Closed scottking2 closed 4 years ago
The value only changes the kubernetes config, not the graphite config. So you have to change the GRAPHITE_URL variable in the graphite config too.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
Is this a request for help?: yes
Is this a BUG REPORT or FEATURE REQUEST? (choose one):
Version of Helm and Kubernetes: helm: v3.1.2 k8s: v1.15.5
Which chart in which version: graphite: 0.5.0
What happened: I changed the service.port from the default 8080 to port 8000 and the pod gets stuck in a crashloopbackoff with the liveness and readiness probes failing to connect.
What you expected to happen: I expected the service.port to be configurable to a different port number.
How to reproduce it (as minimally and precisely as possible): I ran the following on my mac terminal connected with docker desktop and k8s through docker desktop.
helm install graphite --set service.port=8000 --set service.type=NodePort kiwigrid/graphite
Anything else we need to know:
Everything works as expected with the service.type=NodePort if I don't change the service.port. Below are the logs when i follow the pod after changing the service.port: