I've been getting pretty comfortable with Docker the past few months, but this video was really what I needed to fill in the gaps! Love all your stuff.
The best Docker video I could find. Every essential topic for the novice, combined with some tips, and up-to-date material - what more can anyone ask for? Just loved it. Keep them coming!
I use docker every now and then and I often forget key points in it that need to be revised or I need to onboard someone onto something that uses Docker, and seeing you upload this video gave me some sense of relief because I know it will be high quality, thanks!
I learnt docker last year but this video is good at checking up my knowledge from self learning docker, now it is all concise for future learner! will definitely send these to colleagues who need to learn this
I've been watching Docker tutorials every now and then since last year (maybe even the year before that), but only now am I starting to dockerize my own applications haha. It's such a convenient tool. Right now with my limited understanding, I use compose for local development, and Dockerfiles for deployment
really great resource, watched it through and will be referring back to this video in the future. thank you for the straightforward, motivated explanations!
Really nice and clear video, as usual! Your channel has been my goto to recommend to my colleagues to learn when they need to dive deep into Python topics (and now docker!) :) Thanks for all the content
Great video! Answering to your demoscript: no, I didn't run it, I executed docker this way: docker run mcodingllc/demoscript:0.0.1 /bin/sh and then used cat to see the contents of run.sh, but good one lol
Excellent tutorial! I wish Docker could help me where I struggle with dependency and compatibility issues most, but I don't think it can. I am an embedded developer (so no support for Docker on our targets), but as such, you end up with various different tool chains and libraries on your computer, often causing compatibility issues or at least quite a headache until you got things configured to somehow cooperate. But with us working in a Windows environment (because our IT department doesn't know Linux🙄) and most of these tools being GUI based, I don't think that Docker could help containerizing things there. For us, it goes so far, that we started a project not too long ago using a 20 year old tool chain (which doesn't even support C99 or C++98 fully), because we can't install a newer version - that would break the old version which we still need to run to support older products. You just can't have both versions on one computer at the same time, it has even been confirmed by the tech support of the tool chain vendor.. (At least I found a way to use VS Code instead of their crappy editor and get things somewhat integrated, but it still sucks...) And of course we don't have the manpower to convert the legacy code to the new environment, which (of course) is not compatible in some finicky ways (let's just say that previous generations of programmers exploited erroneous behavior of this very compiler - so if a newer one does things right, have fun)... And of course being bound to Windows means that you can't just spin up separate VMs for everything either, as for each and every one of them you would have to get licensing straight.
One problem I can't seem to find a solid answer to that has made it difficult using Docker is how do you modify the host's firewall without restarting the daemon? Most firewals use iptables under the hood and the Docker daemon creates its iptables chains and many virutal networks as needed, say for a docker-compose app with some ports exposed to the host. After changing firewall rules (e.g. for exposing a port on the host publicly) by doing `iptables flush` or `iptables-restore ` all networking configuration for Docker is gone and the entire service needs to be restarted with all currently running containers in order to regenerate it and have the networking work again. This is not acceptable as I often need to modify the firewall for development and testing purposes and have long-running applications as well without interruption. Some advice?
I just thought that Docker would provide the isolation to begin with. I did the .venv and it seems to work fine. Then I nested containers without .venv which made it slimmer in terms of resources it’s using. But the same effect as far as I can tell. I imagine in production you’d want that generally speaking.
Thanks! In theory, you use a venv in a docker container for the same reason you use one on the host. Outside a venv, python is your system python, and critical programs required for your operating system to function might depend on the packages already installed. You installing new ones or upgrading them might therefore break your system (inside the container). In practice, I have not had an issue with this even once and I haven't seen anyone use a venv inside docker in production. But in theory, it could be a problem.
Damn I have quite the road ahead of me moving from remote support to get to cloud architect Would love to hear some input on what I'll have to master to get there. I'm 22 years old
Quick question - at 35:55, why copy and install the wheels instead of installing "normally" in the builder and then just copying the python site-packages directory?
Hmm I think either one should work fine. In general if you had build dependencies they would be in your site packages too so you would want to start fresh and install wheels to avoid copying build deps. In this case we didnt have any so I don't see any reason copying the site packages wouldnt work too.
I have Install Docker on separate local disk D using this command start /w "" "Docker Desktop Installer.exe" install -accept-license --installation-dir="D:\Docker\Docker" --wsl-default-data-root="D:\Docker\wsl" --windows-containers-default-data-root="D:\\Docker" But I am still facing storage issues in the local disk C. Does anyone know how to solve this issue
If you are using Docker Engine directly with local volumes, then you can find them as just a plain directory somewhere. With Docker Desktop it becomes slightly more obfuscated because the volumes live within the Docker Desktop VM for performance reasons. If you are using other volume types besides local, like an efs volume, then it becomes increasingly more inconvenient to work with them outside containers!
My "software" is two simple rows, to install I must learn 5 new super complex languages to write on them and install several systems bigger than my own computer?! Never use make files, never use linters, never use tests, never docker, no absurd buggy corporative linux insanity (terminal based) flood bloat or js chaotic insane "frameworks", never!!!
This must be a troll comment, or OP is a teenager. It's fine to think that docker is over engineered; but to suggest that tests, linters and even makefiles are bloat is pure amateurism.
How important and enormous must be my application, to suggest myself and my clients the whole big system - several GB to run my... my 100 rows code?! Because of this I don't use node, react, electron, python environments, almost all js technologies, etc., all wants for each "hello word" project to install hundreds MB again and again. If SSD become 100TB for 1$ reliable for 30 years, I will change my mind 🤣
you dont really need docker, environments and other ci/cd stuff for small hobby projects. But if you are building critical application ability to scale it at please and run it in isolated environment could save man hours
Docker is a tool that was developed to solve problems that arose in real world applications, typically with large scale deployments and complex systems. As you can sense for yourself, Docker is not a one size fits all solution for all situations. It's perfectly possible that Docker isn't the right tool for your current project, and that's okay. Only you know what's best for your projects, so trust your instincts!
Docker is a tool that was developed to solve problems that arose in real world applications, typically with large scale deployments and complex systems. As you can sense for yourself, Docker is not a one size fits all solution for all situations. It's perfectly possible that Docker isn't the right tool for your current project, and that's okay. Only you know what's best for your projects, so trust your instincts!
For each tiny part of software, take new different computer with full system installed haha 🤣🤦🏻 This only linux user can invent 🤣, but Windows tradesmen likes this 😱
James Murphy with the "just what I needed" video once again.
It's a skill!
its like he reads our minds
I've always wanted to learn docker but could never get myself to stick to it. Thank you, mCoding!
This is your sign! Don't put it off!
Not even the docker website could explain for me what it actually is. You've solved this mystery, thank you.
You're very welcome!
mCoding is back at it again with another banger that works better than official docs.
I've been getting pretty comfortable with Docker the past few months, but this video was really what I needed to fill in the gaps! Love all your stuff.
The best Docker video I could find. Every essential topic for the novice, combined with some tips, and up-to-date material - what more can anyone ask for? Just loved it. Keep them coming!
Thanks! I appreciate it!
I use docker every now and then and I often forget key points in it that need to be revised or I need to onboard someone onto something that uses Docker, and seeing you upload this video gave me some sense of relief because I know it will be high quality, thanks!
Excellent tutorial! One of the best I've seen in a while.
Thank you, hope it helps you in the future!
Awesome! And congratulations for the Docker sponsor, that talks a lot about the quality of your channel
Thanks a ton!
I learnt docker last year but this video is good at checking up my knowledge from self learning docker, now it is all concise for future learner! will definitely send these to colleagues who need to learn this
Thanks so much!
I've been watching Docker tutorials every now and then since last year (maybe even the year before that), but only now am I starting to dockerize my own applications haha. It's such a convenient tool.
Right now with my limited understanding, I use compose for local development, and Dockerfiles for deployment
Docker swarm or Kubernetes next please.
really great resource, watched it through and will be referring back to this video in the future. thank you for the straightforward, motivated explanations!
Thanks for that thorough run through. Subscribed.
K8s/Helm video next? :D
This goes not only to my "useful dev" playlist, but also straight to my "TOP" playlist.
Superb work :-)
Much appreciated!
A docker sponsorship… impressive!
because there are many competitors
Finally i could add 10 years of docker experience by watching your video
I haven't even watched the video yet, but I know it'll be good. I needed to learn Docker soon for work, so thanks for this!
Haha hopefully it lives up to expectations!
Really nice and clear video, as usual! Your channel has been my goto to recommend to my colleagues to learn when they need to dive deep into Python topics (and now docker!) :) Thanks for all the content
I think a demo of bind mounts would've been nice, e.g. for automatically restarting the Python API on code changes. Maybe it is out of scope.
Awesome comprehensive training
Thank you very much
love this topic
Great video. Thanks. Can you make one on docker network? You explain very well.
Great video!
Answering to your demoscript: no, I didn't run it, I executed docker this way: docker run mcodingllc/demoscript:0.0.1 /bin/sh and then used cat to see the contents of run.sh, but good one lol
btw, you'll probably need to change the permissions to make it executable
Excellent tutorial!
I wish Docker could help me where I struggle with dependency and compatibility issues most, but I don't think it can.
I am an embedded developer (so no support for Docker on our targets), but as such, you end up with various different tool chains and libraries on your computer, often causing compatibility issues or at least quite a headache until you got things configured to somehow cooperate. But with us working in a Windows environment (because our IT department doesn't know Linux🙄) and most of these tools being GUI based, I don't think that Docker could help containerizing things there.
For us, it goes so far, that we started a project not too long ago using a 20 year old tool chain (which doesn't even support C99 or C++98 fully), because we can't install a newer version - that would break the old version which we still need to run to support older products. You just can't have both versions on one computer at the same time, it has even been confirmed by the tech support of the tool chain vendor.. (At least I found a way to use VS Code instead of their crappy editor and get things somewhat integrated, but it still sucks...) And of course we don't have the manpower to convert the legacy code to the new environment, which (of course) is not compatible in some finicky ways (let's just say that previous generations of programmers exploited erroneous behavior of this very compiler - so if a newer one does things right, have fun)...
And of course being bound to Windows means that you can't just spin up separate VMs for everything either, as for each and every one of them you would have to get licensing straight.
6:23 docker run hello-world
awesome output. niiiice
THE docker tutorial!!
Great video, thanks so much
Kubernetes next please : )
One problem I can't seem to find a solid answer to that has made it difficult using Docker is how do you modify the host's firewall without restarting the daemon?
Most firewals use iptables under the hood and the Docker daemon creates its iptables chains and many virutal networks as needed, say for a docker-compose app with some ports exposed to the host. After changing firewall rules (e.g. for exposing a port on the host publicly) by doing `iptables flush` or `iptables-restore ` all networking configuration for Docker is gone and the entire service needs to be restarted with all currently running containers in order to regenerate it and have the networking work again.
This is not acceptable as I often need to modify the firewall for development and testing purposes and have long-running applications as well without interruption. Some advice?
I just thought that Docker would provide the isolation to begin with. I did the .venv and it seems to work fine. Then I nested containers without .venv which made it slimmer in terms of resources it’s using. But the same effect as far as I can tell. I imagine in production you’d want that generally speaking.
I didnt ask for this but this is realllly welcome!
Congrats on the sponsor that's cool
I use devcontainers at work.
Docker? I don't even know her!
Thanks
had to add pip install uvicorn to get the backend fastapi bit working
Nice video! One thing though, why would you use a .venv inside along with docker containers? Thanks.
Thanks! In theory, you use a venv in a docker container for the same reason you use one on the host. Outside a venv, python is your system python, and critical programs required for your operating system to function might depend on the packages already installed. You installing new ones or upgrading them might therefore break your system (inside the container). In practice, I have not had an issue with this even once and I haven't seen anyone use a venv inside docker in production. But in theory, it could be a problem.
Perfect timing
Random question just to help with engagement, but have you played with Python 3.14 at all?
Too far away, I'm looking out closely for 3.13 though!
Damn I have quite the road ahead of me moving from remote support to get to cloud architect
Would love to hear some input on what I'll have to master to get there. I'm 22 years old
Good~
Kubernetes next!
great
Amazing video, packed with info. Thank you.
Also can you share the requirements.txt that you showed at 29:28?
Will upload later but you can generate for yourself with pip install fastapi motor then pip freeze
when I generate it myself it is much shorter, only containing 10 or so requirements
please can you make video about deployment?
Can do if there's enough interest!
@@mCoding yes please because all youtubers don't cover this topic
Quick question - at 35:55, why copy and install the wheels instead of installing "normally" in the builder and then just copying the python site-packages directory?
Hmm I think either one should work fine. In general if you had build dependencies they would be in your site packages too so you would want to start fresh and install wheels to avoid copying build deps. In this case we didnt have any so I don't see any reason copying the site packages wouldnt work too.
can't find the source code, can you please share it?
Can u make kubernetes in 1 hour if possible please
HE IS BACK ?????
Never left! This vid just took a long time to make :)
@@mCoding that I believe without a doubt. That’s very dense material. Chapeau’s 👍🏻
What's the solution for having to write similar/same details in each .env file?
Secrets managers (mentioned but not explored in video) can help with this by allowing your instances to pull all their secrets at runtime.
❤
Tried to install Docker Desktop inside Ubuntu, but I can't run it.
Check the prerequisites carefully! Not all versions of Ubuntu are supported!
🌹
Kubernetes when ?
Thanks. but this is not for beginners.. You got me started fine and then shortly later, I was lost.. I will keep my nose to the grindstone though.
I have Install Docker on separate local disk D using this command
start /w "" "Docker Desktop Installer.exe" install -accept-license --installation-dir="D:\Docker\Docker" --wsl-default-data-root="D:\Docker\wsl" --windows-containers-default-data-root="D:\\Docker"
But I am still facing storage issues in the local disk C.
Does anyone know how to solve this issue
Well.. the volume files can still be found inside docker. They're not encrypted od compressed. It's just not convenient. So the point still stands.
If you are using Docker Engine directly with local volumes, then you can find them as just a plain directory somewhere. With Docker Desktop it becomes slightly more obfuscated because the volumes live within the Docker Desktop VM for performance reasons. If you are using other volume types besides local, like an efs volume, then it becomes increasingly more inconvenient to work with them outside containers!
incredibly complicated 😅
And what’s the difference to the other hundreds of Docker videos on CZcams?
This one is made by me :)
@@Eisesser01 Watch it and it will become apparent. This content is at a whole different level.
My "software" is two simple rows, to install I must learn 5 new super complex languages to write on them and install several systems bigger than my own computer?! Never use make files, never use linters, never use tests, never docker, no absurd buggy corporative linux insanity (terminal based) flood bloat or js chaotic insane "frameworks", never!!!
This must be a troll comment, or OP is a teenager.
It's fine to think that docker is over engineered; but to suggest that tests, linters and even makefiles are bloat is pure amateurism.
How important and enormous must be my application, to suggest myself and my clients the whole big system - several GB to run my... my 100 rows code?! Because of this I don't use node, react, electron, python environments, almost all js technologies, etc., all wants for each "hello word" project to install hundreds MB again and again. If SSD become 100TB for 1$ reliable for 30 years, I will change my mind 🤣
you dont really need docker, environments and other ci/cd stuff for small hobby projects. But if you are building critical application ability to scale it at please and run it in isolated environment could save man hours
Docker is a tool that was developed to solve problems that arose in real world applications, typically with large scale deployments and complex systems. As you can sense for yourself, Docker is not a one size fits all solution for all situations. It's perfectly possible that Docker isn't the right tool for your current project, and that's okay. Only you know what's best for your projects, so trust your instincts!
Docker is a tool that was developed to solve problems that arose in real world applications, typically with large scale deployments and complex systems. As you can sense for yourself, Docker is not a one size fits all solution for all situations. It's perfectly possible that Docker isn't the right tool for your current project, and that's okay. Only you know what's best for your projects, so trust your instincts!
For each tiny part of software, take new different computer with full system installed haha 🤣🤦🏻
This only linux user can invent 🤣, but Windows tradesmen likes this 😱
I mean its look sane when you need isolated system
As windows user, it's difficult for me to follow every instructions. Please make a new video using windows command prompt. 🙏🏼
can i say phenomenal - short and crisp🫡
Thank you, I appreciate the support!
@@mCoding do you stream ?