github/kubernetes
1
0
Fork 0
mirror of https://github.com/k3s-io/kubernetes.git synced 2025-07-29 22:46:12 +00:00
Code Issues Projects Releases Wiki Activity

e2e: fix flake on loadbalancer tests

Browse Source 
validating that one endpoint is reachable from one part of the cluster
is not enough condition to consider it will be reachable from any node,
as different Services proxies on different nodes will have different
propagation delays for the EndpointSlices and Services information.
This commit is contained in:
Antonio Ojea 2024-07-18 12:50:41 +00:00
parent 7693a7e71a
commit fdbe6912d2
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
test/e2e/network/loadbalancer.go
View File

@ -480,10 +480,10 @@ var _ = common.SIGDescribe("LoadBalancers", feature.LoadBalancer, func() {
e2eservice.TestReachableHTTP(ctx, ingress, svcPort, loadBalancerLagTimeout)
ginkgo.By("checking reachability from pods when LoadBalancerSourceRanges is unset")
// We can use timeout 0 here since we know from above that the service is
// already running (and we aren't waiting for changes to it to propagate).
checkReachabilityFromPod(ctx, true, 0, namespace, acceptPod.Name, ingress)
checkReachabilityFromPod(ctx, true, 0, namespace, dropPod.Name, ingress)
// There are different propagation delay for the APIs for different nodes, so it tries
// a few times, despite previously it was confirmed that the Service was reachable.
checkReachabilityFromPod(ctx, true, e2eservice.KubeProxyEndpointLagTimeout, namespace, acceptPod.Name, ingress)
checkReachabilityFromPod(ctx, true, e2eservice.KubeProxyEndpointLagTimeout, namespace, dropPod.Name, ingress)
// Create source ranges that allow acceptPod but not dropPod or
// cluster-external sources. We assume that the LBSR rules will either see