Hello,

I wanted to connect to my kubernetes (limited availability) cluster from my gitlab repo, firstly I followed this guide and successfully created a user ‘admin-user’ which I could use to sign into Kubernetes Dashboard as an admin

GitLab’s Kubernetes connection page asks for:

  • Cluster name -> Which I got from the kubeconfig file beside 'name’
  • API URL -> From kubeconfig file beside 'server’
  • CA Certificate (PEM) -> From kubeconfig file beside 'certificate-authority-data’ (then base64 decoded)
  • Token -> The token created for my admin-user account
  • Project namespace -> A new unique one as specified by GitLabs Docs

But I had limited success. GItLab reported 'Something went wrong while installing Helm Tiller’, when I clicked install. There is a final experimental settings in the GitLab UI called 'RBAC-enabled cluster’, I have tried with this setting both checked and unchecked, same results.

After looking at the dashboard, I see a 'gitlab-managed-apps’ namespace had appeared which has done a few things like create a tiller-token secret, a config map and there’s some items listed under events

Is this something I’m doing wrong, a bug with GitLab or a bug with DigitalOcean?
I’m relatively new to Kubernetes so if there’s more I can do to debug the issue please let me know

Thanks

EDIT: Oh, very strange, I deleted the tiller-secret and config-map left behind, tried clicking install about 3 more times, and now Helm Tiller has deployed successfully, no more errors. So solved for me, although I suspect more people will encounter this issue. (I had RBAC-enabled cluster checked for anyone looking for the same 'fix’)

These answers are provided by our Community. If you find them useful, show some love by clicking the heart. If you run into issues leave a comment, or add your own answer to help others.

×
4 answers

I had the same issue, I was able to fix it after installing helm on my own, but it also required to create a helm tiller service account first:

kubectl create serviceaccount --namespace kube-system tiller
kubectl create clusterrolebinding tiller-cluster-rule --clusterrole=cluster-admin --serviceaccount=kube-system:tiller
kubectl patch deploy --namespace kube-system tiller-deploy -p '{"spec":{"template":{"spec":{"serviceAccount":"tiller"}}}}' 

then (assuming helm is already installed)

helm reset 
helm init

for some reason it only worked until I removed the namespace gitlab created

kubectl delete namespace gitlab-managed-apps

Then I went to the gitlab cluster configuration, clicked the helm install button and it worked. I think the tiller account is the only prerrequisite I’ll try with a fresh cluster later, the above instructions are for fixing an existing one.

kubectl get pods -n gitlab-managed-apps
NAME                                    READY     STATUS    RESTARTS   AGE
runner-gitlab-runner-7b9c6c6bc9-tw6wg   1/1       Running   0          8m
tiller-deploy-6cc8b46cf-bfbwk           1/1       Running   0          12m

To anyone still having issues here: the above has been resolved, just check the

RBAC-enabled cluster
Enable this setting if using role-based access control (RBAC). This option will allow you to install applications on RBAC clusters.

option in GitLab while you’re add the cluster. To get the API url, try this:

kubectl cluster-info | grep 'Kubernetes master' | awk '/http/ {print $NF}'

Source: GitLab Documentation

If you are getting a Kubernetes Error 401 it means your Token is incorrect as it was in my case. Here I was using a token from an API tab in the Dashboard BUT should really create the Token by command line like recommended. After generating a user and a token using kubectl commands and changing the Token the installation of Helm worked perfectly.
Hope it helps!

Hi Wearing, I am just facing the same issue.

Submit an Answer