init-snapshot VS bocker

At CodeSandbox we use Firecracker to run our VMs (more info here: https://codesandbox.io/blog/how-we-clone-a-running-vm-in-2-s...).

To answer the questions:

> what version of the kernel do you use (the github page says 5.10 but isn't that quite old?)

Right, they have tested with 5.10, but it also works with higher kernel versions. Our host currently runs 5.19 and we're planning to upgrade to 6.1 soon. The guest runs 5.15.63, we use a config very similar to the recommended config by FC team (it's in the FC repo). It's important to mention that we had to disable async pagefaulting (a KVM feature) with more modern kernel versions, as VMs could get stuck waiting for an PF resolve.

> What do you use to build the 'micro' images

We created a CLI that creates a rootfs from a Docker image. It pulls the image, creates a container and then extracts the fs from it to an ext4 disk. For the init, we forked the open sourced init from the Fly team (https://github.com/superfly/init-snapshot) and changed/added some functionality.

> How do you keep timesync of you're not using a timesync daemon?

IIRC we expose the time as a PTP device (handled by kvm) and run phc2sys to sync the time in an interval. Firecracker has some documentation on this, where it recommends chrony. It can also be done with vsock, but it would be more manual.

> Handle kernel and app logs without adding an log daemon, and same through vsocks, etc?

The init forwards stdout/stderr of the command it runs to its own stdout, which Firecracker then logs out by itself. A supervisor reads these and writes the logs to files.

Unless they’ve changed things, there is no containerization within the VM a la kata. They run their own custom init inside the VM and use it to start the entry point. https://github.com/superfly/init-snapshot is the source.

Jerome wrote our init in Rust, and, after being cajoled by Josh Triplett, [we released the code (https://github.com/superfly/init-snapshot), which you can go read.

> Fly.io transforms container images into fleets of micro-VMs running around the world on our hardware.

Oh boy!

> None of us have ever worked for Google, let alone as SREs. So we’re going out on a limb

Oh.... boy.

> We spent some time scaling it with Thanos, and Thanos was a lot, as far as ops hassle goes.

You know, they have these companies now, that will collect your metrics for you, so that you don't have to deal with ops hassle.

Holy shit. I see they even wrote their own init... in Rust. Yes, the thing that is normally a shell script, is now a compiled program in a new language, that mostly just runs mkdir(), mount() and ethtool(). (https://github.com/superfly/init-snapshot/blob/public/src/bi...)

A (joke?) one-liner I came up with while thinking about solutions to centralized container management across multiple SSH hosts. Shame on me.

The name is inspired by bocker [0], albeit this doesn't re-implement docker-compose in bash, I found it to be fitting enough.

I'd love to see someone come up with a smarter and/or shorter way to do this.

[0] https://github.com/p8952/bocker

When I did a talk about docker I also wanted to show a bit of what it does under the hood without going through all the layers and without too much details. This ~120 lines of shell script is really good in providing just an intro into what's needed for containers: https://github.com/p8952/bocker/blob/master/bocker

i tried it and like the concnpt, but until it can be launched via a systemd userspace service (without previously manually booting it) among other problems i will keep using docker (or bocker)

> But you couldn't reimplement podman in a few hundred lines of code.

You don't even need a few hundred: https://github.com/p8952/bocker

And then there's 'dokku' which IIRC, started as a bash version of Heroku.

> Not all ideas have the same quality.

They really do. I've heard all kinds of things in my career, but almost none I would want to dedicate a portion of my life building. Not because they are bad ideas or won't work, but because of the person with the idea or it just didn't interest me. Those people went on to be moderately successful (like hundreds of millions worth) but I'm glad I wasn't on that ride.

Bocker is in this same category...docker clone in bash that's helpful in seeing what's really happening underneath with nsenter, namespaces, network bridging, cgroups, etc.

https://github.com/p8952/bocker

Docker implemented in around 100 lines of bash: https://github.com/p8952/bocker

This is the most mindblowing example for enterprise security teams that think Docker is a new threat on a single tenant Linux host.

No, buddies, all this stuff is already there. If you were fine with your visibility before*, you're still fine. Go find a real problem while we play with our developer dopamine.

* NARRATOR: They shouldn't have been.

Bocker[1] does a reasonably good job of showing the value of Docker was mostly in Docker hub.

[1] https://github.com/p8952/bocker

There is a cool project I've seen called "bocker" (https://github.com/p8952/bocker) which is something of a proof of concept of implementing Docker with bash, which speaks a bit to how Docker is indeed in many ways an amalgam of lower level primitives (such as chroot as you mentioned). Pretty neat!