Live data from Hacker News

Apple Silicon M1 Chips and Docker

docker.com

101–110 of 129 posts

Re: Apple Silicon M1 Chips and Docker

#101
post #42

Earlier quoted context omitted.

I don't want x86 Docker on ARM.

What the point of docker if you can't just run single binary everywhere? That was its promise, right?

No it wasn't, you are thinking of Java.

Docker was snapshots of developer machines you can run in production. They never promised that it would run everywhere.

Re: Apple Silicon M1 Chips and Docker

#102
post #84

If Docker can run on Raspberry Pi's ARM chips, it can run on Mac ARM chips. I expect we'll see this working very soon after the new Macs start to arrive.

Docker uses Linux-specific containerization technology, so getting it to run on MacOS and Windows is a completely different beast of a problem.

Docker runs in a Linux VM if you use Windows or Mac,and nothing will change that here.

The "different beast" is just some go wrappers.

Re: Apple Silicon M1 Chips and Docker

#103

Earlier quoted context omitted.

The question is: why even care? Use an Arm image. For docker, it makes little to no difference

That requires that ARM images are available, which will take time and always have a disconnect. It also violates the idea of being able to run the exact same image on your laptop (ARM Mac) and in prod (x86 server).

> It also violates the idea of being able to run the exact same image on your laptop (ARM Mac) and in prod (x86 server).

99% of people are running node/python/RoR/PHP on Ubuntu/Debian/Arch. It works the same for most cases. (possible caveat: I'm not sure if numpy and TensorFlow is supported already but it probably is)

But for basic stuff, it's a non-issue

Re: Apple Silicon M1 Chips and Docker

#106
post #81

Maybe building containers for one operating system and architecture from a completely different operating system and architecture is not the best idea.

How do you think people write and compile software for any embedded or other low-powered architecture out there? The vast majority of hardware in use around the world today isn't desktop class, and emulation works just fine for building on those.

Cross compilers are nowhere near the same as Docker on ARM Mac building x86 Linux containers. The biggest difference: no one expects you to be able to run your MIPS binary on the machine you built it on.

Edit: although sometimes you can through qemu.

Re: Apple Silicon M1 Chips and Docker

#107
post #98
post #32

Can we give them a full month to sort this out ? I want x64 docker on Arm , as this could also be massive for server power consumption. But things take time . If your some wiz kid who can crack this puzzle you’ll have a billion dollar startup within 6 months As someone who’s published on the App Store , having a single binary run on IPhone , IPad and Mac is a much much bigger deal than most of you realize .

And if some startup is doing it, i know enough companies who will instantly buy a better docker for mac solution if it does have better filesystem performance. Accessing the macs filesystem from the docker container is really slow, there are some solution (e.g. docker-sync) but they all stop working after some time.

Is Apple still offering options that such kind of software could actually be developed by non-apple on _their_ system?

Re: Apple Silicon M1 Chips and Docker

#108

Earlier quoted context omitted.

They mention using Apple's Hypervisor Framework. On Apple Silicon, this runs ARM64 programs only (Rosetta 2 is unavailable to programs running inside Apple's Hypervisor Framework). https://developer.apple.com/documentation/apple_silicon/abou... Unless they plan to run an emulator inside the Hypervisor Framework (a seemingly unnecessary effort since you could run the emulator without Hypervisor) this implies x86 conta…

Docker for Mac already can build and run cross-arch containers: https://docs.docker.com/docker-for-mac/multi-arch/ And if you're on x86, you can run ARM containers with QEMU userspace emulation seamlessly after running docker run --rm --privileged docker/binfmt:a7996909642ee92942dcd6cff44b9b95f08dad64 The inverse would be extremely easy as well.

Seamlessly, but with a heavy performance impact. In my experience it's not really usable in anything other than running specific single tests, or some short demos.

Re: Apple Silicon M1 Chips and Docker

#109

Earlier quoted context omitted.

You can run on ARM servers in production. They're even cheaper than x86 servers. (Though also slower, at least AWS's version.) Realistically, ARM is probably the future, and just needs something to push society in the direction of the future. A lot of people use Macs, so this forces people to become interested in switching, which has some value. Remember when the original iMac had USB instead of ADB ports? Everyone w…

> You can run on ARM servers in production. They're even cheaper than x86 servers. (Though also slower, at least AWS's version.) Where can I buy ARM server comparable with x86 server on performance? I'm not interested with clouds, I want bare metal server. I'm not aware of any server brand like HP or Dell producing ARM servers. I'm also not aware of any CPU manufacturer producing computer desktops with ARM comparable…

> I'm not interested with clouds, I want bare metal server.

You can have both.

https://aws.amazon.com/blogs/aws/now-available-bare-metal-ar...

Re: Apple Silicon M1 Chips and Docker

#110

Earlier quoted context omitted.

What are you going to deploy that ARM image to?

Aws Graviton2 images. It’s exciting to finally have ARM consumer laptops: it’ll encourage people to port applications to ARM which will make it easier to take advantage of cheap ARM servers.

> It’s exciting to finally have ARM consumer laptops:

Apple may be the loudest and now the fastest available, but the ARM consumer laptops existed for some time already. For example https://www.lenovo.com/au/en/laptops/ideapad/ideapad-s-serie...

Post reply on HN