Open kumy opened 1 month ago
BTW the oulogin
plugin works fine to configure our kube config
# ~
$ kubectl oulogin --host=k8sou.ou.xxxxxx
oulogin 0.0.7
Checking for existing issuer https://k8sou.ou.xxxxxx/auth/idp/k8sIdp
Invalid context or does not exist, launching browser to login
Starting OpenID Connect for host k8sou.ou.xxxxxx
kubectl configuration created
# ~
$ k get po
NAME READY STATUS RESTARTS AGE
kube-oidc-proxy-orchestra-6989c8ffbb-wlglv 2/2 Running 0 39m
openunison-operator-6f5d4b7cc6-dn9w7 2/2 Running 0 45m
openunison-orchestra-b459ffb77-tb7vz 2/2 Running 0 39m
ouhtml-orchestra-login-portal-5fd446c89-l2k74 2/2 Running 0 36m
test-orchestra-orchestra 0/1 Completed 0 39m
Does the dashboard work?
We didn't deployed the kube dashboard if it's what you mean
We didn't deployed the kube dashboard if it's what you mean
Understood, what happens if you go to https://k8s.ou.XXX/k8stoken/token/user
, do you get a bunch of JSON? Also, do you have network policies enabled in your values.yaml?
Yes we see json, and no network policies enabled
network_policies:
enabled: false
ingress:
enabled: true
labels: []
monitoring:
enabled: true
labels: []
apiserver:
enabled: true
labels: []
namespace_label: kubernetes.io/metadata.name
That's so odd. The original error comes usually from the virtual gateway not being able to connect or the certificate isn't setup. but you're getting to the OpenUnison service and pod. The "slow" part, generating tokens, works. It appears the issue is just the conection from OpenUnison --> html pod, but you're able to login to the openunison portal. is it possible that the service mesh is blocking the URL?
oh, i just saw your screen shot, that's REALLY old. What version of OpenUnison are you deploying (charts, images, etc)
We're using helm as:
helm template openunison tremolo/openunison-operator --namespace openunison | tee openunison-base.yaml
helm template orchestra tremolo/orchestra --namespace openunison -f values.yaml | tee orchestra-base.yaml
helm template orchestra-login-portal tremolo/orchestra-login-portal --namespace openunison -f values.yaml | tee orchestra-login-portal-base.yaml
from
That deploys images:
$ k get po -o yaml|grep image:|sort -u
image: ghcr.io/openunison/openunison-k8s:1.0.41
image: ghcr.io/openunison/openunison-k8s-html:1.0.0
image: ghcr.io/openunison/openunison-kubernetes-operator:1.0.6
image: ghcr.io/tremolosecurity/kube-oidc-proxy:1.0.7
image: ghcr.io/tremolosecurity/python3:1.0.0
image: upm-istio/proxyv2:1.19.7-distroless
is it possible that the service mesh is blocking the URL?
we were suspecting our base istio rules to block something, but we didn't have any to what/where to start our investigation
ghcr.io/openunison/openunison-k8s-html:1.0.0
odd, where did that come from? Try setting openunison.html.image
to ghcr.io/openunison/openunison-k8s-react:1.0.0
in your values.yaml. also, make sure that openunison.html.legacy
is false
.
we were suspecting our base istio rules to block something, but we didn't have any to what/where to start our investigation
Do you have Kiali deployed? That's usually a good place. What version of istio?
Hi, we're trying to deploy the stack. We followed https://openunison.github.io/deployauth/#alternate-deployment-methods page.
We can successfully login on orchestra using GitHub and were presented with the landing page
When we try to access
https://k8sou.xxx/k8stoken/
we get errorWe didn't find any error in the logs from the pods in the openunison namespace when we hit/refresh that page.
However, we found this one on startup, could it be related :shrug: ?
Any hint on what we should check, or how to enable more debug logs? (We're using Istio as Ingress controller.) Thanks for your help :pray:
Startup logs: openunison-orchestra.txt