docs: Document auto-restart of pods on secret rotation

[ewan-chalmers]

What type of PR is this?

/kind documentation

What this PR does / why we need it: Provides guidance for users of the auto-rotation feature of Secrets Store CSI Driver on how to implement optional auto-restart of workload pods when secrets are updated.

Which issue(s) this PR fixes (optional, using fixes #<issue number>(, fixes #<issue_number>, ...) format, will close the issue(s) when the PR gets merged):
Fixes #1647

Special notes for your reviewer:

TODOs:

• squashed commits
• includes documentation
• adds unit tests

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: ewan-chalmers / name: Ewan Chalmers (35070d9, ba3aa84, bd0c93f, c2ad53a, 4d17d7a, 9d930b1, 251459e, 6b943eb, 06db65b, ea08454, ea92357, ed41ab2, 62be1fb, f85a10a, ab3b512, d830b7f, be216d3, dd84820, b25eeb7, f92bf62, 57cf852)

[k8s-ci-robot]

Welcome @ewan-chalmers!

It looks like this is your first PR to kubernetes-sigs/secrets-store-csi-driver 🎉. Please refer to our pull request process documentation to help your PR have a smooth ride to approval.

You will be prompted by a bot to use commands during the review process. Do not be afraid to follow the prompts! It is okay to experiment. Here is the bot commands documentation.

You can also check if kubernetes-sigs/secrets-store-csi-driver has its own contribution guidelines.

You may want to refer to our testing guide if you run into trouble with your tests not passing.

If you are having difficulty getting your pull request seen, please follow the recommended escalation practices. Also, for tips and tricks in the contribution process you may want to read the Kubernetes contributor cheat sheet. We want to make sure your contribution gets all the attention it needs!

Thank you, and welcome to Kubernetes. 😃

[k8s-ci-robot]

Hi @ewan-chalmers. Thanks for your PR.

I'm waiting for a kubernetes-sigs member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[ewan-chalmers]

I have requested and am awaiting my addition to the kubernetes CCLA

[enj]

/assign aramase nilekhc

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: ewan-chalmers
Once this PR has been reviewed and has the lgtm label, please ask for approval from aramase. For more information see the Kubernetes Code Review Process.

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[ewan-chalmers]

Looking at https://deploy-preview-1648--kubernetes-sigs-secrets-store-csi-driver.netlify.app/topics/auto-restart-on-rotation, I note that mermaid source is not rendered as a diagram in the mdbook output (as it is on github).

Simplest thing is to remove the diagram, but I could also look into whether mdbook configuration can be persuaded to render mermaid diagrams.

Looking at

• provide mermaidjs rendering rust-lang/mdBook#762
• https://github.com/badboy/mdbook-mermaid (via https://github.com/rust-lang/mdBook/wiki/Third-party-plugins)

I'll provisionally attempt to add mermaid support in this same PR with an update to docs/book/Makefile and docs/book/book.toml and see if the diagram renders in the preview.

[ewan-chalmers]

Well... that took a few more tries than I expected. The diagram now renders (FWIW - not a very exciting diagram) - https://deploy-preview-1648--kubernetes-sigs-secrets-store-csi-driver.netlify.app/topics/auto-restart-on-rotation#secretproviderclasspodstatus-custom-resource

While doing this, I notice that we should probably upgrade mbook and mbook-toc versions (as well as this new mbook-mermaid). I had to go quite a way back in mbook-mermaid tags to get one that works with our mbook version.

I'm happy to extract mermaid enabling into a separate PR or remove it altogether (along with the diagram src in this PR) if preferred.

[ewan-chalmers]

I'm unfamiliar with squashing commits and looks like there are loads of different ways to do so :/

[benjaminapetersen]

From an official SIG-Auth project, should we be linking out to this in the formal docs?

[benjaminapetersen]

This seems in tone a bit like a blog post for solving a particular problem, rather than a docs site entry. 🤔

[benjaminapetersen]

Per our discussion in the sync meeting, perhaps we keep, but just note that this is not (officially) supported or tested.

[ewan-chalmers]

You could see this doc as a continuation from https://github.com/kubernetes-sigs/secrets-store-csi-driver/blob/main/docs/book/src/topics/secret-auto-rotation.md.

I aimed for similar tone.

That doc links to Reloader.

I'm open to suggestions.

My motivation here is that the solution I arrived at and implemented was

• not obvious based on existing resources
• fairly straightforward given the necessary pointers

[ewan-chalmers]

@benjaminapetersen I have added a note to the top of the doc. If this is not what you have in mind, could you suggest alternative? Thanks

[ewan-chalmers]

@aramase @ritazh Any suggestion on how I could move this PR to done?

[aramase]

Reading through the implementation details, I believe the controller restart will now generate twice the number of calls for the associated pod. The SecretProviderClassPodStatus is only updated by the driver after it has updated the mounted content with the new secrets. By this point, we’ve already made
𝑂(number of secrets) calls to the external secrets store. When the controller restarts the pod, it will trigger a new mount operation, resulting in another
𝑂(number of secrets) calls. In my opinion, this essentially defeats the purpose of auto rotation.

[aramase]

/retitle docs: Document auto-restart of pods on secret rotation

[ewan-chalmers]

Reading through the implementation details, I believe the controller restart will now generate twice the number of calls for the associated pod. The SecretProviderClassPodStatus is only updated by the driver after it has updated the mounted content with the new secrets. By this point, we’ve already made 𝑂(number of secrets) calls to the external secrets store. When the controller restarts the pod, it will trigger a new mount operation, resulting in another 𝑂(number of secrets) calls. In my opinion, this essentially defeats the purpose of auto rotation.

I may be misunderstanding, but I think you are pointing to a potentially undesirable consequence of this implementation rather than to something which defeats the purpose. The purpose is: if mounted secrets are updated, pod will be restarted; pod code does not need to handle dynamic secret updating. This does work. However, are you pointing to the fact that for one secret update operation in the secrets store, the driver will now read the secrets from there and create mount for the pod twice - because it does this again after controller restarts the pod? If so, do you agree it is not 'defeating the purpose' but rather a consequence which might be undesirable and so should be mentioned?

I have proposed a caveat NOTE to that effect: 57cf852

[ewan-chalmers]

Thanks for review @aramase