| Stuck trying to set up Eclipse Che as an in-browser IDE [message #1851751] |
Thu, 14 April 2022 20:44  |
 Ryan Westlund
Messages: 2
Registered: April 2022 |
Junior Member |
|
|
I'm a noob at both Kubernetes and Eclipse Che who's trying to run Eclipse Che on a k3s cluster. I'm not sure if what I'm trying to do is supported or not; my ultimate goal is to get Eclipse Che integrated into a webpage where you can use it to write code inside a pod being used as a development environment (this pod is not managed by Eclipse).
The cluster is all on localhost for development, so to deploy Che I am running `chectl server:deploy --platform=k8s --domain=172.18.0.3.nip.io --skip-oidc-provider-check`. The output shows that everything succeeds except starting the "Eclipse Che server pod", on which it says "Failed to start a pod, reason: Error, exit code: 137"), and I'm at a loss for how to get any other information about the cause. (The error log file contains nothing else).
Looking at the pods in `kubectl get pod -n eclipse-che` output, I see che-operator, postgres, devfile-registry, plugin-registry, che-dashboard, and just 'che' are running, while che-gateway is in CrashLoopBackOff.
When I check `kubectl get services -n eclipse-che`, I see the services che-host, che-dashboard, che-gateway, postgres, devfile-registry, and plugin-registry. They have ports assigned, but trying to ncat to the k3s node (which is running at 172.18.0.3) on those ports always says connection refused. I notice these services are of type ClusterIP whereas the ones I had set up before involving Che in the project were of type NodePort and had pairs of ports (like 22:31294), whereas these ClusterIP services have only individual ports.
I also tried to reach some of the ClusterIP addressses; the result is always a timeout when connecting, regardless of whether I use the correct port or not.
There's a API running which I only found out about from trying to run `chectl server:delete` and it told me about the server at https://0.0.0.0:33985. I tried hitting this API; it returns a JSON object saying Unauthorized.
When searching for similar issues I encountered some screenshots of people who'd made it to a dashboard and were getting an Unauthorized error at it, so I think there's supposed to be a dashboard I can access that uses this API, but I don't know where that dashboard is at or if it won't be accessible because che-gateway failed.
So I have multiple things that seem like they're not working as intended (gateway pod failing to start, no public endpoint to reach the dashboard), but don't know how to proceed on any of them. I have no idea if these issues could be caused by running it on localhost or by skipping the OIDC provider check (and if so how I would install an OIDC provider.) I have checked all the documentation and Github threads I can find.
|
|
|
|
|
Powered by
FUDForum. Page generated in 0.07825 seconds
] 
Search
Help
Register
Login
Home
