Confused if Service is working

Timothy Sutton

unread,

Jun 13, 2022, 5:27:06 PM6/13/22

to Knative Users

I have set up knative operator, and services following the instructions here https://knative.dev/docs/install/operator/knative-with-operators/#configure-dns. I am using kourier, and deployed the hello world go test service. I am getting a response that doesnt jive with the expected results of the curl command. I just get back an empty html response. I did deploy the magic DNS stuff as well. Below is kind of the state of everything in my cluster.

kn service list
NAME URL LATEST AGE CONDITIONS READY REASON
hello http://hello.default.192.168.79.100.sslip.io hello-00001 17s 3 OK / 3 True

k get services
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes ClusterIP 10.43.0.1 <none> 443/TCP 3h18m
operator-webhook ClusterIP 10.43.151.142 <none> 9090/TCP,8008/TCP,443/TCP 170m
docker-registry NodePort 10.43.172.19 <none> 5000:30059/TCP 15m
hello-00001-private ClusterIP 10.43.26.12 <none> 80/TCP,443/TCP,9090/TCP,9091/TCP,8022/TCP,8012/TCP 9m42s
hello-00001 ClusterIP 10.43.214.200 <none> 80/TCP,443/TCP 9m42s
hello ExternalName <none> kourier-internal.knative-serving.svc.cluster.local 80/TCP 9m39s

kubectl --namespace knative-serving get service
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
autoscaler-bucket-00-of-01 ClusterIP 10.43.80.149 <none> 8080/TCP 170m
activator-service ClusterIP 10.43.211.17 <none> 9090/TCP,8008/TCP,80/TCP,81/TCP,443/TCP 170m
autoscaler ClusterIP 10.43.67.164 <none> 9090/TCP,8008/TCP,8080/TCP 170m
controller ClusterIP 10.43.1.182 <none> 9090/TCP,8008/TCP 170m
domainmapping-webhook ClusterIP 10.43.245.20 <none> 9090/TCP,8008/TCP,443/TCP 170m
webhook ClusterIP 10.43.43.232 <none> 9090/TCP,8008/TCP,443/TCP 170m
autoscaler-hpa ClusterIP 10.43.201.157 <none> 9090/TCP,8008/TCP 170m
net-kourier-controller ClusterIP 10.43.51.42 <none> 18000/TCP 160m
kourier LoadBalancer 10.43.2.79 192.168.79.100,192.168.79.101,192.168.79.102,192.168.79.103 80:31899/TCP,443:31766/TCP 160m
kourier-internal ClusterIP 10.43.240.24 <none> 80/TCP,443/TCP 160m
default-domain-service ClusterIP 10.43.198.58 <none> 80/TCP 82m

This is where I think things aren't configured, expect simple Hello statement back but as you see getting something different and no hello even wrapped in the html. If I try wget I get 503 error.

curl http://hello.default.192.168.79.100.sslip.io
<!DOCTYPE html>
<html>
<head>
</head>
<body>
<iframe src="/fortiadc_error_page/index.html" frameborder="0" width="100%" scrolling="no" onload="function resizeIframe(obj) {
obj.style.height = obj.contentWindow.document.body.scrollHeight + 10 + 'px';
};resizeIframe(this)"></iframe>
</body>
</html>

Mauricio Salatino

unread,

Jun 13, 2022, 5:34:46 PM6/13/22

to Timothy Sutton, Knative Users

Timothy, what flavour of Kubernetes are you using? Is that KinD?

--
You received this message because you are subscribed to the Google Groups "Knative Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to knative-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/knative-users/e5d1c3d6-aef0-488e-9e7d-e93b854a02e9n%40googlegroups.com.

--

- Blog: http://salaboy.com

- Github user: http://github.com/salaboy

- Twitter: http://twitter.com/salaboy

- Mauricio "Salaboy" Salatino -

Timothy Sutton

unread,

Jun 14, 2022, 8:15:17 AM6/14/22

to Knative Users

Sorry, I meant to mention this. It is K3S

Timothy Sutton

unread,

Jun 14, 2022, 11:22:47 AM6/14/22

to Knative Users

As a follow up I have made a little progress but still confused, about where I am misconfigured.

I was deploying the service as such kn service create hello --image gcr.io/knative-samples/helloworld-go --port 8080 --env TARGET=World

I then tried again deploying with kn service create hello --image gcr.io/knative-samples/helloworld-go --port 80 --env TARGET=World

I was able to query the private service, and get expected response.

curl http://hello.default.192.168.79.100.sslip.io
<!DOCTYPE html>
<html>
<head>
</head>
<body>
<iframe src="/fortiadc_error_page/index.html" frameborder="0" width="100%" scrolling="no" onload="function resizeIframe(obj) {
obj.style.height = obj.contentWindow.document.body.scrollHeight + 10 + 'px';
};resizeIframe(this)"></iframe>
</body>

</html>[vagrant@master ~]$ k get service

NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE

kubernetes ClusterIP 10.43.0.1 <none> 443/TCP 21h
operator-webhook ClusterIP 10.43.151.142 <none> 9090/TCP,8008/TCP,443/TCP 20h
docker-registry NodePort 10.43.172.19 <none> 5000:30059/TCP 18h
hello-00001-private ClusterIP 10.43.35.87 <none> 80/TCP,443/TCP,9090/TCP,9091/TCP,8022/TCP,8012/TCP 14s
hello-00001 ClusterIP 10.43.84.250 <none> 80/TCP,443/TCP 14s
hello ExternalName <none> kourier-internal.knative-serving.svc.cluster.local 80/TCP 13s
[vagrant@master ~]$ curl 10.43.35.87
Hello World!

Evan Anderson

unread,

Jun 14, 2022, 3:55:43 PM6/14/22

to Timothy Sutton, Knative Users

Your initial `curl` command suggests that you're hittinga Fortinet web firewall at that IP address, rather than Kourier in your cluster. If you're running locally (particularly on Mac/Windows), this may be because your load balancer IPs aren't actually routable from your desktop shell.

I'd start by trying to get a load balancer service working before getting Knative working (since it's the same network path, but less of it). This blog post might help (untested): https://blog.kubernauts.io/k3s-with-k3d-and-metallb-on-mac-923a3255c36e

From: knativ...@googlegroups.com <knativ...@googlegroups.com> on behalf of Timothy Sutton <timothy...@saabusa.com>
Sent: Tuesday, June 14, 2022 8:22:47 AM
To: Knative Users <knativ...@googlegroups.com>
Subject: [Suspected Spam] Re: Confused if Service is working

⚠ External Email

To view this discussion on the web visit https://groups.google.com/d/msgid/knative-users/ae196f15-74b3-4e00-b196-bcc962b6568an%40googlegroups.com.

⚠ External Email: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender.

Timothy Sutton

unread,

Jun 15, 2022, 8:42:37 AM6/15/22

to Knative Users

So this is definitely some sort of DNS issue. After closer inspection with curl -v, my machine was resolving DNS name to something outside my local network. I edited my /etc/hosts to force resolution of hello.default.192.168.79.100.sslip.io to the ip address 192.168.79.100 (address of my load balancer), and all worked correctly.

Reply all

Reply to author

Forward