Chameth.com - posts like reproducible-builds-docker-images but not why-you-should-be-using-https

Reproducible Builds and Docker Images

2022-02-18T00:00:00Z

XKCD 2347: Dependency

Reproducible builds are builds which you are able to reproduce byte-for-byte, given the same source input. Your initial reaction to that statement might be “Aren’t nearly all builds ‘reproducible builds’, then? If I give my compiler a source file it will always give me the same binary, won’t it?” It sounds simple, like it’s something that should just be fundamentally true unless we go out of our way to break it, but in reality it’s actually quite a challenge. A group of Debian developers have been working on reproducible packages for the best part of a decade and while they’ve made fantastic progress, Debian still isn’t reproducible. Before we talk about why it’s a hard problem, let’s take a minute to ponder why it’s worth that much effort.

On supply chain attacks

Suppose you want to run some open-source software. One of the many benefits of open-source software is that anyone can look at the source and, in theory, spot bugs or malicious code. Some projects even have sponsored audits or penetration tests to affirm that the software is safe. But how do you actually deploy that software? You’re probably not building from source - more likely you’re using a package manager to install a pre-built version, or downloading a binary archive, or running a docker image. How do you know whoever prepared those binary artifacts did so from an un-doctored copy of the source? How do you know a middle-man hasn’t decided to add malware to the binaries to make money?

Even worse: if the software you’re trying to use includes any dependencies, you have the same issue of trust with them. Maybe your supplier isn’t compromising the software, but that doesn’t mean their supplier isn’t. The beauty-cum-horror of a supply chain attack is that it can target the weakest link anywhere along the supply chain. Even if there aren’t any binary files involved, dependencies can still be attacked: what if npmjs.com or proxy.golang.org or github.com return a different version of a dependency-of-a-dependency when the request comes from your IP address? It doesn’t even need to be a modified dependency, it could be a perfectly un-tampered, properly signed copy of the source, just from an older version with a known vulnerability.

Enter stage left: reproducible builds, here to save the day! If the build process is reproducible then you - or anyone else on the internet - can perform the same build on the same source and validate the output has the same checksum or hash. If Debian publish a binary package and an independent re-builder comes up with the exact same build artifact, there’s a reasonably good chance that the build is good. An attacker would have to compromise both the build machine and the re-build machine to do anything nefarious. The more re-builders there are, the less feasible a supply chain attack is.

So why isn’t software just reproducible?

Compilers

As a bit of an experiment, I asked some friends to run the following for me and report the answer:

echo -e "#include <stdio.h>\nint main() { printf(\"Hello\"); return 0; }" | \
  gcc -x c -o hello.out - && \
  sha256sum hello.out

This compiles a super-simple hello world program and then prints the SHA-256 hash of the resulting binary. Here are the results:

Hash	System	GCC
1f62feab5a06861dc575201d807781926d1ae49fb113da018fde8b670a1346f7	Arch	11.2.0
b8e6f2c7082be69f65ffa5e7a3d749eb47866a1b2e1ec19efb63cc59a8b160cd	Debian	8.3.0
cbad2e47a22c234b5e7fa55e029a8db4d64ac7a962e2176bd2e1373d78954088	Debian	8.3.0
e0f6bbc13b29fea8cfa2a975ba4661e781323298aec166c8311d342e6f93c4a6	Alpine	10.3.1
e379156895e06c7a0bf18ac4d648860edcb2655576b0ab9fab172bd6c8b92075	Debian	10.2.1
7ffdaee4eb64e016b89dc5e54d2c8eebab3cebafe2c7aa97de627b5972ecea46	Debian	11.2.0
8ae52cc166743b6ae1eb3e14179ef33de5061a04237f8f97088c896c41a2f698	Arch	11.1.0
8ae52cc166743b6ae1eb3e14179ef33de5061a04237f8f97088c896c41a2f698	Arch	11.1.0

As you can see, there are barely any duplicates. Even the same version of GCC on the same OS sometimes produces different results. And this is the most basic program I could write! Differences arise from the compiler version, the build flags, the libraries installed, and a whole host of other factors. If you compile a Go application instead of a C one, then by default the compiler will include debug information in the binary. This includes the full path to the source file on disk, so building a project in /home/chris/ will produce a different binary to building the same source in /tmp. Future versions of Go are also going to stamp in other meta-data such as VCS info, so building inside and outside a Git repository will produce different binaries.

Iteration order

So we’ve pinned our build environment, we’re manipulating timestamps when adding files to archives, now what? Imagine part of the build process involves looping through all the files in a directory and doing something. What order do these files get iterated in? Well, that very much depends on the filesystem and perhaps when the files themselves were created. To ensure this is reproducible we need to explicitly sort any such operation so that it’s always consistent. This iteration could be happening in a tool that’s called by another tool that’s called by a build script, so the fix isn’t necessarily straight-forward.

Side note: a bug war story

I’ve personally been victim to this kind of non-determinism. I was working on an Android app, and committed a new test that worked fine on my machine, and worked fine on the CI server. But it failed consistently for a colleague.

We both did fresh checkouts of the source, and ran the tests. Mine passed, his failed. He sent me an archive of his checkout in case there was something weird going on there, and the tests passed on my machine. We compared hashes of our checkouts, and they were the same. It was obviously environmental somehow, but everything else worked fine, and the build system went to great pains to ensure things were the same.

After a lot of debugging, I worked out that his test was running with a different version of a library to me, despite the libraries being defined in the build files and the build files being identical. After even more debugging it turned out there were two versions of the library on the classpath, and the ordering of them was different between my machine and his.

The actual issue turned out to be that the build tool generated the classpath by iterating over the library files, and that iteration was done in order of file creation time. The two libraries were added at different points in the project history, so the creation time in your local cache depended on which versions of the app you’d built in the past. With no cache everything worked as expected but there was a slim range of commits where only one library was in use, and if you had run the tests during that period your cache was effectively poisoned.

We fixed the issue by excluding the older version of the library (which was being pulled in as a transient dependency), and filed a bug against the build tool to make the classpath properly deterministic. I think that stands as the most difficult to diagnose bug I’ve ever dealt with.

Interestingly, if you iterate over a map in Go, the iteration is deliberately non-deterministic. That’s an attempt to defeat Hyrum’s Law and prevent developers from relying on whatever the current behaviour happens to be. This actually makes it easier to make things reproducible as the problem is loud and in-your-face, rather than subtle and hard to spot.

Other sources

There’s an awful lot of other places that non-determinism can come from. If the app pulls in dependencies, their versions have to be pinned, otherwise your build changes depending on the latest release of that dependency. If the build process pulls any information from a website, it’s liable to change. Hopefully the website is under your control so that you can version the resource and pin that version. Obviously, anything to do with dates or the current user will probably cause problems. Timezones and locales can cause subtle differences.

What about Docker?

Docker comes with some good and some bad points for reproducibility. The biggest advantage is that it inherently completely describes the build environment; it should work exactly the same from one system to another, even across different OS families. The biggest drawback is it sprays timestamps around like no-one’s business. Each layer in a container image is a .tar.gz file, meaning each file within it is timestamped as discussed above. Making an image involves a lot of copying of files around, so these timestamps invariably end up causing reproducibility issues.

Even worse than timestamps in the filesystem, the image format also contains some meta-data that includes the timestamp at which each layer was built. That means even if you go out of your way to set the timestamp of every single file in your image, the image itself will be different every time you rebuild it. There is no way to deal with this in Docker, which is a very sad state of affairs. Fortunately, Buildah provides a --timestamp flag for its build commands; this not only sets the layer timestamp but also the creation timestamp of any file within the layer.

The other major issue that affects Docker images is the pinning of packages pulled in by package managers. An awful lot of images are based on Alpine or Debian derivatives, and use apk or apt to install dependencies. These need to have a version specified as otherwise the package manager will just pull in the latest at the time of the build. But this isn’t quite enough: you also need to pin the version of any packages that they depend on, recursively. This means flattening the entire package hierarchy and installing all the packages explicitly and with pinned versions.

One more wrinkle in the package management space is that Alpine don’t keep old packages in their main repositories. If you have a Docker image with pinned alpine packages in, it will stop building if the package is updated. This isn’t necessarily fatal to making a reproducible build — as long as it’s reproducible for its useful lifetime, I don’t really see an issue.

Honestly, though, the biggest issue with making Docker images reproducible is getting people to care. Dockerfiles are a relatively new way of packaging software, and there’s no centralised organisation like you find with Linux distributions. There are enough challenges that most casual packagers aren’t going to bother, and no real incentive for them to. That won’t stop me trying, though!

Artisanal Docker images

2022-02-05T00:00:00Z

Artisanal containers…

I run a fair number of services as docker containers. Recently, I’ve been moving away from pre-built images pulled from Docker Hub in favour of those I’ve hand-crafted myself. If you’re thinking “that sounds like a lot of effort”, you’re right. It also comes with a number of advantages, though, and has been a fairly fun journey.

The problems with Docker Hub and its images

Rate limits

For the last few years, I’ve been getting increasingly unhappy with Docker Hub itself. Docker-the-technology is wonderful, but Docker-the-company has been making some rather large missteps. The biggest and most impactful of these has been introducing “pull rate” limits. At the time of writing, if you want to just pull a public image without logging in then you are limited to 100 pulls every 6 hours. If you log in then you’re limited to 200 pulls per 6 hours, but it’s account wide. This might seem like a big enough number, but I repeatedly hit it and there is no way to actually audit what is causing it. I have various containers that may all pull images at arbitrary times (e.g. continuous integration build agents), and the only information you get back from Docker Hub is the number of pulls remaining.

Obviously, I could start paying Docker Hub for a “Pro” plan. That gets you 5,000 pulls per day for $7/month. The downside is that every docker client would have to be authenticated, which presents a fair annoyance in terms of credential management. I also don’t really like how they positioned the service as a public utility with special treatment in the docker software, and then start tightening the ratchet to make money.

“Bad” images

I’m fairly opinionated about what a container image should look like: most importantly it should run just a single process, and only include the bare minimum dependencies required for that. Other people think differently, and it’s very hard to tell at a glance whether an image on Docker Hub contains just the application you want, or whether it also bundles MySQL, Redis, Elasticsearch, and a partridge in a pear tree. Some people want that kind of thing, but I really don’t. It’s also very hard to tell whether an image is officially endorsed by the upstream project, and where the source Dockerfile is. This used to be better because most projects used Docker Hub’s automatic builds, but they’re now a “pro” feature.

I quite often found that I’d be looking for an image for X, and there would be 5-10 images from different users. None of them looked official, some of them were out-of-date, some bundled the kitchen sink. Even when one looked good, it’s a bit of a gamble whether the author is going to keep it updated or not.

Doijanky

The rate limits and other problems were annoying, but they weren’t really annoying enough to force me to do anything about it. The straw that broke the camel’s back came later: I was looking at the official golang images, and noticed that all the tags were pushed by a random user account called “doijanky”:

An ‘official’ Docker Hub image pushed by user ‘doijanky’

I, perhaps naively, assumed that official images were built on Docker Hub’s own infrastructure. Why would all the Golang images be attributed to this user? Checking out their profile, they’re simply identified as a “Community User” like everyone else, with no repositories of their own. The only thing in the profile is their homepage, which is a link to a Jenkins dashboard: https://doi-janky.infosiftr.net/. It appears legitimate: “Infosiftr” are a container consultancy and the dashboard is linked to from the README in the official images git repository, but I find it baffling that they’re using third-party infrastructure and a normal user account (with a dubious name) to push these images. There doesn’t seem to be a good way to verify what you pull corresponds to the Dockerfile it came from; if infosiftr wanted to inject something into the build they could happily do so, and who knows how good their infosec posture is? If someone got access to the “doijanky” account, how long could they upload malicious images before someone noticed?

This little roller-coaster ride from “are all the official images compromised?!” to “oh, no, they’re not, it’s all just awful” finally convinced me to look at building my own images from scratch.

The implementation: templating with contempt

One of the big issues I needed to tackle was how to deal with updates. I didn’t want to have to go and edit a file every time some minor release was made of some software, or every time there was a security vulnerability in a common library. The official images use a shell-scripting based system to check for updates and generate Dockerfiles, I decided to do something similar but with Go templates. The result is a tool called contempt. It takes a template like:

FROM {{image "golang"}} AS build

ARG TAG="{{github_tag "example/project"}}"

RUN apk add --no-cache \
        {{range $key, $value := alpine_packages "git" -}}
        {{$key}}={{$value}}\
        {{end}}; \
# ...

Contempt has support for getting information from a variety of sources. In this case, it’s getting the latest digest of another Docker image, the latest tag from a Git repository, and the latest version of an alpine package and all its dependencies. The resulting Dockerfile looks something like this:

# Generated from https://github.com/csmith/dockerfiles/blob/master/miniflux/Dockerfile.gotpl
# BOM: {"apk:brotli-libs":"1.0.9-r5","apk:busybox":"1.34.1-r4", <snip> }

FROM reg.c5h.io/golang@sha256:ac8fa5f4078b0a697796b5d741... AS build

ARG TAG="2.0.35"

RUN apk add --no-cache \
        brotli-libs=1.0.9-r5\
        busybox=1.34.1-r4\
        # <snip>
        pcre2=10.39-r0\
        zlib=1.2.11-r3\
        ; \
# ...

I’ve cut out the longer parts for readability. You can see that it pins all versions of the alpine packages in use, as well as the base image. This ensures that if you build from the same Dockerfile at a later time it will build the same image (or will fail entirely, as Alpine doesn’t keep their old packages around indefinitely). It also produces a “bill of materials” as a really long JSON-encoded comment. If you let contempt commit the Dockerfile it uses the BOM to generate useful commit messages like: [project] apk:busybox: 1.34.1-r3->1.34.1-r4, so you can see exactly what changed.

Contempt also has support for building and pushing images whenever it changes the Dockerfile. I use it in a GitHub action that runs daily to check all my images are up-to-date and push those that aren’t. It understands the dependencies between images (by pre-analysing the templates) so it will always check and build base images before ones that require them. This means an update to, say, the “alpine” base image will cause anything that depends on it to get updated at the same time, ensuring security updates are rolled out promptly.

The result

You can see my collection of lovingly hand-crafted Dockerfiles in my dockerfiles repository.

There are a number of advantages to handwriting all the images I use. The obvious one is that they’re all built how I want: there are no extraneous dependencies, they’re all based on the same small set of base images (rather than pulling around 10 different versions of debian), nothing tries to also run a DBMS in its container, etc.

This level of customisation goes further, though. Because I’m packaging the software myself, I can tweak how it’s built to fit my needs. A couple of things I run need their own TLS certificates separate from my normal HTTPS setup, so I bake my certwrapper tool in to manage those; I can even set the build flags on certwrapper to only enable the particular DNS provider I personally need (thus avoiding dragging in clients for AWS, GCP, etc). Some software like Hashicorp Vault has an optional web interface that I don’t need, so I simply don’t enable it in the build. These changes save build time, reduce image sizes, in some cases improve runtime performance, and generally reduce the attack surface of what’s running in the container.

It’s also been a great way to learn more about how software is distributed. Writing a Dockerfile is not that distant from writing a PKGBUILD file for an Arch Linux package, or the equivalent for other distributions. In a couple of instances I’ve googled how to solve a particular issue, and found an Arch or Void linux maintainer asking the upstream project about the exact same issue.

Finally, all the images I build I push to my own registry so there are obviously no rate limiting issues. Standing up a service (assuming the Dockerfile has been written!) is amazingly quick because the base layers are all shared and cached, and the registry is a lot physically closer than Docker Hub. Bootstrapping this whole thing becomes an interesting problem because the image for the registry is stored on the registry, but I’ll leave that discussion for another post…