Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[RFC] Flux Bootstrap for OCI-compliant Container Registries #4749

Draft
wants to merge 5 commits into
base: main
Choose a base branch
from
Draft
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
396 changes: 396 additions & 0 deletions rfcs/000X-flux-bootstrap-oci/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,396 @@
# [RFC] Flux Bootstrap for OCI-compliant Container Registries

**Status:** provisional

**Creation date:** 2024-04-27

**Last update:** 2024-04-27

## Summary

Flux should allow a Git-less bootstrap procedure where the cluster desired state is stored in OCI artifacts.

On the client-side, the Flux CLI should offer a command for packaging its own Kubernetes manifests into
an OCI artifact and pushing the artifact to a container registry.

On the server-side, the Flux controllers should be configured to self-update from the registry
and reconcile the cluster state from OCI artifacts stored in the same or a different registry.
Comment on lines +16 to +17
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would imagine that users may wish to choose between pulling from upstream OCI artifact that is published as part of the Flux release or having a full copy of it. If they choose to use a copy, another command may be needed to keep their copy up to date. Does it make sense?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You would always have a copy in your registry that includes customisations, same as with Git, bootstrap means vendoring the Flux manifests that in 99% of cases would need some fine tuning.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah, that's true that in most cases right now people will end up with a copy. How do they bring the copy up to date, e.g. some component has new pod spec fields that have to be set or there is an RBAC change? Do they have to read changelog and implement such changes? If in OCI world this could be avoided by means of referencing an upstream artifact and local changes stored as a patch/kustonization, it might be very nice actually, I guess it wasn't feasible with Git as an upstream.

Copy link
Member Author

@stefanprodan stefanprodan Apr 27, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For both Git and OCI bootstrap, the Flux update is fully automated in CI. See the Story 3 in this RFC. And also the docs here: https://fluxcd.io/flux/flux-gh-action/#automate-flux-updates

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

... bootstrap means vendoring the Flux manifests that in 99% of cases would need some fine tuning.

Just want to clarify, can this fine tuning be done with an in-cluster Kustomization, or has that been proven somehow challenging?

Copy link
Member Author

@stefanprodan stefanprodan Apr 28, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I've added the common flags to the RFC.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Controller selection is already supported via flags common to all bootstrap sub commands.

Yes, I know, what I mean is that it would be kind of less natural specify controllers with kustomize, one would probably need to select bases as there is no meaningful parameters and if or switch statements (at least last time I checked), the CLI offers are more meaningful option.
So because you need to select controllers, you start with CLI that gives you a single file that you kustomize a little, but it's all much complicated then you might have wished and defeats the purpose of putting kustomize in at this stage.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

CUE on the other hand could do better in all of this and you could potentially remove the need for CLI and make custom configs easier to introduce. Does what I said earlier make more sense now?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The way Flux can be customized at bootstrap is all via Kustomize patches and CLI options, this must be 100% compatible with the oci sub-command. I’m not considering using CUE or anything really that would diverge from the current Git bootstrap procedure. Users should be able to migrate from Git to OCI by simply reusing their current flux-system overlay including patches, image overrides, configGenerator, volumes, etc.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That is for sure, I'm just thinking maybe self-managed flux in the future could use CUE for this, possibly even without exposing CUE to the user. CLI could still work the same way on the surface also. I do recall we once spoke of an installer operator too, you could use CUE there and in the CLI. Just an idea :)


## Motivation

After the implementation of [RFC-0003](../0003-kubernetes-oci/README.md) in 2022 and the introduction
of the `OCIRepository` source, we had a recurring ask from users about improving the UX of running
Flux fully decoupled from Git.

Given that OCI registries are evolving into a generic artifact storage solution,
we should assist Flux users who don't want to depend on a Git (for any reason,
including auth and SSH key management) in their production infrastructure to
bootstrap and manage their Kubernetes clusters using OCI artifacts.

To decouple the clusters reconciliation from the Git repositories, Flux allows packaging and publishing
the Kubernetes manifests stored in Git to an OCI registry by running the `flux push artifact`
command in CI pipelines.

### Goals

- Add support to the Flux CLI for bootstrapping with a container registry as the source of truth.
- Make it easy for users to switch from Git repositories to OCI repositories.

### Non-Goals

- Automate the migration of Flux manifests from a Git bootstrap repository to OCI.

## Proposal

Implement the `flux bootstrap oci` command with the following specifications:

```shell
flux bootstrap oci \
--url=oci://<registry-url>/<flux-manifests>:<tag> \
--username=<registry-username> \
--password=<registry-password> \
--kustomization=<local/path/to/kustomization.yaml> \
--cluster-url=oci://<registry-url>/<fleet-manifests>:<tag> \
--cluster-path=<path/inside/oci/artifact>
```

The bootstrap shares the following flags with the `flux install` command:

```text
--cluster-domain string internal cluster domain (default "cluster.local")
--components strings list of components, accepts comma-separated values (default [source-controller,kustomize-controller,helm-controller,notification-controller])
--components-extra strings list of components in addition to those supplied or defaulted, accepts values such as 'image-reflector-controller,image-automation-controller'
--image-pull-secret string Kubernetes secret name used for pulling the toolkit images from a private registry
--log-level logLevel log level, available options are: (debug, info, error) (default info)
--network-policy deny ingress access to the toolkit controllers from other namespaces using network policies (default true)
--registry string container registry where the toolkit images are published (default "ghcr.io/fluxcd")
--toleration-keys strings list of toleration keys used to schedule the components pods onto nodes with matching taints
--version string toolkit version, when specified the manifests are downloaded from https://github.com/fluxcd/flux2/releases
--watch-all-namespaces watch for custom resources in all namespaces, if set to false it will only watch the namespace where the toolkit is installed (default true)
```

The Terraform/OpenTofu counterpart is the `flux_bootstrap_oci` provider that exposes
the same configuration options as the CLI.

The bootstrap operations are split into two phases:

- Install and self-update configuration for the Flux components.
- Cluster state reconciliation configuration.

### Install and self-update configuration

The command performs the following steps based on the `url`, `username`,
`password` and `kustomization` arguments:

1. Logs in to the OCI registry using the provided credentials.
2. Generates an OCI artifact from the Flux components manifests and the `kustomization.yaml` file.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Might be worse mentioning here regarding workload identity mods that may be needed, e.g. on EKS role ARN needs to be set as an annotation, I forgot if that was necessary in GKE also.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We have all of those documented here https://fluxcd.io/flux/installation/configuration/workload-identity/, people will need to read the docs and adapt their kustomization.yaml.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

But perhaps there could be bootstrap argument to specify provider-specific attributes that would be handled accordingly based on provider flag?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There is --provider=<aws|azure|gcp> but this is for the CLI to use the role of the machine where it's running. The IAM of the bastion host may be different from the one that you want to use for Flux source-controller. So we'll need yet another flag with the identity name. This is how Flux AIO works: https://timoni.sh/flux-aio/#__tabbed_1_3

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is probably not the right place to bring this up but the workload identity for GKE is partially incorrect, you dont need to annotate service accounts any more with a GCP SA for GKE workload identity. You grant the Kubernetes service account access to what ever resources it needs via a member statement like below

principal://iam.googleapis.com/projects/PROJECT_NUMBER/locations/global/workloadIdentityPools/PROJECT_ID.svc.id.goog/subject/ns/flux-system/sa/source-controller

I'm going to have a think on how I can update the docs on this one, but thought I'd raise it before I forget

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We've recently adopted Flux for our multi-cloud architecture. In order to support that we're actually overriding env vars, volumes, and volume bindings directly in order to set up OIDC-based auth to each cloud on our workloads. If we can extend those parameters on the source controller, or if we can pass in a pod template that would allow us to easily inject OIDC auth for accessing the OCI backend (I suspect we could do this with the Helm install method, but support in the CLI would be nice, as the bootstrap command is great).

3. Applies the Flux components manifests along with their customisations to the cluster.
4. Pushes the OCI artifact to the container registry using the specified tag.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wonder if it would make sense to have two registry auth. A read-only for the image pull secrets, and read-write for pushing the artifacts to the registry. Storing a read-write secret in the cluster for image pull secrets does not seem like a good idea.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It is even possible to consider more pull secrets with specific permissions (least privileges principle):

  • one with a read-only permission to be used by the Deployments of the Flux controllers to pull images
  • one with a read-only permission (different namespace than images) to be used by the OCIRepository resource to pull OCI artifacts
  • one with write permissions to be used by the bootstrap command line

Then, it is possible to consider different OCI registry: one for images and another for Flux artifacts, because the latter could contain sensible infrastructure information.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wonder if it would make sense to have two registry auth. A read-only for the image pull secrets, and read-write for pushing the artifacts to the registry.

Yes this is something the command could support. Currently our OCI implementation supports reading the Docker config file from the host OS, so we could use that for write operations and the flags for in-cluster secret.

@sestegra the pull secret for the container images is already supported, it's one of the common flags to all bootstrap commands.

5. Generates an image pull secret, an OCIRepository that points to the OCI artifact and
stefanprodan marked this conversation as resolved.
Show resolved Hide resolved
a Flux Kustomization object that reconciles the OCI artifact contents.
6. Applies the image pull secret, OCIRepository and Flux Kustomization to the cluster.
stefanprodan marked this conversation as resolved.
Show resolved Hide resolved

Note that the creation of the image pull secret is skipped when
[Kubernetes Workload Identity](#story-2) is used for authentication to the container registry.

Artifacts pushed to the registry:
- `<registry-url>/<flux-manifests>:<checksum>` (immutable artifact)
- `<registry-url>/<flux-manifests>:<tag>` (tag pointing to the immutable artifact)

The OCI artifact has the following content:

```shell
./flux-system/
├── gotk_components.yaml
└── kustomization.yaml
```

Objects created by the command in the `flux-system` namespace:
- `flux-components` Secret
- `flux-components` OCIRepository
- `flux-components` Kustomization

### Cluster state reconciliation configuration

After the OCIRepository and Flux Kustomization called `flux` become ready, the command
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Does this have to be strictly sequential and synchronous?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes it does, CRDs and controllers must be up and running before the cluster sync is deployed, same procedure as for the Git bootstrap.

continues with the following steps:

1. Logs in to the OCI registry where the cluster artifacts are stored using the provided credentials.
2. If the cluster OCI artifact is not found, an empty artifact is created
and pushed to the registry using the provided tag.
3. Generates an image pull secret, an OCIRepository and a Flux Kustomization object
stefanprodan marked this conversation as resolved.
Show resolved Hide resolved
that reconciles the cluster OCI artifact contents.
4. Applies the image pull secret, OCIRepository and Flux Kustomization to the cluster.

Note that the creation of the image pull secret is skipped when
[Kubernetes Workload Identity](#story-2) is used for authentication to the container registry.

Objects created by the command in the `flux-system` namespace:
- `flux-system` Secret
- `flux-system` OCIRepository
- `flux-system` Kustomization

If the cluster registry is the same as the Flux components registry, the command could reuse the
`flux-components` image pull secret.

If the cluster OCI artifact is not found, the generated one contains the following:

```shell
./clusters/my-cluster/ # taken from --cluster-path
└── kustomization.yaml # empty overlay with no resources
```

### Registry authentication

The `flux bootstrap oci` command supports the following authentication methods:

- Basic authentication with `--username` and `--password`. The credentials are stored in a Kubernetes Secret.
- OIDC authentication with `--provider=<aws|azure|gcp>`. No credentials are stored in the cluster, source-controller
will use Kubernetes Workload Identity to authenticate to the registry.

To avoid passing the credentials as CLI flags, the password can be read from the standard input, e.g.:
`echo <password> | flux bootstrap oci` or using an environment variable `OCI_PASSWORD`.

If the registry is self-hosted and uses a self-signed TLS certificate,
the root CA certificate can be provided with the `--ca-file` flag.

If the registry is exposed on HTTP and not HTTPS, the `--allow-insecure-http`
flag can be used to force non-TLS connections.

### Signing and verification

The `flux bootstrap oci` command supports the following signing and verification methods:

- Cosign
- Notation

TODO: Add more details about the signing and verification methods, flags and options.

### User Stories

#### Story 1

> As a platform operator I want to bootstrap a Kubernetes cluster with Flux
> using OCI artifacts stored in a container registry.

The following example demonstrates how to bootstrap a Flux instance using GitHub Container Registry
as the OCI registry for Flux components and the cluster state.

```shell
flux bootstrap oci \
--url=oci://ghcr.io/stefanprodan/flux-manifests:production \
--username=<ghcr-username> \
--password=<ghcr-token> \
--kustomization=flux-manifests/kustomization.yaml \
--cluster-url=oci://ghcr.io/stefanprodan/fleet-manifests:production \
--cluster-username=<ghcr-username> \
--cluster-password=<ghcr-token> \
--cluster-path=clusters/production
```

Generated OCI artifacts:

- `ghcr.io/stefanprodan/flux-manifests:88b028f`
- `ghcr.io/stefanprodan/flux-manifests:production`
- `ghcr.io/stefanprodan/fleet-manifests:6f7a258`
- `ghcr.io/stefanprodan/fleet-manifests:production`

Objects created in the `flux-system` namespace:

Flux components reconciliation:

```yaml
apiVersion: source.toolkit.fluxcd.io/v1beta2
kind: OCIRepository
metadata:
name: flux-components
namespace: flux-system
spec:
interval: 1m
url: oci://ghcr.io/stefanprodan/flux-manifests
ref:
tag: production
secretRef:
name: flux-components
---
apiVersion: kustomize.toolkit.fluxcd.io/v1
kind: Kustomization
metadata:
name: flux-components
namespace: flux-system
spec:
interval: 1h
retryInterval: 5m
sourceRef:
kind: OCIRepository
name: flux-components
path: ./
prune: true
```

Cluster state reconciliation:

```yaml
apiVersion: source.toolkit.fluxcd.io/v1beta2
kind: OCIRepository
metadata:
name: flux-system
namespace: flux-system
spec:
interval: 1m
url: oci://ghcr.io/stefanprodan/fleet-manifests
ref:
tag: production
secretRef:
name: flux-system
---
apiVersion: kustomize.toolkit.fluxcd.io/v1
kind: Kustomization
metadata:
name: flux-system
namespace: flux-system
spec:
interval: 1h
retryInterval: 5m
sourceRef:
kind: OCIRepository
name: flux-system
path: clusters/production
prune: true
```

#### Story 2

> As a platform operator I want to bootstrap an EKS cluster with Flux
> using OCI artifacts stored in ECR.

The following example demonstrates how to bootstrap a Flux instance using ECR using IAM auth.
Assuming the EKS nodes have read-only access to ECR and the bastion host where
the Flux CLI is running has read and write access to ECR:

```shell
flux bootstrap oci \
--provider=aws \
--url=oci://aws_account_id.dkr.ecr.us-west-2.amazonaws.com/flux-manifests:production \
--kustomization=flux-manifests/kustomization.yaml \
--cluster-url=oci://aws_account_id.dkr.ecr.us-west-2.amazonaws.com/fleet-manifests:production \
--cluster-path=clusters/production
```

Note that when using Kubernetes Workload Identity instead of the worker node IAM role,
the `kustomization.yaml` must contain patches for the source-controller Service Account
as described [here](https://fluxcd.io/flux/installation/configuration/workload-identity/).

#### Story 3

> As a platform operator I want to sync the cluster state with the fleet Git repository.

Push changes from the fleet Git repository to the container registry:

```shell
# clone the fleet Git repository
git clone https://github.com/stefanprodan/fleet.git
cd fleet
git switch main

# push the contents the fleet OCI repository and tag it with the commit short SHA
flux push artifact oci://ghcr.io/stefanprodan/fleet-manifests:$(git rev-parse --short HEAD) \
--path="./" \
--source="$(git config --get remote.origin.url)" \
--revision="$(git branch --show-current)@sha1:$(git rev-parse HEAD)"

# tag the new version for production
flux tag artifact oci://ghcr.io/stefanprodan/fleet-manifests:$(git rev-parse --short HEAD) \
--tag=production
```

This operation can be automated using the Flux GitHub Action.

The Git repository structure would be similar to the
[flux2-kustomize-helm-example](https://github.com/fluxcd/flux2-kustomize-helm-example) with the following changes:

- The `clusters/production/flux-system` directory is no more.
- The Flux Kustomization objects defined in the `clusters/production` directory, such as
`infrastructure.yaml` and `apps.yaml`, have the `.spec.sourceRef` set to
`kind: OCIRepository` and `name: flux-system`.

#### Story 4

> As a platform operator I want to update the Flux controllers on my production cluster
> from CI without access to the Kubernetes API.

Download the latest CLI version and update Flux directly in the registry, without rerunning bootstrap:

```shell
# pull the latest manifests from the registry
flux pull artifact oci://ghcr.io/stefanprodan/flux-manifests:production \
--output=./flux-manifests

# update the Flux components manifests
flux install --export > ./flux-manifests/flux-system/gotk-components.yaml
Comment on lines +324 to +330
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

These are two alternative methods right? It's not very clear from the text at the moment.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If you don't have access to the cluster (what the user story is about), this is the only way. If you have API access, then like with Git bootstrap, you can just rerun it to update. OCI bootstrap behaves the same as Git bootstrap.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I just thought these two commands will write the same kind of output, except that install lets you select a subset of controllers... maybe I am missing something else.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is it the case right now that one has to rerun bootstrap on major/minor releases while patch releases are taken care of by in-cluster image version bumps?

Copy link
Member Author

@stefanprodan stefanprodan Apr 27, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Every time we release Flux, users get a PR opened to update their manifests in Git. For OCI you would need some kind of semver range or some other manual gate e.g. a GitHub workflow dispatch to approve minor bumps and let only patch versions be automatically push to the registry.

Copy link
Member Author

@stefanprodan stefanprodan Apr 27, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I just thought these two commands will write the same kind of output, except that install lets you select a subset of controllers... maybe I am missing something else.

All the flux bootstrap commands have the same args as flux install, so you can pick controllers, etc with bootstrap too. If you bootstrap with flux bootstrap oci --components=source-controller,kustomize-controller, to update your would run flux install --components=source-controller,kustomize-controller --export in CI.

Copy link
Contributor

@errordeveloper errordeveloper Apr 27, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I am talking about'pull artifact' vs 'install --export' (per above)

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You need to pull to preserve the existing kustomization.yaml and any other extra resources you may have added at bootstrap. The install -export command only generates the components YAML.

Copy link
Contributor

@errordeveloper errordeveloper Apr 29, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah, of course, this is a "rebase" ;)
Maybe the text needs to make that clearer.


# calculate the checksum of the manifests
checksum=$(grep -ar -e . ./flux-manifests/ | shasum | cut -c-16)

# extract the Flux version and commit
flux_version=$(flux version --client | awk '{print $2}')
flux_commit=$(go version -m $(which flux) | grep vcs.revisio | awk -F= '{print $NF}')

# push the updated manifests to the registry using the checksum as tag
flux push artifact oci://ghcr.io/stefanprodan/flux-manifests:${checksum} \
--path="./flux-manifests" \
--source="https://github.com/fluxcd/flux2" \
--revision="${flux_version}@sha1:${flux_commit}"

# tag the new version for production
flux tag artifact oci://ghcr.io/stefanprodan/flux-manifests:${checksum} \
--tag=production
```

This operation could be simplified by implementing a dedicated CLI command and/or GitHub Action.

#### Story 5

> As a platform operator I want to update the registry credentials on my clusters.

To rotate the registry credentials, generate a new GitHub token and overwrite the image pull secret:

```shell
flux create secret oci flux-system \
--url=ghcr.io \
--username=<ghcr-username> \
--password=<ghcr-token>
```

Another option is to rerun the bootstrap command with the new credentials.

#### Story 6

> As a platform operator I want to know the Git repository used to generate the OCI artifact
> where a Kubernetes Deployment running on the cluster belongs to.

To determine the source of a Kubernetes object in-cluster, run:

```shell
flux -n default trace deploy/podinfo
```

The trace command will display the OCI artifact URL, tag and digest along
with the Git repository URL, branch and commit.

## Design Details

The bootstrap feature will be implemented as a Go package under `fluxcd/flux2/pkg/bootstrap/oci`
using the [fluxcd/pkg/oci](https://github.com/fluxcd/pkg/tree/main/oci)
library for OCI operations such as auth, push, pull, tag, etc.

Both the Flux CLI and the Terraform/OpenTofu provider will use the `fluxcd/flux2/pkg/bootstrap/oci` package
and expose the same configuration options.

### Enabling the feature

The feature is enabled by default.

## Implementation History

* NONE
Loading