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

In multi-cluster mode, adding topology.kubernetes.io/region label to a new nodes does not trigger the update of envoy cluster endpoints. #51144

Closed
2 tasks done
fizos opened this issue May 20, 2024 · 3 comments
Labels
feature/Multi-cluster issues related with multi-cluster support

Comments

@fizos
Copy link

fizos commented May 20, 2024

Is this the right place to submit this?

  • This is not a security vulnerability or a crashing bug
  • This is not a question about how to use Istio

Bug Description

We have a multi-primary, multi-network Istio cluster. Currently, everything is functioning properly. However, when we add a new worker node (without the topology.kubernetes.io/region label) to one of the clusters and trigger workload scaling, the traffic is not routed to the pods on the new worker node. By using the "istioctl pc endpoints" command, we can see that the pods on the node are not included in the envoy endpoints configuration. After adding the topology.kubernetes.io/region label to the new node, istiod does not update the cluster endpoints configuration in envoy based on the pods on that node. To make it effective, we need to manually restart these pods.

In this scenario, should istiod be responsible for triggering the update of envoy endpoint configuration based on changes in the topology.kubernetes.io/region label?

Version

client version: 1.21.1
control plane version: 1.21.1
data plane version: 1.21.1 (92 proxies)

Client Version: v1.27.6
Kustomize Version: v5.0.1
Server Version: v1.27.6

Additional Information

No response

@istio-policy-bot istio-policy-bot added the feature/Multi-cluster issues related with multi-cluster support label May 20, 2024
@hzxuzhonghu
Copy link
Member

This is expected, we donot have a informer watching on node labels. Because the useful attributes istio cared about rarely change.

While watching it may be costy.

@hzxuzhonghu
Copy link
Member

One workaround is to update the pod labels to retrigger topology recompute

@fizos
Copy link
Author

fizos commented May 20, 2024

OK, thank you for your reply!

@fizos fizos closed this as completed May 30, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
feature/Multi-cluster issues related with multi-cluster support
Projects
None yet
Development

No branches or pull requests

3 participants