tectonic-console pod keeps failing

[Arve Knudsen]

In porting Tectonic Installer to DigitalOcean I am experiencing that tectonic-console pods are failing, with the message 'Provider config sync failed, retrying in 1s: invalid character 'U' looking for beginning of value'.

What is causing this? In any case, it looks to me like a bug where an error message (from the provider in question?) is attempted to be parsed as JSON.

Thanks,

Arve

[Ed Rooth]

This is generally a symptom of Tectonic Identity (Dex) not being available at the configured endpoint Console is using to contact it. Console queries Dex's endpoint to sync the OIDC provider info & public signing keys. If the initial requests fail, Console will eventually exit. The weird error is probably b/c Console is expecting a JSON response but is getting something else like an ingress 503 error page or something (we should detect this and log better messages). You'll need to make sure this URL is configured correctly and available to the Console pod https://github.com/coreos/tectonic-installer/blob/master/modules/tectonic/resources/manifests/console/deployment.yaml#L54

It should be something like: https://baseurl.example.com/identity

and Console will try to query: https://baseurl.example.com/identity/.well-known

It's response should be similar to what Google's OIDC endpoint serves:

https://accounts.google.com/.well-known/openid-configuration

TL;DR

Dex is down, or Console can't reach Dex, or DNS might be misconfigured.

[Arve Knudsen]

Thanks Ed, that makes sense :) Would be great if that error handling were fixed.

Best,

Arve