Theo Baschak 

Sometimes I resist doing things that are probably sensible, for absolutely no reason, for way too long.

#Proxmox was one of those things. WHY was I doing fully manual libvirt/KVM when I could have been doing Proxmox?

Is #docker one of those things too? I"m actively resisting this one HAAARD right now.

What about #postgres? Also resisting this one. Should I stop using mysql for as many things and opt for postgres instead?

I'm pretty sure this is a #NeuroDivergent thing too.

Nov 13, 2022 at 2:51amWeb
Show threadKivaNov 13, 2022

@theo postgres vs mysql is no biggie. use one, be happy.

proxmox is good. but, yeah, you probably want to be working with docker at some level these days.

Show threadTheo Baschak Nov 13, 2022
@kiva I just...... don't get docker at all. Everyone around me is using it, and I'm definitely missing something painfully obvious about how to operate it.
Show threadKivaNov 13, 2022
@theo yeah. I get it. it's a change in thinking. but at the end of the day it's just a different level of abstraction. Once you get over the initial hurdle it's smooth sailing from there. as a multi-decade sysadmin, honestly I think containers are easier to operate than VMs/metal once you get your foot in the door. but the first step isn't easy.
Show threadTheo Baschak Nov 13, 2022

@kiva hopefully with this new server and re-setting up pretty much everything, I'll find some things to run that way and wrap my head around how it works.

I guess my big question was always where was the data stored, not the binaries, but the data itself. How did you configure where it was stored, etc. That seems to be really glossed over in a getting started with docker guide. I just see variations of the same 2 docker CLI lines over and over on every repo.

Show threadKivaNov 13, 2022
@theo there's a few options there depending on how you're running things. but if you're in native docker, look at docs on docker volumes and that should help clear things up. you can also mount specific directories on the host machine as well.
Show threadorchestrated messNov 14, 2022
@theo @kiva A change of perspective might help here. For devs, containers provide a huge advantage when packaging up their stuff. Esp. Python or Ruby code. Just put it into a container and whatever is in there you don’t care anymore. From the ops perspective it’s more a isolation thing: you isolate the host from the clutter of dependencies. The cruft stays inside the container and the host clean 🧽
Show threadHacker MemesNov 13, 2022

@theo still prefer mysql and sqlite to postgres, it has some really nice features but I have never found it to be as performant for most function.

docker i'm not against, but dislike because it's technical inefficiency and its really just a bandage to the problem of #butitworksonmymachine=

Show threadTheo Baschak Nov 13, 2022

@i0null most of the things I'm running for my own lab use or out of interest are generally pretty performant just due mainly to their small size, no matter how they're setup.

I've got this new lab server up now, maybe I'll find some kinda service to want to run via docker instead. I'm already storing a bunch of my infrastructure as git repos. There's a whole ecosystem of kinda interesting docker tooling and utilities that I just need to find the right situation to use to introduce myself to and it'll likely all make sense.

I agree though that a lot of it seems like bandages to make things work. One of the things I like about stuff written in Golang is how portable the binaries are. I see some of that in docker too.

Show threadHacker MemesNov 13, 2022

@theo I agree, the whole point of the ecosystem is to cater to different use cases.

Analysis of large data sets will be different to complex analytics on data subsets.

My advise would be to investigate the differences and deploy according to use case.

Show threadJay SmithNov 14, 2022

@theo I think you might find docker to be worth looking into. I was in the same boat as you until I took a job that used it extensively. I found that it has a lot of cool uses that go beyond what I originally expected it to do. I can't imagine going without it now.

Two pleasant surprises come to mind:
It's great to be able to run somebody else's proof-of-concept app and have the only dependency be docker. It saved me a lot of time while testing out log4shell detection techniques last year.

As a developer, it's also a great tool for simplifying the build process - no more spending time getting the build environment set up perfectly on each developer's machine. I especially love open source projects that give me a docker container that can do the entire build for me.