Skip to content
This repository has been archived by the owner on Jun 14, 2018. It is now read-only.

Use ConfigMap for sidecar initializer deployment configuration #1062

Closed
ayj opened this issue Aug 18, 2017 · 2 comments
Closed

Use ConfigMap for sidecar initializer deployment configuration #1062

ayj opened this issue Aug 18, 2017 · 2 comments

Comments

@ayj
Copy link
Contributor

ayj commented Aug 18, 2017

The sidecar initializer is configured through command line flags. It might be nice to pull the configuration and sidecar template from a k8s ConfigMap (e.g. envoy-initializer.yaml).

NOTE: The sidecar initializer and kube-inject currently share the same injection programmatic template. Adoption of envoy-initializer.yaml style template should ensure these templates don't diverge. kube-inject could use the newer k8s ConfigMap but care should be taken to avoid breaking offline mode, e.g. kube-inject might need to take ConfigMap YAML as argument to retrieve template.

@gyliu513
Copy link
Member

gyliu513 commented Sep 4, 2017

I think that the injected configuration should be handled by pilot but not the configMap as the end user may update the configMap and caused the initializer does not work. @ayj

@ayj
Copy link
Contributor Author

ayj commented Sep 5, 2017

I'm not sure how pushing injection configuration into Pilot would solve that problem. The end user could make the same misconfiguration errors there. Ultimately the injection configuration should be protected and I believe that is possible today with appropriate RBAC settings on the ConfigMap.

Implemented by #1104

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Projects
None yet
Development

No branches or pull requests

3 participants