esc

I think I’ve just never worked in an environment like this it was always either a startup where nobody cares about branding of some tool (also there wouldn’t even be a discussion about it) or places where it wasn’t a priority as long as it worked because menagerial class was limited in numbers so they couldn’t poke where it wasn’t required.

Your work environment is bad if they are so easily distracted by branding and can’t deal with ‘unporfessional’ and ‘weird’ things.

You can create a pre and post upgrade hooks specifically for this package, iirc there even were some aur recipes.

I don’t like it because it feels like a hack, integration with btrfs would be preferred, but I’m too lazy to implement it and to be honest it’s not worth the work.

Prometheus + grafana. It’s overkill tbh, most of the services restart automatically and I mostly ignore it. (it’s an artifact from life where I cared about it)

By saving them beforehand, it’s a very old trick.

Reboot is a recommended choice on any linux system after kernel upgrade (systemd as well).

It’s not a simple package manager and it can do a lot, not on the level of zypper but it’s pretty complete. There are some intentional limitations, true.

Arch package manager isn’t really simplistic? Maintainers can mitigate this easily, just goes counter arch philosophy.

Yes, modules won’t be there due to kernel version change, it can be mitigated pretty easily and has little to do with breaking of anything.

That’s not a broken kernel problem. 😉

Quick search doesn’t show anything, I mean it should be one bullet, what’s the solution? Bind mounts?

You can use pacman hooks to backup and then restore modules pre and post kernel upgrade. I’ve used it for some time, it bothered me so I’ve started to think about how to make it better (use btrfs features) but in the end just returned to reboot after upgrade scheme.

Nix fixes kernel updates? How? I know how it can be done on arch, but it’s sort of hack.

healthchecks.io and matrix.

Official container, I know.

I have no idea what it was, but home assstant is anything but ‘no lock-in’ they have a very narrow idea of what you should and shouldn’t do with it, how and why. Essentially you are (or were didn’t have time lately to check) locked to home assistant os.

Most of the hate came from people that don’t even read logs nor know what they talk about. You can read journal with strings from binutils, you can forward to syslog server and store it in whatever format you prefer, you can do a lot of things with journal tbh. (Journalctl is horribly slow, that’s true)

I have somewhat similar experience (with far more recent kernels tho, 2.6 and newer) gentoo has some gentoo specific things, it won’t really teach you a lot when you have a certain level of knowledge, and how most parts interface with each other. It’s a fun exercise if you have time.

This and some certification guide for sles or rhel (their documentation in general is very good and thorough).

Right, $14-30/m actually adds up to hundreds and is less reliable.

I know, I know super duper clouds, six nines of availability for cloud services etc., yet github (for example) fails more often then my media/git/file/print server on 10 y.o. laptop with tens of terabytes of media and data, that is less maintened and exists in a far worse conditions (comparison is incorrect but the fact stays).

When all I had was a netbook with 1gb of ram and really weak intel atom cpu, I’ve done the same, not for estetic reasons (there was a qtcurve theme that looked like gray windows theme of win 2000 - but worse, but it was extremely configurable and with it netbook turned into something marginally usable with kde 4.x).