pulumi-community #general

general

late-printer-99022

01/15/2020, 8:56 AM

> KUBECONFIG=./kubeconfig.json kubectl get svc --all-namespaces

NAMESPACE           NAME                      TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)         AGE
agromaster          agromaster-plot           NodePort    172.20.88.116    <none>        80:30577/TCP    4d11h
default             kubernetes                ClusterIP   172.20.0.1       <none>        443/TCP         15d
kube-system         kube-dns                  ClusterIP   172.20.0.10      <none>        53/UDP,53/TCP   15d
syngenta-platform   syngenta-platform-ndvi    NodePort    172.20.135.178   <none>        80:30603/TCP    4d11h
syngenta-platform   syngenta-platform-nginx   NodePort    172.20.233.206   <none>        80:30177/TCP    4d11h

late-printer-99022

01/15/2020, 8:57 AM

late-printer-99022

01/15/2020, 9:01 AM

@flat-insurance-25294 trying to follow this blog - https://www.pulumi.com/blog/kubernetes-ingress-with-aws-alb-ingress-controller-and-pulumi-crosswalk/ from last 1 week. still not able to make it work.

flat-insurance-25294

01/15/2020, 9:02 AM

and does your /health return 200?

late-printer-99022

01/15/2020, 9:02 AM

on POD? yes !

flat-insurance-25294

01/15/2020, 9:03 AM

and your container exposes port 80?

late-printer-99022

01/15/2020, 9:04 AM

this much code I have.

flat-insurance-25294

01/15/2020, 9:05 AM

You really should stop dumping code here, use pastebin or gist.

late-printer-99022

01/15/2020, 9:05 AM

flat-insurance-25294

01/15/2020, 9:05 AM

it makes it really hard to filter and search.

flat-insurance-25294

01/15/2020, 9:06 AM

Does your service actually return 200? Do they respond on port 80, Will it work with http? Does it take longer than 10 seconds for your systems to boot?

late-printer-99022

01/15/2020, 9:07 AM

https://gist.github.com/nsisodiya/e1ba779fd5ed6c73d00446c0379ab95b

late-printer-99022

01/15/2020, 9:08 AM

my nodejs based image has container port 3000 but target port is 80. as shown in code.

flat-insurance-25294

01/15/2020, 9:09 AM

How about you make a smaller project without all that cruft, with just 200, hello world and make that works first?

flat-insurance-25294

01/15/2020, 9:10 AM

container port doesn’t actually do anything, if you’re referring to the docker image

EXPOSE

. That’s just for documentation really.

flat-insurance-25294

01/15/2020, 9:12 AM

Remove the healthcheck to see if it works, but you really should just make a hello world playground first

late-printer-99022

01/15/2020, 9:13 AM

This is the smaller project with 2 services and each point to some route like /service1 and /service2 that is all the POC i am doing. POC: 1. Create Cluster - done 2. Run 2 Docker image - 1 create actual service and another is just ngnix - done 3. Run a ALB Ingress controller to route traffic , http://<ALB URL>/ngnix, http://<ALB URL>/ndvi, - pending 4. Add DNS record , http://api.syn-platform.com - pending 5. Make it work with

https

too - pending

flat-insurance-25294

01/15/2020, 9:14 AM

Make an even smaller PoC because obviously you want to make sure the ALB works first.

flat-insurance-25294

01/15/2020, 9:14 AM

kubectl logs --namespace kube-system $(kubectl get po --namespace kube-system | egrep -o [a-zA-Z0–9-]+alb-ingress[a-zA-Z0–9-]+)

Check errors as well.

late-printer-99022

01/15/2020, 9:17 AM

no error

late-printer-99022

01/15/2020, 9:18 AM

@flat-insurance-25294 - just found that- the NGNIX Service Target group is healthy but ndvi service is unhealthy !

flat-insurance-25294

01/15/2020, 9:18 AM

Seriously stop code dumping, lol.

late-printer-99022

01/15/2020, 9:19 AM

You told not to dump code. but not for logs. you mean don't dump log and code both ! 🙂

late-printer-99022

01/15/2020, 9:25 AM

@flat-insurance-25294 can you verify following at-least

> KUBECONFIG=./kubeconfig.json kubectl get svc --all-namespaces

syngenta-platform   syngenta-platform-ndvi    NodePort    172.20.135.178   <none>        80:30603/TCP    4d12h
syngenta-platform   syngenta-platform-nginx   NodePort    172.20.233.206   <none>        80:30177/TCP    4d12h

it is showing port

in both case. is it correct? nginx service has container port 80 and target port 80, but ndvi service has container port 3000 and target port is 80.

late-printer-99022

01/15/2020, 9:29 AM

unfortunately, the example by @narrow-area-11460 show 2080-game and all ports are 80. in real life, our docker container runs on something which is not 80.

late-printer-99022

01/15/2020, 9:29 AM

so, it may be the issue with my configuration.

late-printer-99022

01/15/2020, 10:45 AM

@flat-insurance-25294 problem solved , Instances are showing healthy now.

flat-insurance-25294

01/15/2020, 10:45 AM

nice,

late-printer-99022

01/15/2020, 10:46 AM

but I am stilling getting error like

E0115 10:45:08.543340       1 :0] kubebuilder/controller "msg"="Reconciler error" "error"="failed to reconcile targetGroups due to failed to reconcile targetGroup targets due to Unable to find http port on service syngenta-platform-ndvi"  "controller"="alb-ingress-controller" "request"={"Namespace":"syngenta-platform","Name":"syngenta-platform"}

in Ingress logs, but my Service is working fine.

late-printer-99022

01/15/2020, 1:46 PM

Questions - Why dualstack is added in route53 on Route53 AWS UI it is showing me as - dualstack.e9275192-syngentaplatform--87b5-1276037906.us-east-1.elb.amazonaws.com. but in LoadBalancer url is - e9275192-syngentaplatform--87b5-1276037906.us-east-1.elb.amazonaws.com what is the difference? both gives exactly same response !!