alt.hn

6/10/2026 at 12:29:01 AM

macOS Container Machines

https://github.com/apple/container/blob/main/docs/container-machine.md

by timsneath

6/10/2026 at 1:13:28 AM

To clarify a few comments here: this is not only OCI containers: container machines add support for persistence and filesystem mounting, making container machines a great lightweight Linux environment for developers using macOS. More details here: https://developer.apple.com/videos/play/wwdc2026/389

by timsneath

6/10/2026 at 10:13:11 AM

Still feels like a apple-ified microvm

by CaptainCyber99

6/10/2026 at 4:58:34 AM

> ... highly integrated Linux environment that works seamlessly on your Mac. ...

Which kernel is running, and is it hosted in hypervisor.framework, as is done with UTM (when not using the qemu mode)?

by jjtheblunt

6/10/2026 at 5:22:58 AM

The katas container kernel by default.

by Scarbutt

6/10/2026 at 6:54:39 AM

> filesystem mounting

How is this different to bind mounts

by bogantech

6/10/2026 at 8:31:33 AM

Very different: Linux running in a virtual machine can't bind mount into a macOS host's filesystem. So they use virtiofs.

by jdub

6/10/2026 at 9:12:07 AM

This applies to both containers and container machines though, right?

by lxgr

6/10/2026 at 8:41:35 AM

how does that compare to something like, eg, Orbstack?

by oulipo2

6/10/2026 at 1:34:33 AM

Ah, the Darwin/BSD Subsystem for Linux.

by Onavo

6/10/2026 at 1:55:19 AM

Not quite, it’s still a VM. And while it supports virtio balloon for growing RAM, it doesn’t yet support releasing that RAM back to the host. And there isn’t a convenient way to shrink the sparse disk images as they grow yet, either.

by CGamesPlay

6/10/2026 at 2:00:58 AM

Isn't the Windows subsystem for Linux (the reference there) also a VM?

by AlexB138

6/10/2026 at 2:03:17 AM

Only WSL2; WSL1 was an actual subsystem.

by gsnedders

6/10/2026 at 2:09:20 AM

So this is Darwin/BSD Subsystem for Linux 2.

by selcuka

6/10/2026 at 4:26:11 AM

Yes.

by rvz

6/10/2026 at 2:37:57 AM

WSL1 was so cool, WSL2 made it boring and isolated.

by LoganDark

6/10/2026 at 8:08:38 AM

WSL1 was very conceptually appealing, and ended up working very poorly because of the poor matching between Linux syscalls and the Windows kernel. Git suffered terribly as a result. The inverse is also somewhat true - there have been cases where Wine is much slower than native Windows because Linux simply doesn't provide a simple way to achieve the same outcome, and interestingly the Wine developers have had reasonable (if tediously slow) success in making it possible to express the same semantics to Linux and have it handle things fast. It would be fascinating to know whether WSL1 developers didn't have enough traction to get Windows internals altered to match, or whether it's just way harder to do the same under Windows.

by mjg59

6/10/2026 at 8:34:11 AM

Wine achieves better performance these days due to things like... adding a module to the Linux kernel that implements NT-like synchronization primitives. So, Linux subsystem for NT synchronization basically. (a.k.a. NTSync)

Maybe this works out better because Linux is more flexible, while Windows/NT is more "set in its ways" and therefore more difficult to implement Linux on top of... Maybe?

by LoganDark

6/10/2026 at 2:42:32 AM

Back in my day you to to download a couple GB worth of cygwin, and that wasn't an actual environment, basically just a GNU toolchain compiled for windows. But it got you like....grep and bash and stuff that ran natively on windows which was kinda cool.

by TylerE

6/10/2026 at 5:48:30 AM

Does any older folk here remembers when NT was the Cool New Thing (TM) and it had by design support to multiple subsystems plopped over the NT API, and Win32 was just one of them alongside POSIX (Interix) and OS/2? There was even a _very short_ time span when Interix was actually usable (it was extremely short though)

by qalmakka

6/10/2026 at 9:46:43 AM

I guess that makes me square within the 'older folk' subset - I continued to use the NT core with LiteSTEP alongside the SGI/IRIX Octane2 well after Y2K.

Those days I was working on a rework of the TRO PLATO learning system which was a real beast but essential for the individual learning project of a charter school i was supporting.

PLATO had been taken from it's dedicated mainframe world and made 'runnable' on W95 workstations with an NT server - but it really didn't run well, and the kids could really get behind the interface into regular Windows environment too readily. In combination the workstations were crazy hard to keep running cleanly.

So in the end; we had to take the software out of Windows, wash it clean in the waters of Silicon Graphics System-V with BSD extensions (X11) Unix and BSD - NeXTSTEP, just so we could bring it back to Windows properly using LiteStep.

Life happened and I lost touch with the outcome of it all, moving on to my next project; but, I kept a LiteSTEP desktop until moving entirely over to Linux in 2004.

Haven't used Windows for anything but a gaming load since '05 and stopped doing even that in about 2010, nothing later than XP.

by naishoya

6/10/2026 at 6:27:56 AM

Yes, the only reason I cared for Linux in first place was that the POSIX support wasn't that good.

I am convinced that if POSIX subsystem was UNIX serious, GNU/Linux would never taken off on PC, and the whole would be divided between SGI, HP-UX, Solaris, Aix and Windows NT.

by pjmlp

6/10/2026 at 6:40:50 AM

There were already better free options than Linux when Linux first started gaining traction.

The reason Linux grew in the 90s was because it was part of the hacker culture. Not because better options didn’t exist.

Kids liked the fact that Linux was a free-for-all, anything-goes, platform. It wasn’t stuffy like Unix and it wasn’t proprietary like Windows.

Then those kids grew up and became decision makers themselves. And we started to see Linux replace FreeBSD and commercial Unixes.

by hnlmorg

6/10/2026 at 7:15:12 AM

Which ones? BSD was tied in a lawsuit that left doubts on its future.

Minix was a toy OS for university teachings.

Coherent was commercial.

Nothing else was there on the PC market.

by pjmlp

6/10/2026 at 3:17:36 AM

Cygwin was fun. I'd done zero development on Windows, but about 10 years ago I had to figure out how to deploy some nightly shell scripts across a bunch of local computers in a few dozen offices, where about 80% were MacOS and the rest were Windows. I don't remember exactly how I rigged it, but basically cygwin allowed me to keep the scripts as they were and trigger them in place, with a few small modifications.

I never want to deal with that again ;)

[edit] fwiw, Termux on Android is similarly a fun pseudo-environment. It's a nice and helpful toy.

by noduerme

6/10/2026 at 3:25:13 AM

The biggest issue I remember is directory seperators... windows of course using \ which bash would then interpret as an escape. Cygwin mostly papered over that from what I can recall, but it could lead to some weirdness, like sometimes you'd get C:\\path\\es\\like\\this

by TylerE

6/10/2026 at 3:41:13 AM

You could also use forward slashes, like C:/path/subpath, which has worked since Windows 1.0/DOS 2.0.

That's handy when you're entering paths in a Cygwin/MSYS Bash shell, but might not help much if you're trying to parse or otherwise work with existing patgh variables composed with backslashes.

by rpeden

6/10/2026 at 3:59:11 AM

Yes, you could if you were entering them manually, but some apps that generated file names would screw it up. I think they were using some sort of stdlib function to get the path seperator. Forward slash paths working in native windows apps also wasn't quite a given, either. Keep in mind this was a loooong time ago... like windows xp era maybe, even.

by TylerE

6/10/2026 at 6:07:40 AM

Yeah, I recall directory paths being the biggest PITA with running scripts in cygwin. But I mean, that was a very minor set of things to fix compared to what would've had to be written in anything else available at the time.

Doing retail office deployments of custom code on employee computers is a weird niche, and you find whatever works and hope you can maintain it somehow. Cygwin was awesome though, saved me a ton of time and the client a lot of money for the moment. (The client later stipulated to all future franchisees that they had to buy only Macs, lol)

by noduerme

6/10/2026 at 9:06:25 AM

Always used / and it worked for both cygwin/windows lands.

by coldtea

6/10/2026 at 6:17:37 AM

> Back in my day you to to download a couple GB worth of cygwin

You still can, and it still works exactly the same way.

by kergonath

6/10/2026 at 7:22:09 AM

what do you mean? that's still the only way to work as a human in windows. wsl1 almost replaced it, but obviously they scrapped it.

if you must use windows, it's because you will compile for windows. so you install MSYS, which is a linux distro-ish compiled native for windows. and do your work.

wsl2 (and this apple thing) is just a meme. if you're working in it, you're better of just installing Linux or ssh'ing to a server.

by iririririr

6/10/2026 at 8:23:35 AM

shrug. I haven't owned a Windows machine in years at this point. It's one of those things like PHP that I just decided my life was better off without.

by TylerE

6/10/2026 at 2:49:55 AM

... Now it's just called git bash

by _blk

6/10/2026 at 2:55:41 AM

Just install and use MSYS2, git bash is derived from it anyway, and a regular MSYS2 installation offers a lot more.

by michaelsbradley

6/10/2026 at 6:23:25 AM

It was soooo slow though. Practically unusable for anything i/o heavy.

by kevinminehart

6/10/2026 at 7:13:40 AM

Those issues could have been fixed…

by dented42

6/10/2026 at 5:01:18 AM

WSL 1 is long gone for all practical purposes, yet it still dominates conversations.

Also everyone on FOSS gets it wrong, WSL wasn't a subsystem like classical Windows NT ones.

It was based on Drawbridge research using picoprocesses, a new approach for library OSes.

https://learn.microsoft.com/en-us/archive/blogs/wsl/pico-pro...

by pjmlp

6/10/2026 at 2:03:07 AM

Mac Subsystem for Linux 2

by jayd16

6/10/2026 at 6:23:44 AM

Exactly what I thought. The Mac equivalent to WSL. Which is a great thing for Mac devs. Lots of stuff expects Linux these days, not POSIX. Mach isn’t Linux.

by pseudosavant

6/10/2026 at 8:30:40 AM

I don‘t understand why these tools always advertise about mounting the $HOME inside the container. Isn‘t it better to have a complete isolation? Isn‘t that the point of using such a thing?

by kenanfyi

6/10/2026 at 8:38:38 AM

Containers only got so popular as a tool for developers to make developing/deploying easier. If you want to use them as a security layer that is a completely different goal and has many highly dangerous pitfalls [1]. Just last week there was a post where people were shocked how an AI agent used docker to bypass sudo on a system. I'd imagine this could happen to most people who installed docker. So if you want to use containers for anything but easier development, you need to be much more proficient than the average user already. In that case not exposing $HOME is just a small thing on your config to-do list.

[1] https://cheatsheetseries.owasp.org/cheatsheets/Docker_Securi...

by sigmoid10

6/10/2026 at 8:58:27 AM

I see. Why this interests me is the similar stuff I have been reading lately. All these supply chain attacks regarding npm, Tanstack etc. Therefore I wanted to create a totally isolated sandbox and while considering options I have seen they all by default mount the $HOME. I needed to explicitly tell colima to not do that.

But yeah, I guess my use case is not the main use of such tools or their purpose in general. Thanks for the link, I‘ll take a look at it.

by kenanfyi

6/10/2026 at 9:16:00 AM

> Just last week there was a post where people were shocked how an AI agent used docker to bypass sudo on a system.

This was due to implicitly granting the LLM access to the host docker daemon, which has superuser privileges, not due to a "container breakout". That's arguably a very different scenario, but of course both are worth considering.

> So if you want to use containers for anything but easier development, you need to be much more proficient than the average user already.

I'd disagree. Containers, at least without granting them additional privileges such as CAP_NET_ADMIN and without write-bind-mounting sensitive host directories into the container, offer a reasonable security boundary compared to the counterfactual, despite their bad reputation.

by lxgr

6/10/2026 at 9:17:04 AM

the reason i use this (and just a container with -v $HOME:$HOME before) is to get an environment with all the command line tools i'm familiar with from debian, instead of using something like homebrew. in general, i mostly trust these with access to my home directory. a bonus is that i can throw it away and rebuild it easily if i need to.

i'd still use less permissive containers for things i don't feel comfortable installing on the host, e.g. npm.

by saljam

6/10/2026 at 8:59:37 AM

No, the whole point of machines is their external interfaces? A Linux VM with no interfaces is just a closed box wasting power doing math.

And I think I would caution Apple to consider the lessons of WSL; having shared access to the filesystem is just the bare minimum. Next is networking (and god is this a rabbit hole with WSL), people will want to access their USB devices, X forwarding, GPU passthrough..

by stefan_

6/10/2026 at 9:08:26 AM

The whole point of container machines is their isolation.

If we wanted access to all interfaces, we'd just run it locally.

We want the container as a closed box, "wasting power doing math", i.e. processing what we actually passed to it.

by coldtea

6/10/2026 at 8:37:46 AM

No, the point of using such a thing is to be able to run Linux workloads. For example, I recently used Containerization to generate trace logs from the tup test suite so that I could bring it up to relative parity on macOS. If it had complete isolation, I would have difficulty getting the modified source code into the container and difficulty getting the trace logs back out of the container. Sure, you can paper over this with bind mounts or whatever the fuck but that's annoying

by LoganDark

6/10/2026 at 9:01:09 AM

Understand. And yeah that‘s annoying. I use containers only for development and to keep my main system secure from supply chain attacks. I have almost no build tooling in my Mac anymore. No npm, no cargo, no uv. Nothing. They all live inside the container which is completely isolated.

I guess my use case is not that important for the main user of these tools.

by kenanfyi

6/10/2026 at 9:05:26 AM

I wouldn't say your use case is not important. That is a completely reasonable way to work. I just wouldn't say mounting home directories is an anti-feature. There just should be a way to turn it off.

by LoganDark

6/10/2026 at 9:23:49 AM

Well, maybe I should have used relativity unimportant. And yes there should be a way to turn them off. In OrbStack it was not possible to do that until lately.

by kenanfyi

6/10/2026 at 5:55:52 AM

This is all fine and dandy, but where are the native Darwin Jails Apple? Still scared that people will filling whole rooms of Mac Minis if you allow them to have multiple macOS containers and not only up to two fat VMs per machine?

by qalmakka

6/10/2026 at 7:27:51 AM

[Replied to wrong comment]

by jorisw

6/10/2026 at 7:28:47 AM

That's totally unrelated to what I wrote

by qalmakka

6/10/2026 at 7:28:44 AM

sandbox profiles?

by adastra22

6/10/2026 at 7:33:06 AM

macOS sandboxing is deliberately limited just enough to prevent anyone from truly implement Darwin-on-Darwin containers. People have been discussing about this for a while, see https://github.com/apple/container/discussions/611

In general I understand the rationale behind Apple's decision. They sell hardware, and there's real demand for macOS on servers to run build jobs and other Mac-only tools. Giving you the ability to run multiple containers on a single Mac would end up turning a 10 Mac Mini order into a 2 Mac Minis order for most people. Rest assured, even if it would be technically possible they'd find a way to cap it somehow via the EULA or whatever

by qalmakka

6/10/2026 at 9:10:02 AM

I doubt this insignificant statistically speaking market (compared to the overall units they move) is what prevents them.

by coldtea

6/10/2026 at 9:35:12 AM

Domino theory as applied to business, plus one should never underestimate the lengths to which a company will go to wring the last ounce of profit from a market.

by inejge

6/10/2026 at 1:13:24 AM

OrbStack works really well for me. I wonder how it’s compared to this performance wise

by blahgeek

6/10/2026 at 1:32:34 AM

(OrbStack dev here.) Instead of Virtualization.framework, we have a custom Rust virtualization stack with custom devices and protocols for things like filesystem sharing. It's a highly optimized vertically integrated stack specifically for running our Linux machines and containers.

Our biggest perf/resource gain is dynamic memory, which reduces memory usage a lot by releasing unused memory back to macOS. Nothing else supports this, including Containerization.

I gave Container Machines a try and it seems to be much closer to OCI containers with a default bind mount than OrbStack machines. It has fewer integrations and doesn't run systemd or any other normal init system, so it's hard to run services.

by kdrag0n

6/10/2026 at 9:12:05 AM

Just to give a thumbs up to you and OrbStack. I've been using it every day since the first releases, and it is one of the most stable and performant drop-in replacements that I've seen.

by rubnogueira

6/10/2026 at 4:29:07 AM

just adding a 'hell yeah: orbstack is so good' to the thread. i mainly avoid containers where i can, but when containers need to happen, orbstack is 'just enough' for me. lovely and well considered ui, stable, performant. don't need much else. thank you for your work and care!

by d3v1an7

6/10/2026 at 2:26:16 AM

Super happy orbstack customer. Just curious on your statement:

> I gave Container Machines a try and it seems to be much closer to OCI containers with a default bind mount than OrbStack machines. It has fewer integrations and doesn't run systemd or any other normal init system, so it's hard to run services.

The linked md document says:

> Real Linux services for testing. Run a database or whatever your stack needs as a system service — systemctl start postgresql works on images with systemd installed.

Was that not the case when you used container machines?

by mescalito

6/10/2026 at 2:45:57 AM

That's my bad, I used the example alpine commands and the official alpine doesn't have init. It's supported if you build an image with systemd installed

by kdrag0n

6/10/2026 at 1:47:11 AM

Apple says that `systemctl` is supported... hmm am I missing something?

"Real Linux services for testing. Run a database or whatever your stack needs as a system service — systemctl start postgresql works on images with systemd installed."

by kxxx

6/10/2026 at 1:51:27 AM

Good catch, I tried the example alpine commands and there was no init system. Makes sense if it's based on OCI images

by kdrag0n

6/10/2026 at 1:58:17 AM

Just tested it on on an OCI image with systemd and it works well. I can see the appeal of OrbStack regarding memory reallocation and will stick with it in the time being :)

by kxxx

6/10/2026 at 9:39:48 AM

What kind of virtualization do you use internally? I assume some kind of LXD fork? OrbStack really feels like a single-node Incus host.

by rahen

6/10/2026 at 2:04:01 AM

> Our biggest perf/resource gain is dynamic memory, which reduces memory usage a lot by releasing unused memory back to macOS. Nothing else supports this, including Containerization.

Wow, missed this when reviewing OrbStack. I assumed that you just used Containerization and therefore would have the same limitation.

by CGamesPlay

6/10/2026 at 1:55:10 AM

I know this is off topic, but I do thank you for your Android work, the idea and elegance of fastboot.js and that SafetyNet workaround trick was truly really cool.

by saltamimi

6/10/2026 at 1:57:18 AM

Ahh those were good times, glad you came across it :)

by kdrag0n

6/10/2026 at 7:10:15 AM

I changed over to Orbstack just for local builds and it is one of those apps that makes owning a Mac that much better.

This post reminded me to buy a license, just done it, worth it for the time saved.

by rswail

6/10/2026 at 2:11:22 AM

just dropping in to say orbstack super owns and i use it every day. huge respect to rethinking this experience, for a minute there i thought docker was just going to be the only path. i dont think ive looked back for docker since. orbstack just feels right, and damn its so fast and good with resources, and the UI is just insanely straight forward. props!

by trueno

6/10/2026 at 2:20:29 AM

We love OrbStack too! Thank you for it,

I wanted to make its VM/machine our default secure agent sandbox, but I couldn’t figure out how to isolate this VM from the host properly. This thread prompted me to find the issue though, and I saw this was recently implemented! https://github.com/orbstack/orbstack/issues/169

by TheTaytay

6/10/2026 at 2:23:20 AM

Yep! Still refining it but isolated machines now have fine-grained settings for filesystem mounts, network isolation, SSH agent forwarding, and CPU/memory/disk limits

by kdrag0n

6/10/2026 at 1:54:32 AM

I’ve been using podman on Mac. It’s been a nice fit as the container build files are identical to what I use on my fedora server. I have noticed my 2 virtual core 4 gb Linode vps runs apps faster in the same container as when run on my MacBook Air M2 16 gb. I expected some performance overhead but didn’t think it would be noticeable as it is. Overall happy with podman. How might OrbStack differ?

by jhancock

6/10/2026 at 1:59:41 AM

Having used both, it feels like OrbStack "just works" more than Podman. The main example of this is Supabase.

by thatxliner

6/10/2026 at 3:37:24 AM

When are y’all gonna support sandboxing? Preferably Docker Sandboxes?

by blackqueeriroh

6/10/2026 at 5:51:39 AM

Orbstack plays well with Pycharms BTW.

by bjt12345

6/10/2026 at 2:00:13 AM

I love orbstack, is there any code I could read on the rust side? Seems very interesting

by vsgherzi

6/10/2026 at 2:06:29 AM

I'd like to see a comparison to https://tart.run/ as well.

AFAICT it's pretty similar.

by emmelaich

6/10/2026 at 3:20:08 AM

I like orbstack in theory, but I find it hard to justify a $96/yr license fee for something that has so many open source, free alternatives. As it is, I’d rather use podman or colima

by mpeg

6/10/2026 at 6:16:05 AM

It's free for personal use, and for a company 96/year is absolutely nothing, I'd hope.

by Ghoelian

6/10/2026 at 6:42:09 AM

The alternatives are all broken in some ways is the answer, including the official paid docker enterprise.

Personally I’d rather the company provisioned me MacBook hardware with Linux. Unless Fable or some other ai ports asahi properly to modern hardware I expect to retire before this is possible, orbstack is the next best thing, available today.

by baq

6/10/2026 at 6:44:22 AM

I just wish bind mounts would be more performant/native. I get that this is probably impossible, and probably also sucks on Linux, haven't tried.

But like having containers that need file watchers like vite dev server, or frankenphp in watch mode will overload OrbStack real quick since It seems to fallback to polling instead of listening to fs events.

So I'm stuck running vite dev servers and the like on the host.

by gempir

6/10/2026 at 1:45:10 AM

I really like OrbStack and am also not sure why I'd use Container Machines over it, at the moment...

by kxxx

6/10/2026 at 2:30:02 AM

Not a full docker env, I aimed this as doing builds though you can run dockerd as an option, https://github.com/cpuguy83/crucible uses the containerization framework to run either build kitd or dockerd and wire it up to docker/buildx cli (or whatever client tooling you want to use).

The Containerization framework is a library that sits as a layer on top of the virtualization framework. So each container is its own VM.

Machine is tooling above the containerization framework to run multiple things in a container in a vm.

by cpuguy83

6/10/2026 at 7:26:11 AM

Thank you for mentioning this, I have been suffering under the yoke of docker.

by kiproping

6/10/2026 at 5:08:19 AM

Note that orbstack supports audio and usb pass through, which is super nice

by jbverschoor

6/10/2026 at 5:09:46 AM

I belong to a rare breed of very opportunistic hobby-developers that like to use MacOS but also like to use linux machines or BSDs (rpi etc) sometimes.

I can create docker-images with docker compose, or use something like colima, which this seems to be close to (that should have some advantages over docker, although my hope of circumventing W^X page protection did not pan out).

I was perplexed that the repository does not put these container machines in context. The seem to be close to colima? When should I use which option (docker, collima, container machines ?)

Maybe others wonder too but are ashamed to ask. I have no shame ;)

Thanks for any pointers

by golem14

6/10/2026 at 5:49:34 AM

Why try to circumvent W^X page protection? Some sort of self-modifying program without extra pointer indirections?

by djsavvy

6/10/2026 at 6:56:56 AM

Bad legacy apps like Xilinx ISE

by bogantech

6/10/2026 at 9:27:21 AM

Interesting from a Docker perspective, but I’m more interested in it as a sandbox for AI agents and untrusted code execution.

I wrote about that angle here: https://igorstechnoclub.com/sandbox-exec/

Feels like the spiritual successor to sandbox-exec, but with VM-level isolation.

by Igor_Wiwi

6/10/2026 at 8:19:49 AM

This appears to be an LXC-style alternative for macOS; however, unlike native LXC on Linux, this tool relies on VMs. While Docker and Podman also utilize a VM on macOS, they offer the advantage of the Docker Compose format. In my view, the ability to use YAML for declarative configuration is the most critical feature for any container tool. I have nothing against CLI tools in general, but I prefer avoiding repetitive manual commands that could be easily automated via Docker Compose or Kubernetes manifests.

by emulio

6/10/2026 at 1:05:18 AM

Will this be able to replace docker desktop an equivalents, removing the expensive Linux VM that runs alongside them?

by jaimehrubiks

6/10/2026 at 9:18:00 AM

This explicitly provides a Linux VM, which seems hard to do without providing a Linux VM.

The use case is actually the opposite of what you seem to want (i.e. running Linux containers on macOS without a Linux VM); this uses a Linux-based container implementation of macOS to provide a long-lived Linux VM that looks more like a VM itself than a container.

by lxgr

6/10/2026 at 5:37:12 AM

Linux VMs on doesn't have to be expensive!

by binsquare

6/10/2026 at 1:25:06 AM

My first thought as well, docker desktop overhead is pretty bad, would be awesome to see this land natively in DD. By my estimate this could happen, seeing as Docker has historically tried to improve performance but quickly had to accept platform limitations… would only be natural to settle DD over to containers

by usernametaken29

6/10/2026 at 1:46:55 AM

Well, you can avoid the Docker Desktop tax by not running Docker Desktop. colima is a perfectly usable implementation of Docker for macOS, without the bloat of Docker Desktop.

That said, colima still has the expensive VM that upthread is mentioning.

by deathanatos

6/10/2026 at 1:58:15 AM

OrbStack is great also

by TimTheTinker

6/10/2026 at 9:33:47 AM

until they remove the free tier and raise prices, once their user base is sufficiently reliant on it.

by CarlitosHighway

6/10/2026 at 3:28:59 AM

Postman Desktop too

by phinnaeus

6/10/2026 at 6:01:40 AM

You mean Podman Desktop?

by rawland

6/10/2026 at 6:42:26 AM

Yes, thank you iOS autocorrect.

by phinnaeus

6/10/2026 at 1:09:15 AM

It mostly removes the big shared background VM and replaces it with smaller, more isolated Apple-native VMs.

I did an experiment migrating my Podman workload to Apple's container @ https://gist.github.com/jmonster/39e14585e107dbf990a90966c0f...

TL;DR reduces ram/storage usage; minimizes it's existence

by thejazzman

6/10/2026 at 1:50:45 AM

How does that work, realistically?

> Memory defaults to half of host memory

That's the most expensive part of the whole transaction, b/c AFAIK, RAM is then dedicated to the VM. It can be swapped out, I suppose, but that's not great.

by deathanatos

6/10/2026 at 3:45:32 AM

CGamesPlay said above its balloon memory so it won’t use all that memory by default, but it can’t release balloon memory yet.

by MBCook

6/10/2026 at 2:55:32 AM

Nice, thanks for this. My plan is to swap over to Apple's containers for local dev, and keep using podman quadlets in production.

by nozzlegear

6/10/2026 at 1:31:48 AM

Others here mention it and I’m a new convert to Colima.

The pain of working around Docker Desktop is bad.

by lostlogin

6/10/2026 at 1:08:40 AM

That sure would be nice. I seem to rm -rf ~/.colima every few days.

by trollbridge

6/10/2026 at 6:58:53 AM

Just to clarify, this requires Mac OS 26 Tahoe for "container" doesn't it? So those of us holding out on Sequoia who can't stand the broken glass UI or what's called and the other undesired features need to stick to Docker desktop.

by tannhaeuser

6/10/2026 at 9:12:31 AM

>those of us holding out on Sequoia who can't stand the broken glass UI or what's called

Maybe hold 1 release back, but other than that, I don't think "holding out" on macOS releases has ever been a winning strategy.

In the end, macOS model presupposes users moving to the latest release sooner rather than later.

by coldtea

6/10/2026 at 9:35:18 AM

It seems macOS Golden Gate (the upcoming version) fixes quite a few of the problems with Tahoe, so you might consider skipping Tahoe entirely.

by niek_pas

6/10/2026 at 7:27:03 AM

[dead]

by jorisw

6/10/2026 at 2:17:34 AM

Anyone know why you would use this instead of QEMU+Lima+Colima+Docker/containerd? The latter works on multiple OSes, has a very large ecosystem of tools, images, documentation, and lets you replace pieces as needed

by 0xbadcafebee

6/10/2026 at 9:54:31 AM

From a layman's POV ("I just want to run my containers I need for dev work"), there's no point in switching to this for now. It's just cool that Apple cares enough about containers and might come up with an Apple-like built-in solution some day, this is the groundwork.

I'd stick to Colima, or Orbstack if you trust them enough to not do a rug-pull once their users are reliant on them enough to pay any amount.

by CarlitosHighway

6/10/2026 at 3:38:12 AM

The costs are startup time and image compatibility: dockerhub images don't work as machine images because container machine expects systemd

I am trying it on but its brekaing on homebrew 1.0.0. The formula puts plugins at opt/container/libexec/container-plugins/ and the apiserver looks in libexec/container/plugins/

This can be solved through a symlink or smth

by noobcoder

6/10/2026 at 5:46:24 AM

> dockerhub images don't work as machine images because container machine expects systemd

Are you sure about that? A few comments above a commenter states that they don’t run inits at all (because they ran alpine), multiple people replied that it works fine if you give it an image with an init, and they acknowledged their error.

by masklinn

6/10/2026 at 5:29:08 AM

It's funny that the system config page (https://github.com/apple/container/blob/main/docs/container-...) lists pebibytes for RAM configurations... in this day and age where buying a 16GB stick for workstation would cause me to eat instant ramen for a couple of months because my dentist needs an LLM chatbot on their page to stay competitive!

UX wise it looks kinda neat though!

by rakel_rakel

6/10/2026 at 1:38:42 AM

Is this new? I thought we had this already

In my testing (iirc) filesystem performance was not good enough to be usable with node/rust dev where lots of small files get stat-ed

update: what's new is the `container machine` subcommand. I went to test it out, but container failed to run at all for me: https://github.com/apple/container/issues/1681

by llimllib

6/10/2026 at 2:02:38 AM

Curious if you've tried OrbStack? There's always more work to do (test workloads appreciated!) but we've put a lot of effort into optimizing for small files and other common developer workloads in OrbStack's customized filesystem sharing protocol (not standard virtiofs).

by kdrag0n

6/10/2026 at 2:53:38 AM

Podman is on macOS, FWIW. Uses the existing container framework to run the machine already. Root-full or not.

by ahknight

6/10/2026 at 6:37:54 AM

Did you use their volumes for node_modules or a shared dir? I mounted the whole project directory (with node_modules) inside the container and it seems to work fine (MBA M1 8 GB RAM).

by dchest

6/10/2026 at 7:12:15 AM

Python binary wheels now have to be built for aarch64 for them to work inside the container, unless they are built using the corresponding build system while installing. It is not common for python binary libs to publish arm64 binary wheels, as most often they target amd64.

by LaFolle

6/10/2026 at 7:50:35 AM

Isn't that just expected for modern macOS devices? They have been on arm64 for 6 years now.

by KeplerBoy

6/10/2026 at 9:15:30 AM

This however is for Linux containers on macOS, and for those the deployment target are not usually/necessarily arm.

(Plus, you could always even have amd64 linux containers on macOS AS, with good performance, via Rosetta2).

by coldtea

6/10/2026 at 2:35:54 AM

Is there any reason why macOS doesn't try a WSL1 style approach? I get why that didn't fully work out for windows, but it seems like macOS being another *nix would make a lot of what was hard for windows, easy for mac. It seems like it should be possible to run most linux applications natively on macOS with few additional new APIs.

BSD actually has this already.

by cogman10

6/10/2026 at 5:52:16 AM

FreeBSD has Linuxlator because there is a lot of binary only software that was never and never will be ported to BSD, so it's necessary for them in order to avoid bleeding users away. Conversely, macOS has basically all software ported natively to it, so when you _need_ a Linux environment 95% of the time it isn't because you need $XYZ that only run Linux, but because you need a proper Linux environment with systemd, cgroups etc. Implementing that stuff on top of XNU would probably be extremely expensive and it would arguably defeat the point of having their own kernel in the first place.

by qalmakka

6/10/2026 at 2:47:30 AM

What would be the advantages over a VM infrastructure Apple needs anyway and that has a much simpler, more stable “ABI” compared to the Linux kernel?

by twoodfin

6/10/2026 at 3:00:40 AM

Potentially faster application execution along much lower memory requirements. In the case of docker, even a possibility of shared library loading further reducing runtime costs (For example, containers based on the same base image could load glibc into memory only once).

There's also simply the possibility of using linux software directly in macos without doing OS dependent changes to the software.

by cogman10

6/10/2026 at 3:49:42 AM

Yeah. But in exchange it’s a lot of work to keep up with. For GUI stuff you’re now having to have some sort of Wayland layer/driver.

Running VMs is really really easy and low maintenance demand on Apple. And it’s guaranteed compatibility.

Wasn’t compatibility what really sunk WSL1?

by MBCook

6/10/2026 at 9:18:01 AM

>for GUI stuff you’re now having to have some sort of Wayland layer/driver.

The target for this isn't GUI stuff.

by coldtea

6/10/2026 at 4:39:05 AM

> Wasn’t compatibility what really sunk WSL1?

Yes, but a big part of the problem with WSL1 was the size of the conceptual gap between POSIX and Windows NT that WSL1 had to bridge. An “MSL1” would likely have fewer problems because the gap between macOS and Linux is smaller, given they are both POSIX

The other thing Apple could potentially do, is add Linux-compatible APIs to macOS. IBM wanted to support Kubernetes on their z/OS mainframe operating system, so they implemented on it a clone of Linux namespace APIs, e.g. unshare. Then we could have macOS nodes in a K8S cluster-which might actually be useful for some people, e.g. if you have a Jenkins CI farm, the Linux nodes can run on K8S, but currently macOS nodes (which you need if you are targeting iOS or macOS) can’t, they have to be bare metal or VMs.

More Linux-macOS source compatibility would also benefit macOS by making it less work to port software to it from Linux

by skissane

6/10/2026 at 5:54:42 AM

Linux and the BSDs take APIs one from the other all of the time. The issue with having a Linux ABI is that you don't need just the few APIs you're missing, you need to implement the WHOLE Linux API and it has to be _perfect_, otherwise stuff will randomly break. I loved the original WSL, I had to use it for a time period back in the day when I was stuck on a Windows PC, but it can't be denied it was full of random bugs

by qalmakka

6/10/2026 at 3:10:17 AM

Apple containers are great for providing a sandbox to your AI coding agents

I have made it a MCP so that it's easily discoverable by all the coding agents

https://github.com/instavm/coderunner

by mkagenius

6/10/2026 at 5:25:30 AM

I've looked into replacing Lima with Apple Containers for https://runmachine.dev.

However, unlike Lima, an Apple Container is not a full VM, so you cannot SSH to it, or forward SSH-agent signatures into a machine.

So it's more of a devcontainer story, which is also a great use case. Nice to see Apple creating tooling around their VZ framework.

Edit: referential clarity.

by katspaugh

6/10/2026 at 6:28:34 AM

Most of my team's development happens on beefy desktop machine in incus containers per dev+project (so you run yourname-projname-dev). It has its own tailscale inside so you can open it like regular https website or give to another dev to check out – no need to deploy your branch somewhere, just run it. New dev onboard takes 10 minutes from zero to dev env with VSCode remote development.

I would really love if apple could give inexpensive way to run amd64 containers for situations when dev wants to use their own hardware. We've used LIMA for now, was too much of a hussle. But if there's a more native experience – would give it another try.

by k_bx

6/10/2026 at 7:14:43 AM

How is this different from Virtualbox or similar products with a shared folder with the host machine? I expected that existing virtualization tech for Macs already did that. Maybe the improvement is having nothing to configure.

By the way, is it headless or can it run a full Linux desktop? Use case: buy a Mac, uninistall whatever can be uninstalled, run the Linux VM as primary desktop forgetting MacOS and without going through Asahi and the incomplete hardware support.

by pmontra

6/10/2026 at 7:27:21 AM

it differs by lacking all the cool options that makes vmware and virtualbox good products, but apple users will praise it as a benefit

"bind mounts? I'm better without it"

by iririririr

6/10/2026 at 9:18:53 AM

Funny how confidently people can mock while knowing nothing about the specific tech discussed and the different targets. I'd google: VirtualBox vs containers.

by coldtea

6/10/2026 at 2:10:45 AM

I'm surprised they cared enough to do this. I'd still rather use Linux but MacBook value is incredible.

by osigurdson

6/10/2026 at 2:37:11 AM

I'd always rather use Linux, but sometimes your employer gives you a MacBook. I might use this tool.

by marssaxman

6/10/2026 at 5:46:08 AM

So essentially both macOS and Windows now heavily support developing using Linux on them. They can't more openly admit that they are no match for Linux in that area.

There's some clever advertising in it for Linux, if Linux was advertising.

by cromka

6/10/2026 at 6:47:01 AM

I’d argue they both admin that Linux servers are the target for a lot of applications to run on. Not to develop on.

by rahkiin

6/10/2026 at 7:21:57 AM

Enterprises would do anything to develop on Linux except using an actual Linux distro.

by plutokras

6/10/2026 at 4:50:43 AM

With the BUILD and WWDC 2026 announcements, it is the Year of Linux Containers Desktop.

Which for many folks is good enough for what they are doing, thus the status quo of desktop platforms will hardly change for current form factors.

by pjmlp

6/10/2026 at 8:32:12 AM

I looked at this last year when it came out in Tahoe: https://anil.recoil.org/notes/apple-containerisation and this looks like more of the same.

They've now added a WSL-style virtual machine layer, but there's no x86 container story (Apple's killing Rosetta) so I imagine some qemu shimming will be required.

There's still no equivalent to VPNKit or GVisor for networking so you'll be bridging I think. See: https://cacm.acm.org/research/a-decade-of-docker-containers/ for how Docker for Mac does this

I can't spot any support for dynamic memory ballooning to prevent the hypervisor from gobbling up too much memory. We've had this in Xen since forever! https://xenproject.org/blog/ballooning-rebooting-and-the-fea...

And, most obviously: NO SUPPORT FOR MACOS. This is the single feature that only Apple can do, and they're choosing not to implement it deliberately, and it's so stupid given the pains we all have to go through to implement CI for macOS. In the land of OCaml, we were forced to implement a custom ZFS snapshotter to get reasonably cost effective macOS CI for our package repository: https://tarides.com/blog/2023-08-02-obuilder-on-macos/. This was fun to build, but it sucks to have to maintain it.

Also, I'm really curious what the GPU passthrough story here is for LLMs, since the Apple Silicon -> Linux kernel support is gated on Asahi's support, but that's been lagging beyond M2 due to the efforts of reverse engineering.

Do better for your developers, Apple. This is a half-baked sweep across third-party software without addressing the core needs around your own operating system.

by avsm

6/10/2026 at 7:32:45 AM

This is pretty cool - being able to bring your own container machine image goes a long way to helping it's adoption.

I started using Colima a couple of years ago because I got bored of how bad Docker Desktop was and just started using the CLI / the "Services" tool window in whatever Jetbrains IDE I was using at the time anyway. I can't see myself moving away from it any time - having multiple profiles is an absolute winner of a feature for me there, but maybe the next time I set up a Mac from scratch I'll have a play with this.

by aspeckt_112

6/10/2026 at 7:38:00 AM

Always nice to have more options especially without third party tools

by Havoc

6/10/2026 at 3:59:46 AM

Could this allow us to use proton on mac maybe?

by vachanmn123

6/10/2026 at 4:24:27 AM

This is hilarious. Next year, the PC gamers will be saying "The best Windows gaming experience is win32 on Linux on macOS Containers".

by xd1936

6/10/2026 at 4:40:58 AM

The fastest (Geekbench 6) Windows laptop in the world is actually an M5 Max Macbook running Parallels running Windows.

by aurareturn

6/10/2026 at 6:23:59 AM

Wine works fine on macOS, there is no need for a Linux layer.

by kergonath

6/10/2026 at 4:34:11 AM

I mean at this point literally anything works better than Windows.

by Gigachad

6/10/2026 at 5:03:29 AM

Except game development, hence Proton.

by pjmlp

6/10/2026 at 5:21:42 AM

it always gets a sad chuckle out of me to hear that some native linux ports run worse than the windows version under proton. i think valve games are like that (l4d2 for example) and recently I think Hollow Knight: Silksong was like that

by asimovDev

6/10/2026 at 5:27:44 AM

I think at this point native linux ports are somewhat a thing of the past. The problem was that the ports were usually contracted out to a 3rd party and rarely updated or cared for that much. There was also the issue that they often relied on dynamically linked libraries provided by the distro rather than static linked libraries bundled with the game. So stuff that did work would break on distro updates.

The proton model has the benefit that bugs on linux can be fixed by Valve and the Wine community. While bugs in an official linux port can only be fixed by the game publisher which rarely happened. There also seems to be virtually no downsides to running a Windows game in Proton. These days I don't even bother checking the Wine DB or proton rating because unless the game is deliberately blocking linux via anti cheat, it will just work.

by Gigachad

6/10/2026 at 5:55:57 AM

The irony that without Windows there are no Linux games, eventually Linux folks will learn about OS/2 history in regards to Windows compatibility features.

Linux will stay forever a headless operating system great for embedded, server rooms and containers.

We have all limited time on Earth, and eventually Valve won't be around as it used to be, might even be acquired, sold, whatever, then what in regards to Linux gaming?

by pjmlp

6/10/2026 at 6:06:40 AM

Wine existed before Proton, Valve made it better but the project doesn't rely on Valve. Currently Linux is the best gaming experience. Zero bloat or nagware, everything just works. It's just ironic Wine/Proton ended up being the best platform for gaming on Linux. I don't think anyone expected it to run so well with virtually no performance impact.

Now with the Fex project, it might end up that running Windows games on linux on a modern ARM processor could be the best way to game going forward, especially for mobile platforms like the SteamDeck.

by Gigachad

6/10/2026 at 6:31:21 AM

The best gaming experience are Switch, PlayStation, XBox, iOS, Android, the very definition of everything just works, and no kernel drivers to worry about.

by pjmlp

6/10/2026 at 6:55:25 AM

You just listed concrete hardware (with the exception of Android). That's a category error, of course a fixed hardware with specialized software will have less inconsistencies.

by gf000

6/10/2026 at 7:17:05 AM

You would be happier if I listed the respective OSes instead?

by pjmlp

6/10/2026 at 6:25:24 AM

I don't think so. This is a VM, closer to WSL2.

Proton is based on Wine which translates Windows instructions to Linux.

Besides there's already Wine for mac.

But I would love to be wrong here.

by bel8

6/10/2026 at 8:33:00 AM

And no GPU passthrough? So colima with libkit remains the only method on Apple Silicon?

by throw1234567891

6/10/2026 at 1:48:40 AM

Wouldn’t it be nice if services like Codespaces or Coder or Gitlab would allow you to target running on their hosted/integrated platform, or let you launch that same container completely locally? Sometimes I wanna take my “remote” dev environment off-line but still benefit from the integrated UX.

by numbsafari

6/10/2026 at 1:58:57 AM

If you can express that operation in Terraform, then Coder would let you do that. First problems I can think of are connectivity from the Coder provisioner to your local machine (Tailscale? Local?), and migrating disk images if you want to actually switch a workspace between environments (local provisioner could do this, but no matter what it’ll be slow and janky).

by CGamesPlay

6/10/2026 at 2:04:51 AM

Maybe I don't understand but why doesn't Gitlabs self hosted setup work?

by jayd16

6/10/2026 at 5:58:48 AM

Why did they have to invent their own solution instead of just shipping docker or an equivalent clone ?

by harrouet

6/10/2026 at 7:45:09 AM

Isn't docker on mac os still a large preallocated linux VM that runs the containers inside itself? With this maybe you can separate them.

by nottorp

6/10/2026 at 5:01:39 AM

In the intro it mentions automatically mapping user and home dir. So host files accessible the container. Any settings to control this?

by jzer0cool

6/10/2026 at 3:01:24 AM

We have WSL at home.

by Joyfield

6/10/2026 at 2:59:43 AM

I was wondering if it's possible to have the container volume change to, say, an external drive. I currently use QMEU with qcow2 images to achieve this, works well enough.

by rickstanley

6/10/2026 at 6:07:57 AM

Also works with UTM.

by opengears

6/10/2026 at 8:02:30 AM

This blew up spectacularly when combined with Time Machine, I wonder if that’s fixed.

by rcarmo

6/10/2026 at 7:23:37 AM

I hope this brought us one step closer to being able to run our distros of choice very freely and easily on a Mac.

by shelled

6/10/2026 at 2:34:57 AM

Every time I see Apple flaunting Linux containers I can hardly consider it as anything but admitting defeat. It could easily be Darwin, if they still had the capacity.

by m132

6/10/2026 at 2:51:05 AM

Just change 30 years of internet history

by groundzeros2015

6/10/2026 at 3:54:46 AM

For what it's worth, the first web server was a NeXTcube, and NeXTSTEP was the foundation of macOS.

by al_borland

6/10/2026 at 2:53:18 AM

What is the alternative? They gave up the server market a decade ago and before that they barely actually supported it.

If they were to support darwin containers, what would be the point? Literally nobody would build to it, Linux won.

by tw04

6/10/2026 at 2:59:24 AM

> Literally nobody would build to it

because nobody does ci/cd against macOS or iOS apps right?

by riffic

6/10/2026 at 3:05:35 AM

And what is the revenue stream tied to that ci/cd pipeline they aren’t capturing today? Apple would sell less hardware in order to…?

There aren’t any app developers avoiding the Apple ecosystem because there aren’t Darwin containers. They don’t sell server hardware and by all accounts have no intention of ever reentering that space. So they’d spend a bunch of developer cycles to reduce their own revenue stream with no apparent upside beyond “goodwill” which they’ve never been overly concerned about.

by tw04

6/10/2026 at 3:25:51 AM

Correct me if I'm wrong, but by the same logic, you could also say this whole containerization framework is of no use either.

If they're investing resources into it regardless, they might at least try making something that Docker for macOS and co. haven't solved the same exact way already. Something that, due to their almost unhealthy obsession with "system integrity", only they can realistically make. Like native containers.

by m132

6/10/2026 at 3:35:39 AM

Supporting the containerization framework lets them sell more laptops to Linux devs that may have otherwise bought a Dell or hp or insert brand to run Linux natively on or windows with WSL.

by tw04

6/10/2026 at 3:52:51 AM

Containers are REALLY REALLY popular. This is a a great value add for developers on Mac who need to deal with Linux containers.

Which is a ton of ‘em.

by MBCook

6/10/2026 at 6:35:35 AM

They already support this scenario with XCode Cloud, it is only a market for those that don't want to pay Apple for it.

by pjmlp

6/10/2026 at 2:51:09 AM

[dead]

by ahknight

6/10/2026 at 3:01:58 AM

Apple set itself up for defeat in the server and developer marketplace as soon as they decided macOS was proprietary code.

Why would any serious developer use closed-source code they can't debug and modify? Especially for a production server?

It's the same reason no serious developers or hackers use macOS, like part of the point of being a developer is being able to dig into the code at any layer and debug and fix things.

by TheDong

6/10/2026 at 5:08:59 AM

> It's the same reason no serious developers or hackers use macOS

I know I'm basically taking the bait, but I guess I've not been "seriously" developing stuff for the past decade or two, which is news to me!

by bschwindHN

6/10/2026 at 3:17:53 AM

OpenDarwin was a thing at one point, with mailing lists and other infrastructure hosted by Apple.

That being said, my point isn't that Apple should absolutely focus on making a server OS again. It just saddens me how far behind macOS has fallen as they stopped caring about the fundamentals; back in the day, it would be Linux trailing behind macOS. Nowadays, you can't even have multiple routing tables on the latter, the firewall code was probably last updated in Snow Leopard, and what Apple happily shows off on WWDC is a wrapper around Linux. Something functionally equal can be cobbled up together by anyone sufficiently experienced in minutes, using just Bash, OpenSSH, and QEMU.

I really wish macOS would let me have a similar level of control over applications as Linux with namespaces, without me having to do all the heavy lifting.

by m132

6/10/2026 at 6:08:15 AM

> Nowadays, you can't even have multiple routing tables on the latter, the firewall code was probably last updated in Snow Leopard

Apple uses OpenBSD's Packet Filter [1]; I doubt multiple routing tables are a problem. Back in the Snow Leopard days, it was FreeBSD's IPFW, which is also no slouch.

Whatever a firewall can do, PF can do it.

You can also get a nice GUI for PF [2].

[1]: https://www.openbsd.org/faq/pf/index.html

[2]: https://www.murusfirewall.com/murus/

by alwillis

6/10/2026 at 9:22:50 AM

Yes, I meant pf. Indeed, it was there in the source tree in 10.6 but they only flipped it on it in release builds in 10.7. My bad. Either way, it has hardly changed since then, while the OpenBSD upstream continued to progress.

> I doubt multiple routing tables are a problem.

The lack of them is a limitation for me (complex VM + VPN setup), which requires me to do pretty unholy static routing and address rewriting with pf.

I think even Apple has come across this; they added "scoped routing" (which IMO is a hacky workaround providing some of the functionality you'd get with multiple routing tables) just before iOS shipped with MMS support. Android, for comparison, uses Linux's routing policies and tables to send and receive MMS.

by m132

6/10/2026 at 5:04:25 AM

Apparently game, desktop and app devs aren't serious.

by pjmlp

6/10/2026 at 3:16:15 AM

No offense, but serious developers don’t think this way at all.

by vehemenz

6/10/2026 at 4:30:41 AM

For server side, which I believe is the context here, Linux and open source are king.

Even Microsoft gave up on Windows and just runs Linux most things except niche cases. Heck, even SQL Server which is expensive piece of machinery got ported to Linux and that's the default target now in their docs.

With that said, one can't deny Apple's success on the b2c side of things so it feels wrong to call their strategy a failure.

by bel8

6/10/2026 at 6:33:38 AM

Except the cloud isn't open source, the ones that matter to developers that is.

Which is why so many projects get burned with their license choices.

by pjmlp

6/10/2026 at 6:58:27 AM

I don't see how this comment is relevant to parent's point. Sure, cloud is proprietary. But it is Linux for the vast majority.

by gf000

6/10/2026 at 7:13:58 AM

Which is an Pyrrhic victory, when Linus and other founders are long gone, most of this generation actually, what will subsist are proprietary forks, just like what happened with UNIX System V.

by pjmlp

6/10/2026 at 1:19:57 AM

With colima I can run AMD64 (x86) Linux containers in my Arm64 too. I think this is strictly for Arm64 Linux VMs, or is there some way to run x86 with this too?

by a1o

6/10/2026 at 8:38:48 AM

You can run amd64 binaries inside an aarch64 Linux virtual machine. Although they're not supporting Rosetta for macOS apps from macOS 27, the Rosetta support in Virtualization Framework will remain.

by jdub

6/10/2026 at 5:21:45 AM

What’s the performance when you do that?

by cpach

6/10/2026 at 1:30:03 AM

Rosetta should be supported

by frizlab

6/10/2026 at 4:49:44 AM

Not for long!

by whycombinetor

6/10/2026 at 5:08:17 AM

Very unlikely to lose support for Rosetta for Linux. Maybe just Rosetta 2 for mac apps.

by commandersaki

6/10/2026 at 7:13:39 AM

Im running Multipass on M1 for full linux VMs. Are container machines better?

by sdevonoes

6/10/2026 at 7:26:41 AM

Isn't multiphase is Ubuntu only?

by m1keil

6/10/2026 at 7:57:42 AM

It used to be. Any image is allowed now.

by sdevonoes

6/10/2026 at 1:42:28 AM

Would be cool if you can redirect USB devices to the VM.

by commandersaki

6/10/2026 at 1:45:22 AM

We just released this in OrbStack :) https://docs.orbstack.dev/features/usb

Blog post soon

by kdrag0n

6/10/2026 at 3:41:28 AM

What happened to Orbstack for like 9 months until earlier this year? Suddenly everything went silent for a bit and I was pretty concerned. Glad y’all are back!!!!

by blackqueeriroh

6/10/2026 at 3:02:04 AM

Thank you for sharing this - I looked into OrbStack a few months ago, and this was the reason I didn't use it (as my primary purpose was to have an external wifi adapter for wifi pwnage).

by calebm

6/10/2026 at 2:17:40 AM

Yeah I find this useful for redirecting storage/sdcard*, so you can format linux filesystems or use other tools.

* need a usb sdcard reader for macbook pro cause the builtin is not usb)

by commandersaki

6/10/2026 at 3:10:49 AM

We're working on block device passthrough for the builtin SD reader.

by kdrag0n

6/10/2026 at 3:46:53 AM

I've successfully tinkered with USB/IP with Apple containers, but it does require loading a custom kernel (which they make pretty easy, thankfully). On the host side, macOS also doesn't make it easy to unload a driver that attaches automatically.

by rgovostes

6/10/2026 at 1:49:51 AM

Agreed! There's some good improvements around Accessory Access in virtualization framework this year also - checkout: https://developer.apple.com/videos/play/wwdc2026/224/?time=2...

by egernst

6/10/2026 at 2:30:33 AM

I wonder if the custom virtio can be used to support attaching the built-in sdcard readers on macs which aren't exposed as usb.

by commandersaki

6/10/2026 at 5:10:00 AM

"LXC" for macOS?

by zekrioca

6/10/2026 at 6:11:51 AM

Is this going to be good for AOSP builds on Macs?

by beemboy

6/10/2026 at 4:38:07 AM

I know its not going to be there but wish we had Windows as well.

by CSDude

6/10/2026 at 5:35:44 AM

Install Windows 11 ARM under the macOS "UTM" App. This lets you run x86 Windows programs on Apple silicon.

by Cadwhisker

6/10/2026 at 6:00:10 AM

Sounds like toolbox or distrobox for Mac!

by konaraddi

6/10/2026 at 6:16:09 AM

Can Podman support these eventually?

by ShinyLeftPad

6/10/2026 at 2:00:46 AM

WSL-like implementation on macOS?

by sachinjoseph

6/10/2026 at 1:07:11 AM

Would be nice if they also support Intel based macs, what prevents?

by namegulf

6/10/2026 at 1:37:34 AM

Apple won’t support them with MacOS 27, and it seems they announced this tool as part of this year’s WWDC.

Basically: they’ve moved on.

by MBCook

6/10/2026 at 1:19:44 AM

Allocation of a finite amount of engineering resources.

by danhon

6/10/2026 at 1:22:33 AM

And a legitimate business interest to further incentivize the adoption of Apple Silicon devices. Same with Rosetta deprecation after macOS 27.

by joshuat

6/10/2026 at 1:40:03 AM

> a legitimate business interest to further incentivize the adoption of Apple Silicon devices

Apple has never been about supporting legacy platforms with new features. And with over a quarter of revenue and two fifths of Apple's gross profits coming from services, one could argue the incentives run either way.

by JumpCrisscross

6/10/2026 at 3:08:01 AM

Sure, but to what extent?

Enterprise ARM servers are still a niche product, and so are the ARM developer machines running Linux or Windows. Until this significantly changes, Apple will have to provide good x86 interop - or lose the developer market entirely.

Forcing people towards Apple silicon is of course an attractive approach when targeting the large portion of the market using their MacBooks as Facebook browsing machines, but (especially with the new MacBook Neo) what's going to happen when a large portion of the market for high-end MBPs disappears because it turned from the default no-brainer into a liability?

by crote

6/10/2026 at 3:31:28 AM

> Until this significantly changes, Apple will have to provide good x86 interop - or lose the developer market entirely.

I'm very, very skeptical of this analysis. Certainly "entirely" is hyperbole.

by macintux

6/10/2026 at 6:23:54 AM

That’s a joke right? I’ve been developing software deployed on x86 servers on ARM Macs ever since they were released.

by solarkraft

6/10/2026 at 2:19:01 AM

Rosetta 2. Rosetta was for Intel to emulate 68k, now if you could get Rosetta 2 to run under Rosetta, then you could run 68k, on an ARM, and if you could get the apple ][ emulator...

by ForOldHack

6/10/2026 at 3:14:03 AM

Rosetta 1 was for emulating PPC not 68k

by weikju

6/10/2026 at 8:42:02 AM

The underlying Virtualization Framework works on Intel Macs, but they'll miss out on new features landing in macOS 27 and beyond.

by jdub

6/10/2026 at 1:20:20 AM

[flagged]

by teaearlgraycold

6/10/2026 at 2:50:56 AM

I'll defend, not cringe for everyone.

Daily driver is a 6yo, 32Mb mbp and it might not scream like an M5 or have the miraculous power draw of an M5, it gets my job done.

One nice thing is x86 containers run natively: I run most of my $work landscape which is 40 or 50 k8s pods on top of Kind, which is itself a plain container. That mirrors my prod. That plus slack, zoom, ff with scores of tabs, etc. all while building rust and playing music.

by imglorp

6/10/2026 at 3:54:58 AM

That is a far more useful reply than the GP comment. If they had stated something similar I don’t think they would’ve been downvoted.

by MBCook

6/10/2026 at 4:58:11 AM

Poe's Law and all that, but I was trolling/shitposting.

by teaearlgraycold

6/10/2026 at 2:23:39 AM

More power to ya!

by ncr100

6/10/2026 at 1:29:28 AM

cringe is cringe

by Brian_K_White

6/10/2026 at 5:16:09 AM

Just curious, Apple seems to copy orbstack.. haven’t they made an offer to acquire you guys?

by jbverschoor

6/10/2026 at 5:31:08 AM

Yeah but sitting in the tweak circles just to gather personal data about people to make them lose their minds is no bueno. Otipolfueriborsklineypoo

by itsneulook4

6/10/2026 at 2:35:21 AM

Is this similar to what cygwin was for windows? Could this be an alternative to homebrew?

by t1234s

6/10/2026 at 5:36:08 AM

What FS mounts the Mac drives into the Linux container ?

by tonymet

6/10/2026 at 5:21:39 AM

It was unclear to me, is this a native replacement for docker? I like docker (on mac) but its quite the resource hog.

I usually run like a db, redis, maybe something like rabbitmq/zeromq and have a app that uses these services (makefile/docker-compose).

I would love to switch if this in fact is a lightweight replacement.

by phplovesong

6/10/2026 at 6:59:19 AM

On the one hand yes, on the other hand there are already multiple lighter alternatives to docker on mac.

by masklinn

6/10/2026 at 3:42:17 AM

I saw the video on this this is distrobox basically for Mac. It’s very cool. Seamless with your local files and the container. I’m very keen to try it.

by gigatexal

6/10/2026 at 2:57:09 AM

Can macOS be run as a container machine on macOS?

by michaelsbradley

6/10/2026 at 3:41:55 AM

Yes

by blackqueeriroh

6/10/2026 at 3:55:58 AM

Yep. For a few years. And they keep enhancing it too.

It’s the only legal way to do so, due to the software license on MacOS.

by MBCook

6/10/2026 at 2:01:41 AM

darwin containers when?

by riffic

6/10/2026 at 1:03:54 AM

looks like apple wrote a native docker in swift

you can now run linux containers on your mac

... but it could be better.

what about (totally contrived):

  FROM apple/macos:10.11.6

  RUN xcodebuild -project myapp.xcodeproj -scheme MyScheme -configuration Release

by m463

6/10/2026 at 1:07:28 AM

Nice, but expect to page through a few pages of ToS during the build

by webXL

6/10/2026 at 1:30:55 AM

lol

  ENV XCODE_FRONTEND=unattended
  ENV XCODE_LICENSES=accept,firstborn,applepay,appleid=sjobs@me.com

by m463

6/10/2026 at 1:08:22 AM

Close - but it would be more like this:

  services:
    macos:
      image: dockurr/macos
      container_name: macos
      environment:
        VERSION: "15"
(And indecently slow.)

by trollbridge

6/10/2026 at 1:33:50 AM

It would be wonderful if this ran on older versions of macOS, but according to the README they only support 26.

by windowliker

6/10/2026 at 3:07:11 AM

you do not understand... Not run on, run IN :)

I'm saying the older version of macos could build/run INSIDE the container

just like on a ubuntu 24.04 system you can do:

  FROM ubuntu:16.04
or

  docker run ubuntu:16.04 
and though I haven't tried it, I believe docker can do arm in x86 using an emulator (like rosetta)

by m463

6/10/2026 at 3:57:10 AM

You can already run older versions of macOS inside a VM on macOS.

So it seems like in theory that should be doable if someone just made the container images right?

by MBCook

6/10/2026 at 1:30:25 AM

i wish!

by jadar

6/10/2026 at 5:23:36 AM

[flagged]

by lzwjava

6/10/2026 at 8:55:25 AM

[dead]

by Lapsa

6/10/2026 at 1:22:58 AM

[flagged]

by sourcegrift

6/10/2026 at 2:20:13 AM

macOS only needs to support the hardware it ships on, so of course Linux would have wider hardware support, but that doesn’t really matter in context. The bigger question is what hardware to people actually want? I see most people drool over Apple hardware while not finding any suitable equivalent for the PC that they can install Linux on.

Framework is trying to close that gap with their new release, but we’ll have to see how it is once people get their hands on it. I think it also comes at a price premium. There is always the Thinkpad route, but Lenovo burned just about every bridge with me a decade ago with things like Superfish. Where is the premium Linux laptop OEM that people can trust? Last I heard System76 was just rebranding Clevo hardware. What are people using? Dell? HP?

by al_borland

6/10/2026 at 1:25:09 AM

Sadly, Linux is much much less secure.

by hollerith

6/10/2026 at 1:29:05 AM

This claim is so absurd that I need some sources.

by pixelatedindex

6/10/2026 at 1:56:07 AM

The person you replied to is right, the "security" of Linux might as well be nonexistent compared to macOS and especially iOS/Android. Even the developers of Secureblue (https://secureblue.dev/) state that despite their hardening and mitigations Linux still lags far behind macOS (and possibly Windows) security-wise. The only Linux derivative that has proper security is Android, and even better GrapheneOS.

https://privsec.dev/posts/linux/linux-insecurities/

https://madaidans-insecurities.github.io/linux.html

I also commented here on Linux phones, the same can apply to Linux as a desktop OS: https://news.ycombinator.com/item?id=46997397

Also on top of that Linux/Windows laptops also lack the hardware-backed security that Macs and to an extent some Chromebooks have.

by armadyl

6/10/2026 at 1:44:40 AM

Linux is easier to misconfigure. Macs resists being misconfigured insecurely. At their tightest, I'd say neither is fundamentally more insecure than the other. (The exception would be M5-based Macs, which come with MIE. Though that isn't a macOS vs Linux thing per se.)

by JumpCrisscross

6/10/2026 at 1:56:47 AM

This is incorrect macOS is fundamentally more secure than desktop Linux operating systems and it isn't particularly close.

No amount of Linux hardening will get a system even close to an M-chip Mac. Software insecurities aside, desktop Linux OS systems have almost none of the hardware-backed security benefits that Macs do.

by armadyl

6/10/2026 at 2:01:10 AM

At some point, lack of security becomes a feature. A fully secure, locked-down, T2 attested macOS is able to be controlled not just by Apple, but by increasingly evil governments, with no recourse available to users.

by TimTheTinker

6/10/2026 at 2:07:21 AM

Conversely, a Linux system with no verified boot can be easily tampered with without the user detecting it by people lower than the government such as casual hackers. So in a world where your government is going crazy, you're opting for an operating system that can be penetrated with relative ease (e.g. with persistent root malware) both by a non-government hacker on top of a state backed one.

by armadyl

6/10/2026 at 2:23:24 AM

I'd also guess it's much harder to securely source components for a Linux build in the way Apple is able to.

by JumpCrisscross

6/10/2026 at 3:05:27 AM

It's not really about supply chain security it's about the hardware itself. PC manufacturers in general just can't keep up since they don't have full control/integration over the hardware stack like Apple does. Also CPU, secure element etc security is limited but Qualcomm is catching up pretty quickly I believe if they aren't there already. We won't talk about Intel and AMD. But that's beyond my knowledge so I can't say anything too specific that's just what I have from general knowledge I'm sure someone will jump in with additional info if needed.

I don't think Apple is particularly any more secure against the US government than Intel is with supply chain vulnerabilities but I have nothing to back that up with aside from vibes.

by armadyl

6/10/2026 at 3:07:03 AM

Security by obscurity worked quite well

by dvhh

6/10/2026 at 3:27:16 AM

so basically dockers

by xiaodai

6/10/2026 at 2:56:00 AM

haven't we had hypervisor.framework for like years now?

by jwlake

6/10/2026 at 5:28:03 AM

that thepolfus and the Otis and the bors and the alschweid and pretty much anyone in old the the gs gangstalk or just getting people info to sit in the same room as them to try and makr them go crazy deserve to brave hart quartered

by itsneulook4

6/10/2026 at 5:50:18 AM

try unplugging your keyboard and then plugging it back in

by khazhoux

6/10/2026 at 1:40:43 AM

I found it hard to believe I didn’t have a simple way of staying safe by installing an arbitrary application in a sandbox on macOS. (Restoring using Time Machine doesn’t count! :) )

This is a step in the right direction but requires any given developer’s buy-in first, right?

by Barbing