Closed 120L020430 closed 1 month ago
Let me elaborate on the situation:
I have a local three-node k3s cluster and apply serving-crds.yaml
, serving-core.yaml
, kourier.yaml
, serving-default-domain.yaml
, serving-hpa.yaml
. After deploying example Hello, I get the following result:
watch -n 0.5 kubectl get pod -l serving.knative.dev/service=hello -owide
Observing the autoscaling, everything is fine at this point
#!/bin/bash
for ((i=0;i<=10000;i++)); do {
echo `curl -s http://hello.default.192.168.31.206.sslip.io`
} & done
wait
And then I applyeventing-crds.yaml
,eventing-core.yaml
, eventing-kafka-controller.yaml
, eventing-kafka-channel.yaml
and eventing-kafka-broker.yaml
,and connect it with my local kafka cluster.
The broker is as follows:
apiVersion: eventing.knative.dev/v1
kind: Broker
metadata:
annotations:
# case-sensitive
eventing.knative.dev/broker.class: Kafka
# Optional annotation to point to an externally managed kafka topic:
# kafka.eventing.knative.dev/external.topic: <topic-name>
name: default
namespace: default
spec:
# Configuration specific to this broker.
config:
apiVersion: v1
kind: ConfigMap
name: kafka-broker-config
namespace: knative-eventing
After testing the issue example, the problem arises.
Later I tried to use kafkasource
and deployed this example and it failed.
The log of pod/kafka-source-dispatcher is as follows:
I think the most important is "ERROR","level_value":40000,"stack_trace":"io.netty.resolver.dns.DnsResolveContext$SearchDomainUnknownHostException: Failed to resolve 'event-display.default.svc.cluster.local' [A(1)] and search domain query for configured domains failed as well: [knative-eventing.svc.cluster.local, svc.cluster.local, cluster.local]
Like the original hello, the address of event-display is actually event-display.default.192.168.31.206.sslip.io
, but here it resolves event-display.default.svc.cluster.local
I don't know if I'm missing any files or if there's a mistake in the configuration somewhere.
@120L020430 can you show the full knative service status?
$ kubectl get ksvc hello -oyaml
if there is a URL like event-display.default.svc.cluster.local
can you try to sending a request to it from a pod in the cluster?
there might be something wrong in your knative serving installation
This issue is stale because it has been open for 90 days with no
activity. It will automatically close after 30 more days of
inactivity. Reopen the issue with /reopen
. Mark the issue as
fresh by adding the comment /remove-lifecycle stale
.
Describe the bug A clear and concise description of what the bug is. I'm new to Knative. After installing knative serving and knative eventing, I try the trigger tutorial, but the trigger doesn't work correctly. It doesn't receive the message. The result of cloudevents player is below: I find the SUBSCRIBER_URI of the trigger is different from the service.
My k3s cluster has three nodes. I use kourier+magic DNS as the networking and DNS layer, following the installing turior. I use kafka as the channel and broker layer. It has the same three nodes.
cloudevents-player.yaml
ce-trigger.yaml
Expected behavior A clear and concise description of what you expected to happen.
The trigger can work correctly like the tutorial.
To Reproduce Steps to reproduce the behavior.
Install knative following the YAML-based Installation, and use kourier, magic DNS and kafka. Finally, try the cloudevents player.
Knative release version v1.13.1 Additional context Add any other context about the problem here such as proposed priority