https://rancher.com/ logo
Join Slack
Powered by
This message was deleted.
# random
a
This message was deleted.
m
yes, I have a 3 node RKE cluster that runs Rancher with ingress-nginx as daemonset and a node that runs nginx as a TCP loadbalancer. Is that what you mean? I also have another setup with a cloud loadbalancer for ingress-nginx, which is much simpler...
q
Can you show me how you set it up with a cloud loadbalancer? Ive been struggling with this for days
m
which part? I set up rke with 
external
cloud provider and 
ingress.provider: none
and then installed openstack cloud provider manager and ingress-nginx afterwards
then my 
ingress-nginx
service of Type 
Loadbalancer
was provisioned and I could install rancher with a refular ingress option.
q
I have a clean RKE Cluster running on 3 cloud servers and installed rancher via helm. Ive tried with the option tls=external but i cant reach the interface over port 80 like it should, so i can point my cloud loadbalancer to it
Even when i install it with the default options the interface is not reachable
m
what kind of 
service
do you use? 
Nodeport
?
The question is more like, how do you integrate the cloud loadbalancer into your setup? via a cloud-provider or just a nodeport service and point the LB to all nodes on the port (in the range 30000-32000), what does 
$ kubectl get svc -n cattle-system rancher
give as output?
q
Im talking about the rancher cluster itself just to be clear, so i dont use any service
Or do i need to create a service for the rancher web interface?
Ive only been working with rancher in a single node docker container but i wanna start over with a ha rke rancher installtion
Copy code
root@rancher-fsn1:~# kubectl get svc -n cattle-system rancher
NAME      TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)          AGE
rancher   ClusterIP   10.43.165.105   <none>        80/TCP,443/TCP   4d3h
m
that's the service that provides that webinterface. What ingress controller do you use? 
kubectl get ns |grep ingress
?
q
Copy code
root@rancher-fsn1:~# kubectl get ns |grep ingress
ingress-nginx                            Active   4d11h
m
okay, and how is 
ingress-nginx
configured? 
k get svc -n ingress-nginx
and is the ingress created 
k get ingress -A
?
q
Copy code
NAME                                 TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)   AGE
ingress-nginx-controller-admission   ClusterIP   10.43.125.241   <none>        443/TCP   5d7h
Copy code
NAMESPACE       NAME                        CLASS    HOSTS              ADDRESS                                     PORTS     AGE
cattle-system   cm-acme-http-solver-kv9mr   <none>   <http://myclusterdomain.com|myclusterdomain.com>   65.108.208.205,78.47.146.214,78.47.182.83   80        4d13h
cattle-system   rancher                     nginx    <http://myclusterdomain.com|myclusterdomain.com>   65.108.208.205,78.47.146.214,78.47.182.83   80, 443   5d6h
m
so you have 
ingress-nginx
installed, but without a service? Is it a 
DaemonSet
with hostports? Then you can just point your loadbalancer to all three nodes, port 80 & 443.
q
How can i check if its a DaemonSet?
you can see on http://65.108.208.205 for example it just shows 404
m
k get pod,ds,deploy -n ingress-nginx
shows you the pod, daemonsets and deployments in the namespace _ingress-nginx`. I'm not surprised it gives you a 404, the ingress controller listens on the 
host
in the ingress resource.
4 Views
Open in Slack
PreviousNext
Powered by