Regression around local chart .tar.gz files

[ffMathy]

What happened?

Before #2410, local chart .tar.gz files could be loaded. Now it has to be an extracted chart.

Expected Behavior

N/A

Steps to reproduce

N/A

Output of pulumi about

N/A

Additional context

No response

Contributing

Vote on this issue by adding a 👍 reaction.
To contribute a fix for this issue, leave a comment (and link to your pull request, if you've opened one already).

[mikhailshilkov]

@rquitales @kpitzen Have we run a postmortem on this regression? Is there an issue?

[kpitzen]

@rquitales @kpitzen Have we run a postmortem on this regression? Is there an issue?

We haven't yet - I wasn't sure if that was reserved for P0s v P1s. If it's a process we want to do for both, then we should run one.

[ffMathy]

I also found it a bit weird that this wasn't caught with (for instance) integration tests.

Seems like a pretty central feature.

So I'd love to see the postmortem too! 😍

[mikhailshilkov]

@kpitzen We run postmortems for any significant user-facing regression (p1 and impact/regression combined), any p0, or other notable misses. My OKRs say this explicitly: "Postmortem call runs for every user-facing P0 or regression within 7 days after a fix". This means this Thu is the deadline for a post-mortem on this issue.

[kpitzen]

Ack - I'll create the doc and set the meeting. @rquitales could you please fill the doc out once it's created, and lead the postmortem itself?

[ffMathy]

This broke again in #2532.

Are there no tests for this? Why does this keep breaking?

CC @mikhailshilkov @kpitzen

[lblackstone]

I'll take a look at this today and figure out why it regressed again. Sorry for the trouble!

[lblackstone]

@ffMathy I just checked the relevant test (added in https://github.com/pulumi/pulumi-kubernetes/pull/2428/files#diff-cbad9e75635545189006ad7411a41379245a4c5d6b4a05b16a764d6bfb6e6036), and it appears to be working as expected. I was able to successfully create the helm.Release resource targeting a local .tgz file using the v4.1.1 release.

Can you provide further details about what behavior you're seeing, and how you determined that it's related to #2532? Thanks!

[ffMathy]

The regression is around using extracted helm charts (so folders) as paths. This worked before 😊

Sorry for not pointing that out ❤️

[lblackstone]

@ffMathy

The regression is around using extracted helm charts (so folders) as paths.

https://github.com/pulumi/pulumi-kubernetes/tree/v4.1.1/tests/sdk/go/helm-release-local/step1 covers this scenario, and I wasn't able to reproduce the reported regression on my end in this case either.

My best guess is that you're somehow running an older version of the provider code, perhaps related to a replacement that was triggered by the change in #2532. If you need further help troubleshooting, I think it would be easier to track down on slack.

[ffMathy]

Thank you. For now, I have just switched to the compressed format.

Sorry for the inconvenience this report may have caused in that case. You guys rock.