Support Tekton OCI bundles without feature flag (i.e. promote to beta)

[bobcatfish]

Feature request

In #3492 we put Tekton OCI bundles behind a feature flag to indicate that they are still "alpha" and to give us a chance to look into some issues @vdemeester discovered while using them.

This issue is to track the work to promote these bundles to a feature that you don't have to opt into.

TODO:

• Identify what we need to do to promote this feature
• Identify and solve performance issues
• Deal with complications around where you refer to Tasks at runtime (Add support for referencing Tasks in git #2298 (comment)) - e.g. if you are making changes to a Task referenced by a Pipeline, how do you point that Pipeline to your in progress Task (currently you'd have to edit the Pipeline or publish the Task)

[vdemeester]

/assign

[tekton-robot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale with a justification.
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 with a justification.
If this issue should be exempted, mark the issue as frozen with /lifecycle frozen with a justification.

/lifecycle stale

Send feedback to tektoncd/plumbing.

[nikhil-thomas]

/remove-lifecycle stale

[tekton-robot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale with a justification.
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 with a justification.
If this issue should be exempted, mark the issue as frozen with /lifecycle frozen with a justification.

/lifecycle stale

Send feedback to tektoncd/plumbing.

[vdemeester]

/remove-lifecycle stale

[tekton-robot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale with a justification.
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 with a justification.
If this issue should be exempted, mark the issue as frozen with /lifecycle frozen with a justification.

/lifecycle stale

Send feedback to tektoncd/plumbing.

[vdemeester]

/lifecycle frozen
This is part of the roadmap 🙃

[lbernick]

Instead of promoting OCI bundles to beta, could we instead go ahead with promoting remote resolution to beta (#4710) and have people use the "bundle" resolver? as far as I can tell it provides the same functionality.

[vdemeester]

@lbernick yes, we need to decide which one to promote, but my feeling is that we should aim to promote remote resolution with the bundle resolver "built-in" probably, and deprecate the current way we refer bundles. It is gonna take some time though (to go from point A to point B), so we should decide this rather sooner.

[abayer]

I agree with @vdemeester - it makes sense to have one approach for resolving bundles, and remote resolution provides us with more than just bundle resolution, so I'd go with that.

[vdemeester]

Two things around this issue:

• I added this to the v1 board in order to make sure we go through it and we take a decision here.
• The current take is to put our efforts into the remote resolvers, see here, that would also support oci bundle, making this a "duplicate" way of refering to an OCI bundle.

I think we should make sure we mark the field (taskref.bundle) as deprecated as soon as possible (0.39 ?) so that we do not bring it in v1 and we rely on remote resolvers instead. Both are alpha today, so it shouldn't be a big deal to deprecate it (we could even remove it "abruptely" but because we use it in dogfooding as well as internally at Red Hat, I feel we should go the deprecation route).

/assign
(Assigning myself to prepare a pull-request deprecated the field)

[lbernick]

Proposed in tektoncd/community#755

[lbernick]

Decision has been made to deprecate OCI bundles syntax, so I'm going to close this issue. @vdemeester feel free to reopen if there's anything else to track here-- deprecation is being tracked in #5514