Skip Navigation
Selfhosted @lemmy.world
Jofus @lemmy.world

What is Docker?

Hi! Im new to self hosting. Currently i am running a Jellyfin server on an old laptop. I am very curious to host other things in the future like immich or other services. I see a lot of mention of a program called docker.

search this on The internet I am still Not very clear what it does.

Could someone explain this to me like im stupid? What does it do and why would I need it?

Also what are other services that might be interesting to self host in The future?

Many thanks!

EDIT: Wow! thanks for all the detailed and super quick replies! I've been reading all the comments here and am concluding that (even though I am currently running only one service) it might be interesting to start using Docker to run all (future) services seperately on the server!

92 comments

  • Docker enables you to create instances of an operating system running within a “container” which doesn’t access the host computer unless it is explicitly requested. This is done using a Dockerfile, which is a file that describes in detail all of the settings and parameters for said instance of the operating system. This might be packages to install ahead of time, or commands to create users, compile code, execute code, and more.

    This instance of an operating system, usually a “server,” is great because you can throw the server away at any time and rebuild it with practically zero effort. It will be just like new. There are many reasons to want to do that; who doesn’t love a fresh install with the bare necessities?

    On the surface (and the rabbit hole is deep!), Docker enables you to create an easily repeated formula for building a server so that you don’t get emotionally attached to a server.

  • Thanks for asking this question. These replies are so much more helpful in understanding the basic premise than anything I've come across.

  • It's the platform that runs all of your services in containers. This means they are separated from your system.

    Also what are other services that might be interesting to self host in The future?

    Nextcloud, the Arr stack, your future app, etc etc.

  • Learn Docker even if you have a single app. I do the same with a Minecraft server.

    • No dependency issues
    • All configuration (storage/network/application management) can be done via a single file (compose file)
    • Easy roll-backs possible
    • Maintain multiple versions of the app while keeping them separate
    • Recreate the server on a different server/machine using only the single configuration file
    • Config is standardized so easy to read

    You will save a huge amount of time managing your app.

    PS: I would like to give a shout out to podman as the rootless version of Docker

  • Wow! Thank you all for the civilized responses. This all sounds so great. I am older and I feel like I've already seen enough ads for one lifetime and I hate all this fascist tracking crap.

    But how does that work? Is it just a network on which you store your stuff in a way that you can download it anywhere or can it do more? I mean, to me that's just a home network. Hosting sounds like it's designed for other people to access. Can I put my website on there? If so, how do I go about registering my domain each year. I'm not computer illiterate but this sounds kind of beyond my skill level. I'll go search Jellyfin, weird name, and see what I can find. Thanks again!

    • You're asking a lot of questions at one time and will be better served understanding you're knocking at the door of a very deep rabbit hole.

      That said, I'll try to give you the basic idea here and anyone who can correct me, please do so! I doubt I'll get everything correct and will probably forget some stuff lol.

      So, self hosting really just means running the services you use on your own machine. There's some debate about whether hosting on a cloud server - where someone else owns and has physical access to the machine - counts as self hosting. For the sake of education, and because I'm not a fan of gatekeeping, I say it does count.

      Anyway, when you're running a server (a machine, real or virtualized, that is running a program connected to a network that can - usually - be accessed by other machines connected to that network), who and what you share with other machines on your network or other networks, is ultimately up to you.

      When using a "hosted" service, which is where another entity manages the server (not just the hardware, but the software and administration too, and is therefore the opposite of self hosting. Think Netflix, as opposed to Jellyfin), your data and everything you do on or with that service on that network belongs to the service provider and network owners. Your "saved" info is stored on their disks in their data center. There are of course exceptions and companies who will offer better infrastructure and privacy options but that's the gist of non-self-hosted services.

      To your specific questions:

      But how does that work?

      Hopefully the above helps, but this question is pretty open ended lol. Your next few questions are more pointed, so I'll try to answer them better.

      Is it just a network on which you store your stuff in a way that you can download it anywhere or can it do more?

      Well, kind of. If you're hosting on a physical machine that you own, your services will be accessible to any other machine on your home network (unless you segment your network, which is another conversation for another time) and should not, by default, be accessible from the internet. You will need to be at home, on your own network to access anything you host, by default.

      As for storage of your data, self hosted services almost always default to local storage. This means, you can save anything you're doing on the hard-drive of the machine the server is running on. Alternatively if you have a network drive, you can store it on another machine on your network. Some services will allow you to connect to cloud storage (on someone else's machine somewhere else). The beauty is that you decide where your data lives.

      I mean, to me that's just a home network. Hosting sounds like it's designed for other people to access. Can I put my website on there?

      Like almost anything with computers and networking, the defaults are changeable. You can certainly host a service on the internet for others to access. This usually involves purchasing the rights to a domain name, setting that domain up to link to your private IP address, and forwarding a port on your router so people can connect to your machine. This can be extremely dangerous if you don't know what you're doing an isn't recommended without learning a lot more about network and cyber security.

      That said, there are safer ways to connect from afar. Personally, I use a software called Wireguard. This software allows devices I approve (like my phone, or my girlfirend's laptop) to connect to my network when away from home though what is called an "encrypted tunnel" or a "Virtual Private Network (VPN) ". These can be a pain to set up for the first time if you're new to the tech and there are easier solutions I've heard of but haven't tried, namely Tailscale, and Netbird, both of which use Wireguard but try to make the administration easier.

      You can also look into reverse proxies, and services like cloudflare for accessing things away from home. These involve internet hostng, and security should be considered, like above. Anything that allows remote access will come with unique pros and cons that you'll need to weigh and sort for yourself.

      If so, how do I go about registering my domain each year.

      Personally, I use Porkbun.com for cheap domains, but there are tons of different providers. You'll just have to shop around. To actually use the domain, I'm gonna be linking some resources lower in the post. If I remember correctly, landchad.net was a good resource for learning about configuring a domain but idk. There will be a few links below.

      I'm not computer illiterate but this sounds kind of beyond my skill level.

      It was beyond my skill level when I started too. It's been nearly a year now and I have a service that automatically downloads media I want, such as movies, shows, music, and books. It stores them locally on a stack of hard drives, I can access them outside of my house with wireguard as well. Further, I've got some smaller services, like a recipe book I share with my girlfriend and soon with friends and family. I've also started hosting my own AI, a network wide ad-blocker, a replacement for Google photos, a filesharing server, and some other things that are escaping me right now.

      The point is that it's only a steep hill while you're at the bottom looking up. Personally, the hike has been more rejuvenating than tiresome, though I admit it takes patience, a bit of effort, and a willingness to learn, try new things, and fail sometimes.

      Never sweat the time it takes to accomplish a task. The time will pass either way and at the end of it you can either have accomplished something, or you'll look back and say, "damn I could've been done by now."

      I'll go search Jellyfin, weird name, and see what I can find. Thanks again!

      Also check these out, if you're diving in:

      YouTube:

      Guides:

      Tools:

      Hopefully this helps someone. Good luck!

  • Now compare Docker vs LXC vs Chroot vs Jails and the performance and security differences. I feel a lot of people here are biased without knowing the differences (pros and cons).

  • I'm not sure how familiar you are with computers in general, but I think the best way to explain Docker is to explain the problem it's looking to solve. I'll try and keep it simple.

    Imagine you have a computer program. It could be any program; the details aren't important. What is important, though, is that the program runs perfectly fine on your computer, but constantly errors or crashes on your friend's computer.

    Reproducibility is really important in computing, especially if you're the one actually programming the software. You have to be certain that your software is stable enough for other people to run without issues.

    Docker helps massively simplify this dilemma by running the program inside a 'container', which is basically a way to run the same exact program, with the same exact operating system and 'system components' installed (if you're more tech savvy, this would be packages, libraries, dependencies, etc.), so that your program will be able to run on (best-case scenario) as many different computers as possible. You wouldn't have to worry about if your friend forgot to install some specific system component to get the program running, because Docker handles it for you. There is nuance here of course, like CPU architecture, but for the most part, Docker solves this 'reproducibility' problem.

    Docker is also nice when it comes to simply compiling the software in addition to running it. You might have a program that requires 30 different steps to compile, and messing up even one step means that the program won't compile. And then you'd run into the same exact problem where it compiles on your machine, but not your friend's. Docker can also help solve this problem. Not only can it dumb down a 30-step process into 1 or 2 commands for your friend to run, but it makes compiling the code much less prone to failure. This is usually what the Dockerfile accomplishes, if you ever happen to see those out in the wild in all sorts of software.

    Also, since Docker puts things in 'containers', it also limits what resources that program can access on your machine (but this can be very useful). You can set it so that all the files it creates are saved inside the container and don't affect your 'host' computer. Or maybe you only want to give permission to a few very specific files. Maybe you want to do something like share your computer's timezone with a Docker container, or prevent your Docker containers from being directly exposed to the internet.

    There's plenty of other things that make Docker useful, but I'd say those are the most important ones--reproducibility, ease of setup, containerization, and configurable permissions.

    One last thing--Docker is comparable to something like a virtual machine, but the reason why you'd want to use Docker over a virtual machine is much less resource overhead. A VM might require you to allocate gigabytes of memory, multiple CPU cores, even a GPU, but Docker is designed to be much more lightweight in comparison.

  • good answers already so i will give you a different example.

    my basic understanding of it is that docker was created originally for developers. im not sure if anyone planned for it to be a way to package up software for end users.

    before docker existed you would have this issue where devs would be working on an app, say jellyfin, but each dev might be on a different platform (windows, mac, linux), or be using a different OS version, or different versions of whatever software... which meant it happened often that the app would work for one dev but not another. maybe one dev updated C# to version 2.3 and told everyone else to update, but someone missed the memo and is still running version 2.2 and now jellyfin wont work for them and time would be wasted trying to figure out where the mismatch was

    so docker was a way to fix that "version hell" problem. every single thing that is needed for the app to run is kept inside the container. one dev will update something to a new version, then that container is shared to all other devs and each dev only has to worry about updating to the newest container before they start working on something.

    app settings are kept in a separate location and the app data in another. in the case of jellyfin, the app data would be the movies or tv shows folder for example. then when you start the docker container, it will symlink those 2 locations/folders inside the container and the jellyfin app can access them as if they were folders that were actually stored inside the container.

    so having the settings and data separate like that makes it very easy to update the container to a new version, or for a developer is probably useful being able to rollback to an older container for testing. its similar to how say windows puts the program files in one location and settings in the appdata folder

    for end users its handy if theres a new version of jellyfin or whatever that isnt released yet but you want try it out, you can run 2 containers at the same time and both of them can access the same settings and data. (maybe with the newer one in read-only mode so it doesnt mess up your settings or data!)

92 comments