Generate and register new storage backend about apiserver-builder-alpha

[pwittrock]

@pires Would you be interested in implementing this?

[pires]

Yes. I should be able to pick it up during the last couple weeks of August.

[klustria]

Any updates on this? Is this kind of related to this issue/question I posted: #226 ?

[pires]

Yes it is. And I think I can spare some cycles in the next couple months for this.

[klustria]

Appreciate your response. If you don't mind, can you give me an overview of what you intend to do? Details would be good too. I'm trying to look around for any links that can provide me ideas how to implement but I can't seem to find any. If you have any of those, would appreciate if you can share it with me. I've been trying to look for this topic to justify the use of API-Server aggregation in our project as some people in the team are kind of reluctant with ETCD, so our plan is to try other persistence mechanism. So would appreciate if you can share your ideas about this. TIA.

[klustria]

As I'm reviewing the k8s.io/apiserver code, it looks like storage.Interface is the one used for implementing the storage. pkg/registry/generic/NewRawStorage() is the one that instantiates the actual storage. However from looking at factory.Create in NewRawStorage(), it looks like it only supports ETCD and ETCD3. Unless my understanding is wrong, etcd seems to be hardcoded into the apiserver code without an opportunity to pass a custom implementation of storage.Interface. So as a question, is there another way to do this, i.e. implement your own custom storage.Interface and be able use it in the aggregated api server? Would appreciate if you can share your thoughts if have ideas regarding this already. I'm really eager to get some answers to this question, so I can justify the use of the aggregated api-server to my team who wants to use another persistent storage instead of etcd. TIA.

[pires]

Not really, @klustria. While, conceptually, supporting a different storage engine makes sense, the fact is kubernetes/apiserver library is very opinionated about using etcd.

But here's an idea which might be worth a try. There's also a related gdoc link you should read.

[pires]

It seems the metrics-server proves it can be done! In this case, metrics are stored in-memory and no etcd is involved.

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[pires]

If anyone is still interested in following-up, maybe inspiration can be taken from how k3s relies on kvsql to provide support for different storage backends.

[yue9944882]

/remove-lifecycle rotten

@pires interesting point, i opened #362 to track this feature

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

[k8s-ci-robot]

@fejta-bot: Closing this issue.

In response to this:

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

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/test-infra repository.

[drewwells]

Documentation would at least be useful here. There's a single example of non-etcd but there's no documentation of how to use it.