-
Notifications
You must be signed in to change notification settings - Fork 1.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
GCP Che installation guide does not work #20726
Comments
@npw3202 thank you for opening this issue. We are sunsetting installation via the helm chart but have not updated the doc yet. Can you please try installing Che using the operator? You can find some details here. |
@l0rd, Thanks for the follow up, I ran server:delete and then the recommended command using the operator. I got
but when I tried to visit the user's dashboard at https://che-eclipse-che.ide.***.com, I got image suggesting that the nginx proxy might be misconfigured. When I port-forward to the che-dashboard service, the 404'd resources show up on port 8080 (e.g. http://localhost:8080/dashboard/editor.worker.js 200s as expected but https://che-eclipse-che.ide.***.com/dashboard/editor.worker.js 404s). Is the recommended setup steps for the operator's ingress proxy the same as the recommended setup steps in the GCP installation guide? if so, is there any recommended way of debugging this? |
pls try: |
kubectl describe ingress -n eclipse-che yields
Dashboard's logs are
When visiting the /dashboard/ path, the nginx's logs give
it appears like the 404'd requests were not routed to 10.16.6.157:8080 because the endpoint described in ingress is /dashboard/ and not /dashboard/. I'll try to change the spec to route to /dashboard/, but I suspect that this will result in a circular redirect (ERR_TOO_MANY_REDIRECTS). |
Ok I think I have a handle of what's going on. For some reason, the default nginx config routes The che-host redirects |
I appear to be unable to manually alter the ingress rules (the operator reverts the change before it can be picked up by nginx). I'm unfamiliar with the codebase (I could also be completely wrong about that being the culprit), but at a first glance, could the culprit be this? |
I think it is the same issue ingress path is To check this:
The problem is that it is not possible to configure ingress path ^( We need to add such ability |
Thanks for the response, actually it looks like the solution is a little different than suggested. Interestingly enough, that recommended solution didn't work (resulted in an infinite redirect), but changing the ingress path to "/.*" only was also not enough. The annotation
per https://kubernetes.github.io/ingress-nginx/user-guide/ingress-path-matching/ was also required. Even with GCP, the nginx ingress of "/.*" works (making me think the custom configuration is not needed, but rather, just the addition of the regex annotation is sufficient). |
looks like all is WAI now, lmk if you want me to send a PR to fix this in the operator code / send a PR for documentation fix. Thanks so much for your help. |
@npw3202 |
sorry, somehow this didn't show up on my GitHub issues :-/ I'll have a PR sent by EOD hopefully |
@npw3202 |
@npw3202 |
The GCP docs have been removed. |
Describe the bug
I suspect that the GCP installation guide is out of date. I tried following it several times and always got ERR_TOO_MANY_REDIRECTS. I also tried changing from /dashboard/* -> /dashboard/ in the helm patch file and it seems to give another error (404s the manifest.json making me think the nginx is misconfigured)
Che version
7.38@latest
Steps to reproduce
Follow the GCP Che instructions on https://www.eclipse.org/che/docs/che-7/installation-guide/installing-che-on-google-cloud-platform/#creating-a-service-account-secret-on-google-cloud-platform_che
Expected behavior
An eclipse che instancec to be spun up.
Runtime
other (please specify in additional context)
Screenshots
image
Installation method
chectl/latest
Environment
GCE
Eclipse Che Logs
Eclipse Che itself appears to be functioning (installation all went off without a hitch, when I port forward 8080 from the dashboard svc, it appears like the 404'd assets are available).
Additional context
No response
The text was updated successfully, but these errors were encountered: