You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Discovered this during AKS multi-cluster guide validation...
global NodePort services only work as expected if cilium is replacing kube-proxy.
See comments in: cilium/cilium#23266
Is your feature request related to a problem?
This results in spurious connectivity test failures on multi-clusters with kube-proxy running, leads to bad experience for people running through the documented multi-cluster guided path.
Describe the solution you'd like
By default, auto-detect if kube-proxy replacement is enabled and if it is run the global NodePort service tests. If kube-proxy replacement is not active, skip this tests..as they are expected to fail.
The text was updated successfully, but these errors were encountered:
Global NodePort services are not supported when they are managed by
iptables/ipvs, since they do not know about remote endpoints. Hence,
let's skip the related connectivity tests when running in multicluster
mode and KPR NodePort support is disabled, to prevent spurious failures.
Related: cilium/cilium#23128
Related: cilium/cilium#23266Fixes: cilium#1352
Signed-off-by: Marco Iorio <[email protected]>
Global NodePort services are not supported when they are managed by
iptables/ipvs, since they do not know about remote endpoints. Hence,
let's skip the related connectivity tests when running in multicluster
mode and KPR NodePort support is disabled, to prevent spurious failures.
Related: cilium/cilium#23128
Related: cilium/cilium#23266Fixes: #1352
Signed-off-by: Marco Iorio <[email protected]>
Proposal / RFE
Discovered this during AKS multi-cluster guide validation...
global NodePort services only work as expected if cilium is replacing kube-proxy.
See comments in: cilium/cilium#23266
Is your feature request related to a problem?
This results in spurious connectivity test failures on multi-clusters with kube-proxy running, leads to bad experience for people running through the documented multi-cluster guided path.
Describe the solution you'd like
By default, auto-detect if kube-proxy replacement is enabled and if it is run the global NodePort service tests. If kube-proxy replacement is not active, skip this tests..as they are expected to fail.
The text was updated successfully, but these errors were encountered: