Join GitHub today
GitHub is home to over 40 million developers working together to host and review code, manage projects, and build software together.Sign up
Some times Pilot is skipping sending endpoint updates for a new pod #19025
Steps to reproduce the bug
Version (include the output of
How was Istio installed?
Environment where bug was observed (cloud vendor, OS, etc)
Here is what I found when this happened
Does the above make sense? I think in general, we should trigger Eds update when labels change because localities + other things might have changed?
@hzxuzhonghu I see. But should n't the standard labels like
Or are you suggesting I should validate for change in those labels only?
IIUC, pod cache is eventually consistent, if pod is not available for this new endpoint IP here https://github.com/istio/istio/blob/master/pilot/pkg/serviceregistry/kube/controller/controller.go#L938 - instead of skipping that endpoint, should we refresh the pod list by calling