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

Restore specific groups for peering selectors #741

Merged
merged 2 commits into from
Apr 27, 2021
Merged

Conversation

nolar
Copy link
Owner

@nolar nolar commented Apr 25, 2021

In #644, Kopf's peering selectors were generalised only to "any name", with groups/versions removed — in order to satisfy both old (zalando.org) and new (kopf.dev) groups of the CRD.

This caused Kopf to scan all(!) groups to discover resources, not only the specific ones, which caused some issues in some cluster setups, e.g. where metrics.k8s.io API group is "503 Unavailable" because the extension is scaled down to zero (#740 #738). There is no strong reason to scan all available resources indeed. More on that, that is an unnoticed side-effect with extra work done by the operator.

This PR restores the specific groups of the peering resources (and versions — because versions were there too, originally). But makes it so that BOTH old & new groups are supported. In practice, usually, only one of two groups will exist. If both groups exist, both peering resources will be served in parallel.

Fixes #740.

TODOs:

  • Test it manually with old & new CRDs.
  • Verify that the whole-cluster group scanning ends.

@nolar nolar added the bug Something isn't working label Apr 25, 2021
Signed-off-by: Sergey Vasilyev <nolar@nolar.info>
Signed-off-by: Sergey Vasilyev <nolar@nolar.info>
@nolar nolar merged commit ad8d6d7 into main Apr 27, 2021
@nolar nolar deleted the specific-peering-groups branch April 27, 2021 18:59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Group_filter in scan_resources is always to None
1 participant