cloud because next cloud is a PHP application.

31:13

And Nix has great support for nginx. And so you

31:15

can just kind of have like, there's

31:17

a lot of stuff you need to do to interface with

31:20

nginx and PHP to get it all to execute together. Or

31:22

if you want to customize things, it's

31:24

already in the Nix strength. Like something we know we

31:26

have to change is we need to set a larger

31:28

upload limit and things like that. You can set that

31:30

in there. Some of those are PHP config options that

31:33

you're just setting in this module. I

31:35

think the trickiest thing is trying to figure

31:37

out which module to use. But most of

31:39

them just live on the Nix GitHub

31:42

project page. So you just go to their

31:44

page. And there's official packaged modules for you.

31:46

So it's pretty easy to narrow them

31:48

down. You and I, I

31:50

think, were the newest to this process. Wes was already

31:52

familiar with modules when we started. And

31:55

you're looking at this probably from the perspective

31:57

of I'd like to maybe replace my Snap package one

31:59

day. I'm curious what your impressions

32:01

were and your thoughts about it. My

32:03

initial thought was that this is going to

32:05

be, again, way too complex for

32:08

someone like me who can probably be defined

32:10

as quite a home user. You know, I've

32:12

never had experience professionally

32:14

as a system administrator or anything

32:17

like that. So learning

32:19

some of these technologies for me, like Alex can

32:21

say, reverse proxies has been

32:23

a challenge for me, right?

32:25

It's a thing I haven't learned. You don't give

32:27

yourself enough credit is another thing I'll say right

32:29

now. When we hack

32:31

together, Brent, you know a surprising amount of

32:34

stuff. So, I will agree with

32:36

that. Oh, well, now you got me

32:38

blushing. But I think what attracts

32:40

me to some of

32:42

these projects that are already done for you,

32:44

like the all in one or the snap,

32:48

is that I don't have the background

32:50

necessarily to have a bunch of best

32:52

practices. And so I really appreciate the

32:54

like, here's a good place

32:56

to start. And everything's already kind of well defined

32:58

for you. And you're not

33:00

going to expose things that you didn't intend

33:02

to because, you know, I just don't know

33:04

any better. Although maybe with the snap, I'm

33:06

doing that, but who knows? So

33:09

that was my fear with going

33:11

into this project together to nixify

33:13

our next cloud was

33:15

that I figured it would just end up

33:17

being too complex. But

33:19

it turns out now I

33:21

had like these massive light bulb moments going

33:24

off using these modules because it's a super

33:27

nice blend of using

33:29

best practices. Someone's already really thought about

33:31

this. And that's, you know,

33:33

to become a nix module, there's

33:35

a review process. And there's a team of

33:37

people looking at this stuff. And

33:41

yet, you can use

33:43

those defaults, but then you can also change

33:45

them all. So if you're a more advanced

33:47

user, you can use that baseline and just

33:49

really customize it the way some

33:51

people say with Docker

33:53

images and stuff like that is a really nice

33:56

thing. Yeah, and we did just that too. I

33:58

would argue though that I don't

34:00

mean to sound gatekeeper here, it's kind of a

34:03

fine line, but I would argue that that

34:05

simplicity of jumping off that

34:07

you get through an all-in-one

34:09

container, you're kind of just kicking that

34:12

can down the road and by not

34:15

fully understanding what it is quite that

34:17

you're deploying, you do run the risk

34:19

of further down the road being locked

34:21

into a deployment methodology

34:24

that is difficult to migrate off of or

34:27

is vendor locked like a snap or

34:29

etc. etc. You get the general

34:31

idea and what really thrills me about

34:33

these Nix configs is the fact that

34:36

you can't, at least to my knowledge in a

34:39

Docker compose declaration, say

34:42

I want these applications installed inside

34:44

the app that I'm deploying. I

34:46

mean you can if the container

34:50

manufacturer, that's not the word, if the

34:52

container maintainer exposes those

34:54

knobs and whistles to you,

34:57

but the fact that the

34:59

next cloud module in

35:01

Nix for example, lets Chris

35:03

deploy QO notes and the

35:05

cookbook and tasks and you

35:08

get the idea, there's what six or

35:10

eight apps deployed inside the application

35:12

you're deploying here, the

35:15

entire application is defined

35:17

in this file, declared in this

35:19

file and this sounds

35:21

like a broken record, but this is where the magic

35:23

of Nix really starts to come home, it's like everything

35:26

is there and like you said Chris, if

35:28

you're looking at this file in two years

35:30

time, there's no ambiguity about

35:33

well what did I do, what random

35:35

command did I run that's now left

35:37

my bash history etc.

35:39

etc. It's literally in

35:42

the config, that is the source of truth.

35:44

Yeah and it also

35:47

means that I can be very intentional about

35:50

when I rev the next cloud environment

35:52

because we have one line in

35:55

here, one line and

35:57

it just is next cloud 28. And

36:00

I'm going to keep that at Nextcloud 28

36:02

for a couple of months after Nextcloud 29 comes out.

36:06

And then I just go in there, I rev that

36:08

to 29, do a rebuild, and

36:10

now I've got all the new apps for Nextcloud, and

36:12

I got the new Nextcloud install and any kind of

36:14

Redis updates that might need to be done. All

36:17

of that happens at that point. And

36:21

I like the intentionality of that because Nextcloud

36:23

is the back end to manage my phones.

36:25

It manages a lot of the family calendar

36:28

stuff. It manages the cookbook stuff. So it's

36:30

like dinner is kind of important to

36:32

me and I want that to work. So I want

36:34

to be very intentional when I actually rev that Nextcloud

36:36

version, but I don't want it to be a big

36:39

hoopla where I have to go in there and

36:41

do a bunch of commands that I execute through

36:43

Docker and all that kind of stuff. Yeah, if

36:45

you do choose to opt into the declarative apps

36:47

too, you get the benefit of the app sort

36:49

of saying which versions of Nextcloud they support, and

36:51

that gets encoded into Nix. And

36:53

so if one of your apps wasn't yet

36:55

updated for the new release and it

36:57

wasn't packaged for that release in Nix, your

37:01

build would just break and you'd be like, oh, okay, I guess

37:03

I'm not updating this. I'll wait another week or

37:05

two and try it out then. It sounds like that's a little

37:08

bit of a sense of security too. Chris,

37:10

I at first was really confused by that

37:12

because the other modules that

37:15

I've used previously and just like the Nix

37:18

config on my laptop have

37:20

never had version numbers tied to them. Yeah. And

37:23

so I thought that was a little bit strange, but I

37:25

kind of expected to be rolling at first too. I just

37:27

thought it would be always current Nextcloud. But

37:29

I now after thinking about it, I don't

37:32

think that's actually what I want for

37:34

my production Nextcloud. Well,

37:36

and it's kind of decoupling your

37:38

Nextcloud, which is a somewhat

37:40

complex beast to

37:43

your actual system. I guess

37:45

you were saying that and that was one of

37:47

the light bulbs that went off for me in

37:49

this specific module with these specific version numbers to

37:52

it is like, actually, you want that

37:54

to state. You

37:56

know, one of the reasons that upgrades

37:58

with Nextcloud in particular. can

38:00

be so dicey is that when you

38:03

do a next-cloud upgrade it's changing the

38:05

database schema underneath so it goes into

38:07

the MySQL or the SQLite or whatever

38:10

and it will add or remove certain

38:12

columns and change you know data types

38:14

or whatever like typical stuff

38:16

in software upgrade release cycles

38:18

and I used to work for

38:21

a bank in London and I that we're talking you know

38:23

they did like 60% of the

38:25

UK's credit and debit card transactions and so

38:27

you can imagine the database that ran that

38:29

thing was was pretty beefy and it was

38:31

an Oracle thing and the amount

38:33

of time these guys spent on building

38:35

rollback scripts and you know all this

38:37

kind of stuff for this database because

38:40

even a minute of downtime for a bank

38:42

is just not an option and

38:45

for next cloud every time

38:47

I've had an issue with an upgrade it's

38:49

been some kind of database wonkiness and

38:52

the reason that's so important and I mention it

38:54

here is because rolling back unless you have a

38:56

snapshot you don't really know what

38:58

state the database is in unless you go

39:01

through each column in each field manually

39:03

one by one and I ain't got time for that so

39:06

take a snapshot of your database before you make any

39:08

changes and then that way it's kind of like a

39:11

risk free thing but then you know Nix

39:13

takes it a level further because you're basically

39:15

building a new environment you're switching

39:17

into and you can just roll back to

39:19

the last one and you're good to go

39:22

so I think comparing and contrasting this

39:25

process to the traditional process I

39:27

jokingly describe it as it's

39:29

a one-click deployment after you do 30 ish

39:32

various different things to make it work so

39:34

it's a lot of front work you'd go

39:37

through this thing not building you change paths you

39:39

figure out where you need to put stuff on

39:41

your file system for like a secrets file you

39:44

spend the time figuring out how to make a module work if

39:46

it's new to you and so maybe an

39:48

hour right or two whatever it is and

39:50

then you build it and you're done

39:53

it's the weirdest thing where typically when

39:55

you're deploying server software on Linux you

39:58

deploy it And. Then you go configure

40:00

it. And. Then you spend. A.

40:03

Lot of your time actually configuring the

40:05

software. And. You're setting up accounts and

40:07

sign up all the apps Arabia is whenever. maybe whatever

40:09

your to points. But. In this

40:11

scenario it's the reverse. You do all the

40:13

work ahead of time to figure to figure

40:15

out which need to do and configure the

40:17

everything. And. Then you tell to build. And.

40:20

It's done. And not only that done but

40:22

like of set up the auto back up with the

40:24

script that we have a daughter backups the post rest database

40:26

you gotta get that five there but it's it takes

40:28

care of the backups for you, it takes care of the

40:31

Ssl cert for you what does it renews assert for us

40:33

to I can't remember but I think it doesn't demand

40:35

is yes it even renew the search. Was

40:37

more than a sense to up in about. the

40:39

Next up is a lot of things and Up

40:41

and Next O S as as you know it

40:43

renders out the various template files or conflict files

40:45

that it that it spits out. Those are everything

40:47

goes in the store. Sienna know that and then

40:49

it gets all tied together. A

40:51

while it's just by system these ultimately you end up with.

40:54

System. D Timers and system D Services that.

40:57

You can look at you can look at the

40:59

logs for eating of view the actual unify old

41:01

underneath. And. It's just

41:03

you know, a bunch of next stuff tied

41:05

together with standard systems south the make Linux

41:07

work, make a network. So if

41:09

you're looking of where to find more

41:11

knicks modules search.mix so as.org and if

41:13

you go across to the which one

41:15

is it is it. The options are

41:17

packages of i forget which one exactly

41:20

options yeah since and he he just

41:22

typed services dot next cloud you see

41:24

all of the different options but there

41:26

are hundreds of different reasons that these

41:28

modules expose and then it's a reusable

41:30

patents. If you wanted to play for

41:32

me Cs next week you've kind of

41:34

got that same deployment patton and you

41:36

can start to reuse a lot of

41:38

these principles across different applications. And the

41:40

reason I say that is because in

41:42

all chat right now in a folks

41:45

are saying that by initial build off

41:47

the next got image is a big

41:49

hurdle specifically for newbies and I would

41:51

argue yes that very first one is

41:53

a big lift but it gets significantly

41:55

easier x each time you do it

41:57

after that because you to freezing those

41:59

same. Yes, absolutely true.

42:02

You can really just kind of go look at your old

42:04

configs and be like, okay, how did I do? Okay, so

42:06

that's what I'm gonna do for this application now. That

42:09

is very true. And you know as as Nix

42:12

OS has gotten more popular these last years,

42:14

there's more and more example configs I

42:16

mean we're gonna have have ours out

42:18

there And so, you know, maybe

42:21

you can get a little bit more of a foothold find

42:23

something that pretty much works You just got to change like

42:25

some API credentials or you know, any of the little secret

42:27

bits But something you can play with

42:29

in a VM where there's like no cost You don't need it

42:32

to work and you can actually take take the

42:34

hour or two that it probably will take to try to get

42:36

a first It does initially take a bit. Yeah, but

42:38

it is like they like they both are making the point it

42:41

is worth that investment Um, I you know one thing

42:43

I wanted. Yeah, I was curious about here is like

42:45

what were the edge cases It's easy to get just

42:48

like a next cloud going. Yeah, like you're running nice

42:50

cloud the standard service But like we knew we wanted

42:52

some apps and so having the ability to add custom

42:54

apps if we need to ones that were not necessarily

42:56

Packaged by the Nix community. Yeah, how hard was that

42:58

gonna be? Yeah, that turned out to be

43:01

really easy Yeah so that knocked it down right there

43:03

and then I think it was just nice going through

43:05

sort of using a bit and and and Configuring it

43:07

as you're talking about because you know We

43:09

checked the little self-check that nice cloud very helpfully

43:11

has about like hey How does your installation look

43:14

and it had some complaints for us? And so

43:16

being able to go set easily the PHP options

43:19

in Nix is like was that gonna be a whole thing where we had to Custom

43:22

render out a file somewhere in here

43:24

or was it gonna be literally like

43:26

PHP options dot option name equals The

43:28

new value that next cloud recommended

43:31

to us and it worked and it's so

43:33

nice because sometimes you run it with areas

43:35

And you know these these Nix services are

43:38

more or less sophisticated Sometimes it's just sort of like

43:40

making a cron job for you and setting up

43:43

a default installation And sometimes it's managing a lot

43:45

of the details like here with next cloud You

43:48

know You don't quite know where that support necessarily is gonna

43:50

end and when you're gonna have to do a little

43:52

more custom Nix yourself But with next cloud it

43:54

seems like it's it's very well covered. So I'll

43:56

tell you my my takeaway

43:59

and why I'm going to redo my

44:01

home setup like this, besides

44:04

the fact that it's sustainable and

44:06

self-documenting. For me,

44:08

I think it has the advantage over, say, like

44:12

the Nextcloud Pi or all-in-one images, in that I

44:14

can get Nextcloud releases a little bit faster,

44:18

because those bundled, for good reasons, those bundled distributions,

44:20

wait a little bit to test, to

44:24

make sure they have application compatibility, to do all of that stuff, which is

44:27

fine, good, I'm glad they do. So I can have

44:29

the same kind of, almost, it's my own custom

44:31

all-in-one setup, using

44:33

all open components, well-documented, and I can update

44:35

it on my schedule, as soon as the

44:37

applications are published in the Nix package repository

44:39

or, you know, doing the outside

44:41

of the repository. I think that's a big advantage.

44:44

Brent, though, I suspect your

44:47

big advantage, the thing that I think

44:49

impressed you, because that was mine, might have been the performance

44:51

of the thing. Yeah, I,

44:54

you know, we went through that, at least, as you're

44:56

talking about, about, like, figuring it, and

44:59

also deploying it at the same time. So

45:01

we went through a few iterations of, you

45:03

know, doing a Nix

45:06

OS rebuild switch and complaining about something.

45:08

But once we got that, it

45:11

was, like, amazing how quickly everything was just

45:13

like, oh yeah, the site's ready at the URL

45:15

that we have, and

45:17

that it was the performance that really hit

45:19

me. Like, I, as you might imagine, used

45:22

a few Nix cloud deployments,

45:24

my own, the one we

45:26

use here at JBE, and our internal one

45:28

at Nix cloud itself,

45:31

and including some, like, technical

45:33

previews that we have. And

45:36

I gotta say, this one was the most

45:38

snappy, performant one I've used to date. Like,

45:40

it was also empty. But

45:42

I don't know what it was. I just

45:45

had this, like, feeling that it was more

45:47

performant than anything I'd ever tried

45:49

before. And I'm curious if that's how you

45:51

guys felt, or I'll be curious to see

45:53

if anyone tries our config, if that's how

45:55

you feel as well on various hardware. Yeah,

45:58

that's a great question. I would

46:00

like to know if people grab our config and

46:02

they experience that, please let us know. Because what

46:04

else I think was a little like, it seems fine.

46:07

But you and me were like, this is the fastest

46:09

Nextcloud I've ever used. And I made the comment to

46:11

Brent this morning before the show, that

46:13

if my experience with Nextcloud has always been this

46:15

performant, I think I would have been more inclined

46:18

to use the web apps than

46:20

I have been. I traditionally use Nextcloud more

46:22

as like an API endpoint and file sync.

46:24

A little more asynchronous. Yeah, I don't interact

46:26

with the web applications because I

46:29

barely can tolerate how fast native applications

46:31

launch, right? So, God forbid, a

46:33

web application. But this is something

46:35

else. And it's either, maybe it's the box.

46:38

It is a nice box we have it

46:40

on. Don't forget that quality Postgres database. But

46:43

he's accessing it remotely

46:45

from Canada. And he's

46:47

experiencing the same thing I'm experiencing on the LAN

46:49

performance wise. And I think

46:52

maybe it's just some of the optimizations

46:54

in the PHP config. I don't know what it

46:56

is. But I'd like to know

46:58

if other people, is it Redis? That could

47:00

be it. I mean, I'm not using Redis in mind. I think that

47:03

could be it too. I do think it could be that. But

47:06

that just brings me back to the point of

47:08

having some well thought out defaults built into these

47:10

modules. I was like, I don't even

47:12

really know what Redis is. Like I kind of do now.

47:15

But that's just there by default. And I know

47:17

that's the case in some of these pre-made

47:21

like all in one and such. But I

47:24

don't know. It just seems so easy here.

47:26

And fair enough, we had a WaaS helping

47:28

us out to build this thing. But

47:31

now that's built, like the entire JB community

47:33

has access to it. And we'd love to

47:35

hear optimizations that

47:37

you're adding to it and like change. Maybe we

47:39

did something silly you don't think we should have

47:41

done. That would be amazing feedback.

47:43

Yeah, or maybe there's something we should be adding.

47:47

I mean, it's on our GitHub. So pull

47:49

requests, like bring them on. We'd

47:51

love to see them. Let's get it. Yeah, and

47:53

also, we'd love your suggestions on future applications. We

47:55

should see if they could mix. They don't necessarily

47:57

need to be server stuff. It could be desk.

48:00

Software. It's really just trying to look at. You.

48:02

Know is there a better way to do this in Linux? Final:

48:05

Cut pro ana it

48:07

you bastard. Alex

48:12

I'm curious. did you give our

48:14

little next cloud can seek a

48:16

try and if you did, how

48:18

was your experience vs doing it

48:20

to traditional i don't know answerable

48:22

way for instance. He

48:24

and any teabagger like thoughts on. Or.

48:27

Might might additional answer. the way

48:29

is actually just docker compose file

48:31

way. I've my instances like sake

48:33

of seven like a State started

48:35

in London and it's the same

48:37

databases, the state and it's the

48:39

same files since then. All.

48:42

Of my photos go through it. All

48:44

of my contacts calendars like on the stuff

48:46

go for as well so is probably

48:48

getting and it'll crusty so much like

48:50

Chris and he self like. There's.

48:52

Probably some database clean up the could improve

48:55

performance or that she just looking at the

48:57

yeah the men caching stuff when you mentioned

48:59

read is that for example allows you to

49:01

take cash. Some of the queues of things

49:04

coming in, some of the requests he memory

49:06

that's how it but by a speed things

49:08

up their little bit was hoping to dig

49:10

into before I spoke again was whether the

49:13

knicks O S module uses some of those

49:15

men cast things to speed things up underneath

49:17

in P x P configuration falco. That would

49:19

explain why it feels so snappy when I

49:21

tried it out. alive.com/unplugged

49:26

if you're an I T

49:29

security area particularly. You.

49:31

Know you've been dealing with these kinds of

49:33

recurring problem for a while. Data breaches the

49:35

come from the end users, me, see babies,

49:38

his credentials, whatever it might be some even

49:40

necessarily their fault. That. Man

49:42

is it just.a constant nuisance and

49:44

requires I died. I tease attention

49:46

all the time. All these little tickets

49:48

the come in to solve a problem that. Really?

49:51

Could be handled better. Even.

49:53

Something simple as just making sure everybody software's

49:55

up to date before they connect your cloud

49:57

as. Well. that's where

49:59

collide It's a solution to this challenge.

50:02

It's something that finally takes care of this layer, so

50:04

that way IT can be freed up and

50:06

employees can be empowered to fix it

50:08

directly through your existing messaging system following

50:11

your procedures. Collide

50:13

works to make sure that devices can only

50:15

access your apps after they've passed the checks.

50:17

They also give you a dashboard where you

50:19

can see your Mac, Windows, and Linux machines

50:21

from one pane of glass, manage

50:24

everything, check the state of things, run

50:26

reports, run audits. I

50:29

think you should go experience this

50:31

firsthand. They got a video set

50:33

up at collide.com. That's kolide.com/unplugged. You

50:35

go over there, you see what I'm talking about,

50:38

and you support the show, too. It's not

50:40

really their fault. They just need better tooling. They

50:42

need better solutions, and they don't necessarily – I'm

50:44

talking about end users – need to be burdening

50:46

IT with every single problem. If

50:49

I had this when I was in IT, I think it would

50:51

have been a game changer. I'm passionate about this because I

50:53

think it could make your life easier, too. collide.com/unplugged.

50:56

That's kolide.com/unplugged. Go check out their

50:58

videos, support the show, and see

51:00

what I'm talking about. collide.com/unplugged.

51:09

If you can believe it, boys, there's five episodes

51:13

left until we get to

51:15

scale. It's coming up really soon. That

51:17

also means NixCon North America. Oh, boy.

51:19

All kicks off March 14th in

51:22

Pasadena, California, and

51:24

our audience made this trip possible, so we

51:26

will have a full livestream schedule on the

51:28

way down while we're there before, after Nix,

51:30

all of it, plus a live Linux unplugged.

51:33

We'll be putting all of that out more, but just get

51:35

yourselves ready, mentally prepared for that. Even if you're not going

51:38

to be there, you'll get a little taste of it. If

51:40

you are going to be there, get the full thing in your mouth.

51:43

We'll be doing lunch on the

51:45

16th. meetup.com/Jupiter Broadcasting for those details.

51:48

Who doesn't want more Chris Fisher in their

51:50

mouth, huh? Oh, yeah. Big, big burger named

51:52

Chris? No. And then we could all complain

51:54

about it in the scale chat. We'll have

51:56

a matrix chat set up for that, too.

52:00

gonna be a good time Alex is gonna be

52:02

there so we're gonna have a little

52:04

Airbnb down there where the crew we do some

52:06

stuff and out some more we do some barbecuing

52:08

maybe yeah burn some more yeah the same one

52:10

no no I don't know cuz that had such

52:12

a great grill last it did but also had

52:15

that grease fire yeah but you know flavor yeah

52:17

you work with it think of the charl I

52:19

was impressed that's when I knew the man the

52:21

man was a grill artist where he leaned into

52:23

the grease fire and just worked with it yeah

52:25

I was impressed no

52:28

we're getting something a little closer to the

52:30

venue this time I think so we'll be

52:32

like within walking distance of scale which will

52:34

be nice nice yeah it's good little yeah

52:36

downtown Pasadena is supposed to be quite nice

52:38

yeah in the spring I think so friendly

52:41

we got some emails this week we wanted to

52:43

get to we got sent a lot of suggestions

52:45

on how both you would run a next

52:48

cloud and also how you're currently doing it so

52:50

thanks everyone but we did get a little piece

52:52

of mail this week from thought criminal that I

52:55

thought actually was just perfect

52:57

and you couldn't write it any better hey

52:59

guys this is probably too long for the

53:01

show but too coincidental not to share the

53:03

next challenge came up in my busy season

53:06

so I was always gonna be late to

53:08

the party things finally slowed down

53:10

for me though and so I ordered a little

53:12

Dell micro PC with the intention of putting next

53:14

OS on it being a lazy

53:16

admin type I'd settled to going with

53:19

snowflake and to my delight you guys

53:21

ran a whole show on that very

53:23

topic I finally got

53:25

my hardware in order tamed the

53:27

Medusa of eye watering sneeze inducing

53:29

cables under my workspace disaster area

53:32

and then tamed the disaster area

53:34

itself because I've learned the Linux

53:36

gods demand a good blood sneeze

53:38

sweat and tears sacrifice before a

53:41

good journey a good journey fun

53:43

being honest cheat coat snowflake

53:45

did all the heavy lifting and I just

53:48

stacked myself a quick win and I also

53:50

slapped each top on there because you know

53:52

Linux unplugged as I lucked into

53:54

services and started wrapping my head around the idea

53:56

of one config file to rule them all the

53:58

magic of the deal finally clicked for me.

54:01

I started to realize that this just

54:03

might be the mythical distro of power

54:06

that finally helps me slay the dragon

54:08

that has been my own many

54:11

next-cloud implementations, each

54:13

more glorious than the last,

54:15

all poorly documented in some

54:17

fashion hacky and inevitably fatally

54:19

flawed. And then, as

54:21

if the whole thing had been preordained, what

54:24

should come up on the very next Linux

54:26

Unplugged episode? It's like you guys are right

54:28

inside my head. I cannot wait to see

54:30

what this community comes up with on this

54:32

episode. That's amazing thought, Criminal. Thank you very

54:35

much. I'm glad we could be of service there. And let

54:37

us know how it works, too. I

54:39

always like, you know, we talk about this stuff, but what's the long term?

54:41

You know, what's the longer? Yeah, do you

54:43

stick with it? Do you go somewhere else? It's hilarious. Yeah.

54:46

He's going to write it in six

54:48

months and say, I've grown Gandalf, Link,

54:50

Beard, and... Yeah, that's how it goes with

54:52

it. He was right back in the day. I'm redoing the whole thing. It's

54:54

all going to be based on flakes. Now

54:57

the maintainer over there. Flakes

55:00

are a whole thing. They're worth it. I promise

55:02

you about that. That's pretty difficult. And

55:06

now it is time for the boost.

55:09

And our baller booster this week is

55:11

listener Jeff. He comes in with 81,361

55:13

sats. And

55:20

the biggest one there was, of course, the birthday boost

55:22

with 50,000 sats. He says,

55:24

surely I'm not the only late birthday

55:26

boost. Happy birthday. The answer to the

55:29

ultimate question. Thank you, Jeff. Appreciate that.

55:31

You know, I feel like you were like, oh, crap, these guys are

55:33

going to be down here at my place here in a couple of

55:36

weeks. And I forgot to say something, but I appreciate it. I

55:38

thought Albie was broken, so I'm just glad it got through. Yeah,

55:41

me too. Me too. We did get some

55:43

repeats there, but I printed that out. Yeah,

55:45

sorry. You cleaning them up? No, no, that's

55:47

just more sad. So thank you. He

55:50

writes, though, he's got a Linux truth for us.

55:52

I'm a desktop user. I merely dabble in server

55:54

stuff like self hosting things. I

55:57

got more into it in the last five years or so. I

55:59

gave my Tink or I create and edit all formats

56:01

of media. I install, play, break, and fix things on

56:03

my PCs. And I always have both Windows and Linux

56:05

until just a few years back when I finally wiped

56:08

my last Windows partition for good. But

56:11

why? I game, why would I limit

56:13

myself to just the issues that Linux has, particularly

56:15

around gaming? Because

56:18

Windows has worse. Yeah,

56:21

good, I'm glad to hear this. I

56:23

spend more time fixing stupid little issues

56:25

around my Windows system than I got

56:27

to game. Windows updates will break things,

56:30

driver updates break things. There's a constant

56:32

need for specific DirectX versions. Oh yeah,

56:34

that's ringing a bell. Flashbacks. Yeah, that's

56:36

all ringing a bell, Jeff. Oh

56:39

man, not to even get into the accessories. Like you need

56:42

the software to run the mice and the keyboards and all

56:44

the different headsets. All of them which run in the background.

56:47

All of them want to automatically update. And

56:49

then, God forbid, your Windows install goes for

56:52

more than six months. Yes,

56:55

we do have issues in Linux land, but pick

56:57

your poison, nothing just works. Yeah,

56:59

yeah. Even on the

57:01

supported platforms, it just makes me laugh when my

57:03

Windows gamer buddies tell me they're reinstalling Windows again.

57:06

And it's just normal in every six month routine. My

57:09

last main system was running the same install

57:11

of Arch for over 10 years, even outlasting

57:13

its own hardware support on Windows, of course.

57:16

My biggest caveat is I don't do

57:18

professional work on it of any kind. My desktop is just a

57:20

toy where I can learn, create, and have fun. But

57:23

for me, I find it easier and faster

57:25

to manage when new issues arise. Don't

57:28

use Linux if you value your time. I use

57:30

Linux because I value my

57:32

time. That's a, Jeff,

57:34

that should be a sticker. That right

57:36

there, or the front of a shirt. I don't use Linux if

57:38

you value your time. Could be the front of the shirt. And

57:40

everybody would be like, what? How dare

57:43

you say that? Like we bring these to scale, and the back

57:45

says, I

57:47

use Linux because, all in uppercase, I

57:49

value my time. Right? Beautiful.

57:51

It's a shirt, Jeff. Damn straight. I'm telling

57:54

you what. Coming to an Etsy, still a knee, either. Jeff

57:58

goes on writing us about. his

58:00

Nextcloud setup. As for

58:02

Nextcloud and NixOS, one of my

58:04

Nextcloud instances is a NixOS Viya

58:06

Docker Compose with an

58:08

image found on Nextcloud's GitHub, Nextcloud 28. I

58:10

found I

58:13

needed to fix the cron thing that the

58:16

cloud cries about in a very Nix way.

58:18

I had to make a systemd timer that

58:20

lives in the configuration.nix file, which runs a

58:22

simple bash script. All the

58:24

script does is run Nextcloud's cron.php Viya

58:26

Docker exec. I think that's exactly the

58:29

kind of thing where sometimes

58:31

in setups you're going to need to figure out

58:33

how Nix does that. But once you do, there's

58:36

a lot of times where something similar to

58:38

that kind of just bridges what NixOS already

58:40

has and can fill the solution in for you.

58:42

And honestly, it's kind of, it's well, it might

58:44

not be what you were originally going to do.

58:47

It's not a bad practice. And again,

58:50

it's just going to work forever now. It

58:52

really threw me for a loop, because I've

58:54

never used systemd timers either. So looking up

58:57

how to use systemd timers the normal way,

58:59

not the Nix way, you know, it's just

59:01

that extra little bit of learning on a

59:03

system that's not very well documented. Yes, the

59:06

dude abides boost in with 50,000 sets. I

59:09

hoard that with your kind cover.

59:12

Been running Nextcloud on a scale way

59:14

VPS running docu since 2018. Hey,

59:18

all right. Yeah, docu's got pre and

59:20

post deployment hooks. So you can run

59:22

the OCC commands without even entering the

59:25

container. I'm using a simple

59:27

Docker file based on the official Apache one.

59:29

I run into some upgrade issues through the years,

59:32

but it's pretty solid in general. And

59:34

now you reminded me, I have to upgrade.

59:37

Well, at least you got those pre and post hooks. Yeah,

59:40

sounds like a nice setup. Yeah, you can throw a

59:42

little backup in there too on the pre one, you

59:44

know, we'll do a little database dump there and boom,

59:46

boom, boom, little database are saying

59:49

boom, as Richard boosted in with 40

59:51

6038 Satoshi's. I'd like to see how to set

59:55

up a Minecraft server in the will

59:58

it Nick segment, mostly because I'm going

1:00:00

to be doing that very soon. Feels like

1:00:02

it's probably also a good place to get

1:00:04

into self-hosting. Just a reminder,

1:00:07

this is a ZIP code boost, and PS,

1:00:09

when does Nix get its own theme music

1:00:11

like Rust? Okay,

1:00:14

that's a fair question. It better be something you like to hear a

1:00:16

lot on. I'm sure

1:00:18

you'll. Sorry everybody. Okay,

1:00:21

so four, four, six,

1:00:23

zero, three, eight. Postal

1:00:26

code in Hamilton County, Indiana. Oh

1:00:29

yeah. With cities like Fishers. Yeah,

1:00:31

my town. And Noblesville. Yeah, hello

1:00:34

Fishers and Noblesville, Indiana. Thank you

1:00:36

for boosting in there. Boy,

1:00:40

if you have any suggestions for Nix, seeing that you

1:00:42

weren't mine here in a couple of times, go

1:00:45

ahead, send those in. Well,

1:00:48

we talked a lot about modules

1:00:50

today for Nextcloud. You'll be delighted

1:00:52

to know that Minecraft Server is

1:00:54

also available as a Nix OS

1:00:56

module. Perfect. Maybe we should give

1:00:58

it a go. Actually,

1:01:00

it would be kind of nice because I'm running

1:01:02

one right now via Linnodes deployment script, which is

1:01:04

great, but they're using like some sort of management

1:01:06

program, which I bet you I just

1:01:08

could get away with not having that if I did it

1:01:10

that way. Worth a shot. Maybe. Martin

1:01:13

de Barre comes in, but de Barre comes in

1:01:15

with 44,444 stats. I

1:01:18

think it's a bunch of mik ducks to tell you

1:01:20

the truth. He's out looking up for all the duck.

1:01:22

And he writes, hey guys, coming in

1:01:24

from Podverse, my uncomfortable truth about using Linux

1:01:27

are related to gaming and office work. Some

1:01:30

game launchers, Ubisoft, are not

1:01:32

working, so I dual boot to play

1:01:34

Assassin's Creed games. The Tomb Raider series works just

1:01:36

fine, though. Best workaround is

1:01:38

a cheap Windows 10 license key for office

1:01:40

work, LibreOffice to Office 365 conversions. Never

1:01:44

quite perfect. So I export everything

1:01:46

to PDF instead of sending ODF docs. Those

1:01:48

are my truths. Best regards. It's

1:01:50

true. I do feel that. I do

1:01:53

feel that. Also, he sends some thoughts

1:01:55

with the Podverse team feeling the burnout recently.

1:01:57

He says, I'd try maybe an AI chat

1:01:59

bot. to some bugs and feature requests

1:02:02

and then try to find a community member to do the

1:02:04

triaging and add those items to a backlog. But

1:02:06

don't over commit to the responsiveness. His focus really should

1:02:08

be on the business side, so making money

1:02:10

from memberships, expanding the podcast market, and

1:02:13

prioritizing feature development.

1:02:15

Good insights, Martin. And a copy of that

1:02:18

boost got sent to Mitch over there, so

1:02:20

he is receiving those. Hybrid sarcasm

1:02:23

boost sent with $42,000. In

1:02:26

the heat of battle during this production

1:02:28

cutover weekend, it's great to have the JB crew

1:02:30

in my ears. Oh, I want to know what's going

1:02:32

on. Yeah, good luck. Hope everything goes smoothly. Yeah, I

1:02:34

know what you mean. Thanks for the boost. Those weekends,

1:02:37

you know, they're like, you want to

1:02:39

just crash, but then of course you have to be

1:02:42

around on Monday because that's when all the users are

1:02:44

going to be there. And then God

1:02:46

forbid things don't go well. We'll be

1:02:48

thinking about you, Hybrid. Good luck. Nick

1:02:50

ZP came in with 25,000 sats

1:02:53

using Castamatic. Thank

1:02:55

you for the discussions and criticisms of gaming

1:02:57

on Linux. It's my only sticking point, so

1:02:59

I use Windows for gaming only. The area

1:03:02

that broke the camel's back was

1:03:05

Mirror's Edge. I had the very

1:03:07

same monitor troubles that Chris talked

1:03:09

about with CSGO, so my partnering

1:03:11

had to be left turns only.

1:03:14

I challenge anyone to treat Mirror's Edge

1:03:16

like a NASCAR race. Not fun. No,

1:03:18

it does not sound fun. Yeah,

1:03:22

you know, and it's like, I'm sure it went through

1:03:24

your head at the time, except it's like, well, I

1:03:27

guess I could just unplug all my

1:03:29

monitors. But that's so embarrassing. Like, hold

1:03:31

on, son. Dad's going to go

1:03:33

unplug your screen so he can play. I

1:03:36

just bailed, to tell you the truth. But I

1:03:38

wonder if it crossed your mind. Thank you very much for that boost.

1:03:41

Purple Dog came in with 5,000 Satteroos,

1:03:43

and he says I was running the

1:03:45

Linux server I.O. container for years, talking

1:03:48

about Next Club. But then

1:03:50

I switched to Nix. My server's config is a

1:03:52

flake up. There it is. Then I've got my

1:03:54

own Next Club module based on someone else's that

1:03:56

I found on GitHub. But

1:03:59

what he likes? about is it includes a

1:04:01

collaborate container for Office. It's been working great. In

1:04:03

general, I've been trying to move out of Docker

1:04:05

and onto just using Nix, so it's

1:04:08

one less thing to maintain. Image is the

1:04:10

last one left. Maybe you could do

1:04:12

that in a willed Nix. Oh,

1:04:14

yeah. That'd be a great shout, particularly given the

1:04:17

news of Image this week of their switching from

1:04:19

MIT to a GPL-based license. Hmm, that's

1:04:22

a really good contender. I'm going to

1:04:24

write that down, guys, because, yeah, yeah,

1:04:27

that's a very complicated set of containers, too.

1:04:29

There's a lot going on with Image. It

1:04:31

changes a lot. There's a lot of breaking

1:04:33

changes with each Image release at the moment.

1:04:36

Yeah, you could see why maybe you'd want to have a

1:04:38

little more control over that. Hmm, hmm.

1:04:41

Good suggestion, Purple Dog. Thank you for the boost, too. An

1:04:44

anonymous fountain user boost in with $10,000, just

1:04:47

to say. Happy to hear you went podcasting 2.0.

1:04:49

Oh, yeah, yeah. The feed's going to be getting,

1:04:52

the main feed will be getting updated soon, and

1:04:54

the members' feed has already been upgraded. We're looking

1:04:56

for feedback there. Let us

1:04:58

know if you're a member with a new podcasting app.

1:05:00

And then, of course, grab one because scale's coming up,

1:05:02

and our stream's going to be live in the podcast

1:05:04

app. One of the features of podcasting 2.0 that

1:05:06

I am very excited about is

1:05:08

live streams can be in your podcast app. Like, they

1:05:10

always should have been. We've never should have been sending

1:05:13

you to YouTube and Twitch to listen to a podcast.

1:05:15

And now you can just be subscribed to the Linux unplug

1:05:17

feed. And when it's a 2.0 feed,

1:05:20

when we go live, it'll just show up in your

1:05:22

feed like another entry. And you can just

1:05:24

tap and listen live, like a regular old episode. It's

1:05:27

such a beautiful thing. BAMham182 sent in

1:05:29

two boosts for a total of 6,666 sets. I have

1:05:31

all of my services, including

1:05:36

Nextcloud, set up as Nix

1:05:38

configured system D services, which

1:05:40

fire Podman compose commands. Not

1:05:43

the most Nix way, but it is how

1:05:45

I've been managing containers for a long time,

1:05:48

and it still works just fine. So I

1:05:50

don't see a reason to switch it. I

1:05:52

also don't need to wait for Nix packages

1:05:54

to get the newest updates. I just system

1:05:56

CTL restart if I didn't pin the version.

1:05:59

I forgot to mention... I'm using the official

1:06:02

unofficial container in the next cloud slash

1:06:04

next cloud namespace on Docker

1:06:06

Hub Even though that means the health page

1:06:09

nags me about the totally unnecessary BZ

1:06:12

not being installed Yeah,

1:06:14

fair. Um, I like a

1:06:16

lot that we got a pod man Shot out in

1:06:19

here because that's absolutely a valid way that we didn't

1:06:21

even touch on this episode It's not because we don't

1:06:23

know about it It's just not one that any of

1:06:25

us really have a lot of experience with so Getting

1:06:28

some experience from the audience with using pod man is really good

1:06:30

to hear. He says like a nice setup BAM

1:06:32

ham appreciate that Thank you.

1:06:34

Talk criminal. He's coming in with

1:06:37

a row of dukes You

1:06:39

nixpilled me. Thanks Saturday

1:06:42

fedora also got a row of ducks and

1:06:45

he says keep it up Mitch. You're doing great Agreed.

1:06:48

Thank you. Thank you to the podverse team out there.

1:06:50

How was right comes in with? 4200

1:06:52

cents I Was

1:06:54

curious about running a monitoring service on Nick's

1:06:56

OS I tried Nagios, but I failed to

1:06:58

get the web interface working on a local

1:07:00

host. So I'm trying Gruff up. No How

1:07:04

about Ln bits with cashew or fetiment

1:07:06

on Nick's OS? That would be

1:07:08

really great to know how to do but

1:07:10

the software is mostly experimental Oh

1:07:13

and then a follow-up boost another 2100 here

1:07:15

Mitch could apply for an open source Contribution

1:07:18

from open sats or geyser dot fund. Maybe

1:07:20

that makes some sense for all parties That

1:07:23

is a really good suggestion Especially

1:07:25

the open sats one. I'm

1:07:27

gonna pass that on personally to Mitch. How thank you Yeah,

1:07:30

a monitoring package great idea. We are maybe cooking

1:07:32

something up with the Ln bits direction We may

1:07:34

have more on that in the future But

1:07:37

I'm writing that down. I think an update

1:07:40

is well supported on Nick's OS right? Oh,

1:07:42

yeah. Oh, yeah. I Have

1:07:44

it running on my home one already. So,

1:07:46

you know, I really should just bite the

1:07:49

bullet and Really just

1:07:51

get into their whole learning mechanism. I really should

1:07:53

because I got it running on two different systems

1:07:55

already zakutak comes in

1:07:57

with 13,000 three hundred and fifty

1:07:59

six That's using the index send them

1:08:01

from the podcast index and

1:08:04

I really appreciate this boost from

1:08:06

Zach because We talked very positively about

1:08:08

nicks and I think it's good to also show the other side

1:08:10

of the coin He gave it a go and

1:08:12

he writes I tried nicks for a week and it wasn't for

1:08:14

me. I just got way

1:08:16

too many timers Okay,

1:08:19

I may try snowflake OS when it gets more

1:08:21

stable I ended up wandering again, and I'm trying

1:08:23

pop OS for the first time I do like

1:08:25

how they've tweaked them and for someone who wants

1:08:27

tiling. It's great I'm nervous that their

1:08:30

gooey won't be as smooth to get

1:08:32

going But we'll see and

1:08:34

an uncomfortable Linux truth for me. You'll always

1:08:36

need Windows installed on the rails somewhere I

1:08:39

can get away with a VM for a bit,

1:08:41

but things like certification testing. Well, they just really

1:08:44

hate VMS. I Hadn't

1:08:46

thought about training. I've been very

1:08:48

fortunate that I haven't really had to deal

1:08:50

with that kind of stuff He says

1:08:52

I forgot to mention how I nix cloud. I'm

1:08:54

using the all-in-one Docker image on rocky Linux It's

1:08:56

got all the tools I need has been really

1:08:58

solid and I'm trying out podverse again because

1:09:00

why not if nix won't work for Me I'll

1:09:03

make a podcasting 2.0 app work That's

1:09:06

the spirit, thank you, sir. Really appreciate that Zach

1:09:09

I think that's fair to you know If you're

1:09:11

just trying to get some stuff done Nix will

1:09:13

probably get in your way for a while because

1:09:15

you kind of have to bend yourself to it

1:09:17

to get anything done Yeah, don't pull a Linus

1:09:19

and I mean Linus from YouTube and try to

1:09:21

switch when you're super busy So you don't have

1:09:23

the time to deal with the nuances and you

1:09:26

just hit yes continue continue. I'm busy I don't

1:09:28

got time for this. That's not the

1:09:30

right time to make a switch. It's just not gene

1:09:32

being boost in with 4444

1:09:34

cents I'm

1:09:36

just about to migrate next cloud to nix

1:09:39

and was considering doing the built-in way But

1:09:41

will absolutely be waiting until I hear the

1:09:43

results of this quest and

1:09:46

a follow-up Row a ducks

1:09:48

here if only there was cart play

1:09:50

support in audiobook shell Yes,

1:09:53

but in car play you can have pretty

1:09:56

easy control over just the audio that's playing

1:09:58

so it's The canyon is

1:10:00

still kind of manageable. If we don't

1:10:02

I O s though that means is gonna access to

1:10:05

Prologue which I'm pretty sure does work on car play

1:10:07

assuming he's got a plex. Backend: Of

1:10:09

course. The. I think there would be great

1:10:11

is that the Native Audiobooks health amp actually got.

1:10:14

Car. Play support but it doesn't even have a

1:10:16

citizenry. I o s toward a decent has fly

1:10:18

and a hard to get and right. Yeah.

1:10:21

Lunar. Night boosted and ten thousand Sats

1:10:23

from Fountain. Gonna second the Twenty thirteen.

1:10:25

Mack Trash can love that thing. As

1:10:27

such a monster, Sit is at my

1:10:29

old work. We had one on the

1:10:32

test track that ran linux and it's

1:10:34

friggin' chew through everything, even all the

1:10:36

way through to Twenty Twenty. By far

1:10:38

the Bcs machine we ever had and

1:10:41

it never had any stability or hardware

1:10:43

issues. Isn't. A funny how.

1:10:46

When. These computers come out you don't really, and

1:10:48

we just don't imagine it ends up like this.

1:10:51

And it's such a weird one off from

1:10:53

Apple that are never going to make again.

1:10:55

And. Is I make an ex Any six

1:10:58

machines that hear this thing is was the

1:11:00

and an easy see ram and silent operation

1:11:02

wilde just. Just. Yeah, it's a heck

1:11:04

of a machine is it has Very impressed. Jordan

1:11:06

Bravo comes in with thirteen thousand, three

1:11:08

hundred and forty five sets. The regarding

1:11:10

next had several next. Oh, as I'd

1:11:12

like to learn how others are doing

1:11:15

automated backups. I'm. In the process setting

1:11:17

this up with board back up to run in the

1:11:19

middle of the night. but my questions are this number

1:11:21

one. What? Exactly to back

1:11:23

up and which directories and number two does

1:11:25

the database need to be locked while back

1:11:27

in up? If. So. How.

1:11:30

Now. That is a very good couple of

1:11:32

questions or to take the top one their.

1:11:35

The. Way I backup is a backup the underlying sales

1:11:37

that on the file system for my next thought

1:11:40

instance and then I back of the database each

1:11:42

night. And what I really

1:11:44

like about our module is the

1:11:46

module. Use. Of assistant, a timer.

1:11:49

To actually properly do a postscript database dump

1:11:51

and drop that on your file system. And.

1:11:54

Then you could just grab that with anything. a

1:11:56

file system snapshot are seeing. A duplicate.

1:11:58

He. Whatever. It. The The Next Car

1:12:00

project has some.switch we can link about backup or

1:12:03

they say you need a backup. The convict folder.

1:12:05

The data folder. The same folder. you know if

1:12:07

that theme is important to you. and yeah, of

1:12:09

course that database. And then they do mention that

1:12:12

there's a maintenance mode that you can. The. Do.

1:12:14

With Lox sessions have logged in users and

1:12:16

prevents new loggins. We want to just like

1:12:18

freeze things with the application player. At.

1:12:20

Us like and oh C C Command. So

1:12:23

if if the you did want to go that route

1:12:25

you can put an end maintenance mode of and do

1:12:27

whatever backup stuff you need. Butter. That

1:12:29

are little at ways to do it. Yeah, I.

1:12:31

Like I like what we're doing, but I'd love to

1:12:33

get more input because I think there's that's always something

1:12:35

you wanna get right? In.

1:12:38

The early next president has summer sample a

1:12:40

database dumped commands to a little to look

1:12:42

at. You come and join me in the

1:12:44

is at a fresh coat and are doing

1:12:47

at data on the set of a snapshots

1:12:49

and that does. He. Have

1:12:51

to worry about stopping. Maria de Be

1:12:53

Agree and maintenance mode her tonight. That

1:12:55

just. With. The magic of said if s

1:12:57

really. No. One ever got fired

1:12:59

for deploying Cfs. in the he said

1:13:01

that they differ things that instead of

1:13:04

the other feedback I get on my

1:13:06

eyes. And. My pool brain

1:13:08

so confused stands out stands out and when we

1:13:10

to cup or linux teamster comes in with ten

1:13:12

case as tells about his next card set up

1:13:14

using Linux server ios and up above any had

1:13:16

to give ago he had to take a go

1:13:19

at me with i appreciate he says have never

1:13:21

had a bad upgrade with next cloud. But.

1:13:23

That might be cut be because I've never

1:13:25

used sequel light only Maria De B C.

1:13:27

Unlike Chris I know how to follow direct

1:13:29

share this is the first anyways about a

1:13:31

year ago migrated to I Like see container

1:13:34

I'm approximates mainly because they wanted to keep

1:13:36

tweaking way the docker container didn't always like

1:13:38

love ya professor at to send you to

1:13:40

the burn unit Chris offer out why now

1:13:42

teams are you got me proceed the boost

1:13:44

ambient noise came with three thousand Nine Hundred

1:13:46

and thirty south's as a tried next O

1:13:49

s on a new mb me as get

1:13:51

install anyways. And. Will stable channel but

1:13:53

concern little buddy on my Pc it

1:13:55

refuses to suspend. I. Found out it's

1:13:57

a know next o S issue with gigabyte motherboards

1:13:59

and. NVMe drives. I apply

1:14:01

the fix, the PC suspends. Now most

1:14:04

of my other issues were either fixed or

1:14:06

unstable, so I changed to the unstable branch

1:14:08

and now everything's working and it's fast.

1:14:11

However now my computer tries multiple times to

1:14:13

re-suspend itself when I wake it up. Thanks

1:14:16

for recommending a moody teenager as an OS. I

1:14:20

think you have a moody kernel, right? Because that's probably

1:14:22

where most that's coming from is the kernel layer. I

1:14:25

wonder if you could actually,

1:14:28

just for fun, I mean this is

1:14:30

just clearly for fun noise, but

1:14:32

if you could roll back to the different

1:14:34

states of breakage with the rollbacks just to

1:14:36

see what that's like. Could you go back

1:14:39

a couple of snapshots and go to a

1:14:41

different broken state? That's fun, right?

1:14:44

That's fun. With Nix you could

1:14:46

actually end up going down different, you could

1:14:48

have multiple broken paths to pick from. Totally.

1:14:50

Not a fun idea. Anonymous

1:14:53

boosts in with 12,345 sets. So the combination is one, two,

1:14:55

three, four, five. That's

1:15:02

the stupidest combination I've ever heard in my

1:15:05

life. I'm currently running my Nix Cloud instance

1:15:07

in an un-raid app, but I'm

1:15:09

considering moving my apps to a Nix

1:15:11

VM. Maybe

1:15:13

this episode will help change your mind. With it? Yeah,

1:15:15

let us know. And I'd also like to know Anonymous

1:15:18

why. Unraid was one

1:15:20

that I did see in our Matrix chat room, brought up a couple

1:15:22

of times. It seemed like people were pretty happy with it, so I'd

1:15:24

be curious to know your feedback. So if

1:15:26

you're running it as an un-raid app or a container,

1:15:28

something like that, there'll be

1:15:30

some upstream Docker image somewhere that you're running.

1:15:32

Let us know what that is. Well

1:15:35

Fuzzy Mistborn comes in with one,

1:15:37

two, three, four, five satoshis.

1:15:39

One, two, three, four, five. Yes! That's

1:15:42

amazing! I've got the same combination on

1:15:44

my luggage! He

1:16:00

believes in the steam deck. I tried

1:16:02

all kinds of solutions to switch resolution

1:16:04

use dummy a H D M I

1:16:07

plugs tried X Eleven vs. Whalen, etc.

1:16:09

Nothing weren't. Installed. Windows Found

1:16:11

a plugin for Sunshine that auto switches

1:16:13

resolutions and it just worked. Said to

1:16:15

not use Linux regaining, as I've generally

1:16:18

been happy with it, but in this

1:16:20

instance Windows was just easier. He

1:16:22

I understand miss Born at. My kind of

1:16:24

compromise was when I just couldn't get to

1:16:26

work. Just. Went over to G

1:16:28

Force now streaming I got one of those old

1:16:30

grandfathered in subscriptions to say. I. Pay

1:16:33

like a semi reasonable price. I don't

1:16:35

want to cancel it and I'm. I'm

1:16:37

always doing the math on it but then so it is

1:16:39

nice. I. Do have those problems to have

1:16:42

the streaming service. It. It

1:16:44

has it's own unique issues. special

1:16:46

Starling but it's Eric Centers a

1:16:48

bunch of feedback on I his.

1:16:51

Next. Cloud our knicks journey and one of the

1:16:53

most amazing things was it is trying to get

1:16:55

Calabro working and he couldn't figure out why can

1:16:57

get at where gain it turned out you know

1:16:59

years having connection issues. all of that. As.

1:17:02

He went on though he decided to go

1:17:04

what he calls hashtag next Hundred Full has

1:17:06

stagnant Canada and package collaborat Ford. Next he

1:17:08

says it builds a runtime a little towns

1:17:10

him because clever expected to run of each

1:17:12

process in his own to root incitement. ah

1:17:14

but if anyone wants to contribute of data

1:17:16

link and on will put that the notes

1:17:18

and hit a wall for now. But maybe

1:17:20

we can continue in the future. but a

1:17:22

fun challenge when he had a package for

1:17:24

next. Next up. Sterling. Pts. It's

1:17:27

already packed up. But. It's not

1:17:29

committed. It. Just needs that service module

1:17:31

now that a lot to gray apps.

1:17:33

damning Pdf we taught by itself hosted

1:17:35

recently is just be mvp of Pdf

1:17:37

life if you've never tried it. Will.

1:17:40

Definitely underscore that I've really enjoyed having. That's

1:17:43

the day every they the we did that

1:17:45

we require that episode of sulfur I installed

1:17:47

that might and I've used it several times

1:17:49

since. Them. Don't need Adobe

1:17:51

anymore. Samurai. Yeti Bousson

1:17:53

with a rubber ducks. i'd

1:17:57

love to hear her you guys would set up

1:17:59

a coding environment using NixOS and say, tail

1:18:01

scale. I've not had

1:18:03

the time to dig into it myself, but

1:18:05

having access to a self-hosted, open source cloud

1:18:08

development environment anywhere on any device that supports

1:18:10

a modern web browser, all while

1:18:12

enjoying the stability that an immutable operating system

1:18:14

could offer, well, that sounds like

1:18:16

an absolute game changer. I will add,

1:18:19

it becomes even more important when you start thinking

1:18:21

about different types of devices you might use in

1:18:23

the future. You know, I'm not

1:18:25

Mr. Web App Guy, but I've been

1:18:27

really grateful to have some web apps when I've been using

1:18:29

the native VR apps to do work. So

1:18:31

you never know what future device you might be using.

1:18:34

As far as putting like, you know,

1:18:36

coding set up on tail scale and

1:18:39

NixOS, that's a good topic. I can see I

1:18:41

was taking them on self-hosted too. That could be one

1:18:43

we could chew on. Noodles has a little

1:18:45

advice for Nix beginners with 5,000 SATs.

1:18:49

A while ago, you were talking about trying

1:18:51

out Nix and many people were worried about

1:18:53

moving their main system to NixOS. Well, try

1:18:55

it in a VM. The beauty of

1:18:58

Nix and NixOS is that configuration.nix file

1:19:00

that can be copied almost verbatim to

1:19:02

your next machine and it will work

1:19:04

exactly the same. If you can

1:19:06

get it working in a VM, chances are very

1:19:08

high of transferring that to your main system

1:19:11

and having it work there too. That reproducibility

1:19:13

is amazing, especially in disaster recovery situations. And

1:19:15

to have a VM environment where you can

1:19:17

tweak things, not worry about it and be,

1:19:19

you know, a quick little reboot is also

1:19:21

very handy. And it's a great way to

1:19:23

try things. Even like the

1:19:25

new version of genome or plasma. Genome,

1:19:28

sorry. Thank you Noodles, appreciate

1:19:30

that. Evil Emperor Zurg comes in with a

1:19:32

row of dicks. And I like this one.

1:19:34

This I'm gonna write down too. I'm taking

1:19:36

some notes this segment. Have you ever done

1:19:38

a RAM only

1:19:41

distro challenge before? I looked and

1:19:44

there is an overlap with some of these 32-bit distros,

1:19:46

but you know, this could be a fun distro plus

1:19:48

hardware challenge. Maybe the winner would be the one getting

1:19:51

the most performance with the least amount of RAM. A

1:19:54

RAM only OS for a week. This

1:19:57

could be a fun one. Yeah. If Anybody

1:19:59

would like to. Underscore: this one or cove

1:20:01

sponsored this idea. boost in an

1:20:03

as a ram only. Humans.

1:20:07

Give. Us some ideas. I called up and put

1:20:09

down the list as a contender. I immediately thought

1:20:11

of a few things we could try. And.

1:20:14

Us are thinking which which says which one of

1:20:16

my systems has for hims yeah think I'm gonna

1:20:18

run Hannah Montana Let us hope your height and

1:20:20

co op you know the have to install is

1:20:22

that that to ram flag on the yacht in

1:20:24

the ram a fast We just have to all

1:20:26

live there the whole week. There might also be

1:20:28

a way to do it I don't recall exactly

1:20:30

so they don't use a very often but isn't

1:20:32

their way to do that with then toy to

1:20:34

was just about almost any iso. See.

1:20:36

Could just about make anything a rammell

1:20:38

environment with Enjoy! Mazer Bousson

1:20:40

with or twelve thousand, Three

1:20:42

Hundred and Forty Five said

1:20:44

All right. So ludicrous. Free.

1:20:47

I. Just use the built in this O

1:20:49

S Next cloud service and the. They.

1:20:51

Also sent us a short right up which

1:20:53

is nice. Can a details of the things

1:20:56

they're using job very similar to what we're

1:20:58

doing that my sequels that post press have

1:21:00

also done. Some custom daddy worker not using

1:21:02

Under Next and had to come up with

1:21:04

the right matching rules to his caddie Butter

1:21:06

Got a little working in the. Only.

1:21:08

To the consensus. Some. To second after

1:21:10

the show I included the receipts and willing to

1:21:13

thank you. That's not only is a great to

1:21:15

does have that for something we can kind of

1:21:17

compared to what we just a mazer but every

1:21:19

appreciate the ability to share that with everybody. To.

1:21:22

Max. Power came in with a

1:21:24

space balls boost. Hell was that

1:21:26

space for one Cd. Sales

1:21:30

on a bone to pick with us.

1:21:33

Your love for next breaks down might

1:21:35

destroy hopping defenses and I end up

1:21:37

installing it. Happened twice now. And.

1:21:39

Love Next Shell and the ability to try

1:21:41

out software but I ran into trouble with

1:21:44

the documentation when I start reading about flakes

1:21:46

my as is kind of glaze over and

1:21:48

I go download the arch I so there

1:21:50

were gives us to good and it's this

1:21:53

old man's comfort zone. insert. Old

1:21:55

man yells at cloud just discuss

1:21:57

would have guessed. That

1:21:59

way. Audiobookshelf plus libation is amazing. So

1:22:01

thank you. Yeah plus one to that

1:22:04

libation is amazing Appreciate

1:22:06

everybody sitting that in I think the

1:22:08

arch wiki still is the gold standard in distro

1:22:10

documentation I feel like the Knicks wiki is pretty

1:22:12

good for a lot of things and it kind

1:22:14

of gets a bad rap But how

1:22:16

do you hold? How do you

1:22:18

hold a comparison to to the King right?

1:22:21

I I agree and also I get

1:22:23

the idea of like Arch has

1:22:25

this brilliant simplicity to it that

1:22:27

Nick's Well, Nick's has a simplicity

1:22:30

to it, but it's a different kind of thing. It's a

1:22:32

it's not once you get flakes involved,

1:22:34

right? Yeah, it's simplicity through complexity But

1:22:38

it's just so different right like arches. Everything's

1:22:40

just a pretty much thin wrapper on upstream

1:22:42

So it's like yeah, it is just the

1:22:45

standard components the freshest versions of them you

1:22:47

good luck. Here's how they work The Thing

1:22:51

is about a flake in particular I I'm with

1:22:53

you that eyes glaze over and there was a

1:22:55

there was a weekend or or a week You

1:22:57

know last late last year where I finally fought

1:22:59

through the glaze and figured it out If

1:23:02

you are a software developer by trade And

1:23:04

I think this is where's probably why you

1:23:07

find it so easy compared to the rest

1:23:09

of us Your brain already works in the

1:23:11

way of flakes like inputs outputs, you know

1:23:13

You're dealing with passing data around and and

1:23:15

all that kind of stuff inheritance type stuff

1:23:17

If you're not a developer and you don't write software

1:23:20

for a living you You're

1:23:22

gonna have to learn some of that stuff for flakes to make

1:23:24

sense to you And I think

1:23:26

the issue with the flake documentation is

1:23:28

it's written by the people that understand

1:23:30

it who are Developers for the most

1:23:32

part and so there isn't really that

1:23:34

kind of easy jumping off point yet

1:23:37

for flakes in particular, so If

1:23:40

if Nick's has anywhere to focus on in my

1:23:42

opinion over the next Year say

1:23:44

it's gonna be that jumping off point for flakes

1:23:46

as they become the kind of less

1:23:49

experimental Arm of the

1:23:51

distro and I think things like flake hub are

1:23:54

gonna help that and and that kind of adoption

1:23:56

I'll put a link. I really like to nicks

1:23:58

from first principles flake it edition by Tony Finn,

1:24:01

which is a sort of series. It is

1:24:03

kind of development oriented. You start thinking about like how

1:24:05

Nix works is if you were going to package something

1:24:07

just like starting with a pure raw Nixling and then

1:24:10

you build up from there. But seeing all the parts

1:24:12

laid out kind of makes things fit together at the

1:24:14

end, I think. Again, I just want

1:24:16

to remark, we seem to be witnessing a phenomenon here. This

1:24:19

is an organic thing that seems to be developing. And yes,

1:24:21

we're talking about it a lot. I appreciate that. But this

1:24:23

has also been the history of the show. When

1:24:26

these things are developing in the community, the show ends

1:24:28

up kind of really honing in on this stuff.

1:24:30

And then what we discover years later is

1:24:33

that we were right in the middle of a

1:24:35

new era being developed. We have this with WireGuard.

1:24:37

We have this with System D. We have this

1:24:39

with Pulse Audio. There has been

1:24:42

these events in the Linux landscape during the

1:24:44

history of the show where

1:24:46

the show becomes very focused for a period of time.

1:24:48

And I've described this before as Linux

1:24:50

Unplugged is the DS9 of Jupiter Broadcasting

1:24:52

podcast, where a storyline will continue across

1:24:54

multiple episodes for a while. And you'll

1:24:56

have a season arc that lasts two

1:24:58

or three seasons. It's not the

1:25:00

show. It's not a defining thing about the show,

1:25:03

but it'll be here for a while. When we

1:25:05

get some historical perspective on all of this, I

1:25:08

think we're going to look back and realize that

1:25:10

we were witnessing a transition take place that

1:25:12

is going to redefine the Linux community. I don't know

1:25:14

the extent of that transition. We're in the middle of

1:25:16

it right now, so I couldn't even measure it for

1:25:18

you. But I can feel something's happening. And

1:25:21

I think that's part of the phenomena as to why it

1:25:23

comes up so often on this show. But

1:25:25

thank you very much, Max, for that boost. Appreciate

1:25:27

it. Todd comes in from North Virginia with 11,101

1:25:30

sats and says, Happy

1:25:33

Sunday. Anonymous user 42 from

1:25:35

Fountain sent us two boosts,

1:25:37

3,000 sats total. Kind

1:25:39

of wrapping up here and says, I started listening

1:25:41

to the show a month ago, and I just

1:25:44

moved my ARTS install on my laptop from

1:25:46

ButterFS to BcashFS. Okay, now

1:25:49

this guy, user

1:25:51

42, you are an early adopter, sir.

1:25:53

He goes on to say, I added

1:25:55

a USB hard drive as a background

1:25:57

to give myself some more storage. I

1:25:59

moved my Windows into a VM and

1:26:01

I created a Mac OS VM as well.

1:26:03

This is quite the little role. I know

1:26:05

dude And it says I was wondering why

1:26:08

butterfs copy on right is bad for a

1:26:10

VM And if this is an issue

1:26:12

for B cache of s I heard it is

1:26:14

because of fragmentation But would that

1:26:16

be an issue on an SSD? I do use

1:26:18

encryption So turning off copy on right is not

1:26:21

an option for me Now

1:26:23

the issue you really have I believe and boys

1:26:25

correct me if I'm wrong with a copy on

1:26:27

right file system and storing a VM Image there

1:26:29

is the right amplification issue You could potentially have

1:26:31

in that scenario and you could end

1:26:34

up with a lot of extra rights to that disk I

1:26:37

was doing some rough quick reading and research before

1:26:39

the show to try to give you a little

1:26:41

value back because I appreciate that you Sent in

1:26:43

user 42 and what I was seeing

1:26:45

in early benchmarks is it

1:26:47

is a there's a lot more right

1:26:50

overhead Using copy on

1:26:52

right on butterfs in the same exact

1:26:54

setup with B cache FS I

1:26:56

don't have a technical understanding as to why that is

1:26:58

you still have some right amplification on B cache FS

1:27:00

But it seems to be significantly less Suppose

1:27:03

it also be curious You know as I wondered are

1:27:05

you worried for the in the academic

1:27:07

sense or are you having problems with your workloads?

1:27:09

I'd be curious to know how are

1:27:11

they performing for you out of the gate? Yeah

1:27:13

And please please keep us updated on how be

1:27:16

cash FS works out for you good or bad

1:27:18

If you stick with it or not because you're

1:27:20

one of the few listeners I know out in the wild using

1:27:23

it right now Yeah, I know there. I think there are

1:27:25

at least plans to have you know, the no cow path

1:27:27

for B cache FS So you can yeah disable that I

1:27:29

think in a pretty granular level But

1:27:32

I don't know if that's a mature feature as yet.

1:27:35

Yeah your trail really believes her dude. Thank

1:27:37

you for the boost to thank you Everybody that boosted

1:27:39

a huge huge boost segment. Thank you. We enjoyed all

1:27:42

of that very very much It's sometimes

1:27:44

our very favorite part of the show. We had 36 boosters.

1:27:46

We stacked 897

1:27:50

sets Also,

1:27:52

thank you everybody who streams why they listen we see you and

1:27:54

we appreciate you and it's often the highlight of our day When

1:27:57

we open up that dashboard. So thank you very

1:27:59

much for streaming those as you

1:28:01

listen. And thank you everybody who supports the show

1:28:03

either through a boost, through the membership program,

1:28:05

we do have the Unplugged Core, or by

1:28:07

even just recommending the show to somebody or

1:28:09

participating with one of our sponsors offers. That

1:28:12

all means a great deal to us and keeps us

1:28:14

independent. We

1:28:16

have some picks today and

1:28:19

we have some really great ones and I

1:28:21

want to start with one that I have not yet checked out

1:28:23

but both Brent and Wes

1:28:26

put this one in the doc. It's

1:28:28

called Arianne? A-R-I-O-N?

1:28:33

I like it. Yeah. Arianne? Arianne? Yeah sure.

1:28:35

Okay what is this? Wes, what is this?

1:28:38

Tell me and why are we linking it

1:28:40

in today's episode? Well Brent found it. I've

1:28:42

uh but I have played with it before

1:28:44

so. Well I found it in our inbox

1:28:46

because listener Charin, C-A-A,

1:28:49

sent this in and says, hey great

1:28:51

fall black for when you're too lazy

1:28:53

to properly hook something up to run

1:28:56

on Nix. There's a thin wrapper around

1:28:58

Docker Compose that lets you configure your

1:29:00

services in Nix rather than in YAML.

1:29:02

Uses Podband by default and can support

1:29:05

Docker as well. That creates system deservices

1:29:07

for your projects automatically. So I find

1:29:09

myself using it often when some service

1:29:11

I want to spin up requires MySQL

1:29:13

or Postgres and I'm just too lazy

1:29:16

to hook it up natively. So Arianne

1:29:19

is a tool basically that takes

1:29:21

the old ways of Docker Compose. Oh

1:29:23

sorry Alex I said old. Takes

1:29:26

the ways of Docker Compose and just kind

1:29:28

of like quickly Nix-ifies them for you until

1:29:30

you have time to actually do that yourself. Okay it's

1:29:33

like it's a resource to do that for you. Yeah

1:29:35

you can run it just like you would with Docker

1:29:37

Compose where you just have like say a standalone folder

1:29:39

somewhere on the file system where you go run instead

1:29:41

of Docker Compose up it's Arianne up and

1:29:43

it handles it for you and you have like a little

1:29:46

Nix file instead of a YAML file or you can integrate

1:29:48

it a little more deeply like have it set up in

1:29:50

your configuration.Nix to go call Arianne and spin things up. There

1:29:52

you go. I haven't used it a ton because we've been

1:29:54

trying you know just running stuff with Nix for the most

1:29:56

part but when we were first playing with Nix OS

1:29:59

I gave it a go. And I mean, it really

1:30:01

just worked. It has

1:30:03

pretty much all the escape patches you need to go

1:30:05

do any sort of custom stuff that you might be

1:30:07

trying to do in the underlying Docker Compose. If you

1:30:09

have some fancy network features or volume stuff that you

1:30:11

need to set up, seem to support pretty

1:30:14

much all of it. I at least tried moving some of

1:30:16

our existing thing, example Docker

1:30:18

Compose is out in the field. They all

1:30:20

worked. That sounds really handy.

1:30:22

I always like to jump straight to the FAQ

1:30:24

and a project like this. And there's one thing

1:30:26

here that just gives me just a hint of

1:30:28

pause. The

1:30:31

question they ask is, what is messing with

1:30:33

my environment variables? And obviously

1:30:35

Docker Compose performs its own

1:30:38

environmental variable substitution. And

1:30:40

so there are a couple of little gotchas with

1:30:42

the syntax there, for

1:30:45

a straight lift and shift. That's

1:30:48

a good rule, Thumb. Check out the

1:30:50

FAQ. I like that. It's

1:30:54

like the next, it's the modern version, Alex, of RTFM.

1:30:57

It's RTFF. Well, you're relying on the project

1:31:00

to be honest about their own shortcomings, but. Yeah,

1:31:02

that's true. Before we get to

1:31:04

the end of the show, I have a big question for all of

1:31:06

you guys. Did it nix? Oh,

1:31:09

it nixed. I think it nixed

1:31:11

well. I think if I would have had just

1:31:13

a couple extra hours this weekend, I would have

1:31:16

ripped out my existing install and nixed it. And

1:31:18

I think we're gonna transition to

1:31:20

this in production too, for the JB stuff. It's

1:31:22

just a matter of kind of timing that. But

1:31:25

what about you? Did it nix for you, Brent? I mean, you

1:31:27

got the snap. You might have just as a pending

1:31:29

need to switch as I do. Now, I

1:31:31

know we are running this on like a

1:31:34

bit of a beast of a computer,

1:31:36

but just seeing the performance differences, like

1:31:38

basically one end of the spectrum to

1:31:40

the other, it gave

1:31:42

me the motivation I think I need to switch this over.

1:31:44

So I think it nixed for me. Yeah.

1:31:47

Wes, did it nix for you? Oh, yeah. Yeah?

1:31:50

Yeah? Good. I'm

1:31:52

glad to hear that. And Alex, what are your takeaways? Nixed and

1:31:54

blended. Very

1:31:56

good. I'd like to know what the audience thinks. Also,

1:31:58

in the back. of our minds, we're always

1:32:01

worried that we're going too deep with this stuff. We

1:32:03

may be getting too technical, going on too much about

1:32:05

modules and nicks and whatnot. That's always something we're pretty

1:32:07

sensitive to. So give us your feedback on

1:32:09

that, either by boosting in, which you can send us

1:32:11

a little value to do, or by going to linuxunplug.com

1:32:14

slash contact. Also don't forget, we

1:32:16

want to hear about if you

1:32:18

even want me to get into what it was

1:32:20

like getting desktop Linux into VR and

1:32:23

how you have solved a tiered storage setup or

1:32:25

how you would build it if you

1:32:27

were going to solve it. Next week is episode 550 too. Nothing really

1:32:33

planned yet. I think we're going to take the week off. See

1:32:36

you next week. Same bad time, same

1:32:38

bad steam. Well that guy did it.

1:32:40

Yeah, we should have updated the recording. What were we thinking?

1:32:42

Now we gotta do it. Okay. Alright. We'll

1:32:44

figure it out. No, I kid. I

1:32:46

really just appreciate you making it. Of course we'll be live

1:32:49

if you'd like to join us for 550. We do it

1:32:51

at noon Pacific 3pm

1:32:53

Eastern over at jblive.tv. You

1:32:56

can also subscribe in the Jupiter station. It's a lit feed for

1:32:58

now and it'll be coming to the main feed very soon. Links

1:33:01

to what we talked about,

1:33:03

linuxunplug.com slash, well you probably guessed it,

1:33:05

554. Right? Is that what

1:33:08

it is? Or 49? You didn't guess it because I got

1:33:10

it wrong. I

1:33:13

probably need a nap now. You see, Brent is here

1:33:15

in the pre-show confusing everyone. It's Brent's fault. See you

1:33:17

next Tuesday.

1:34:00

Thank you.