Forum broken for Admin - Assets missing?

Product Support

AI-generated summary

Forum broken for Admin - Assets missing?

Octoberon is experiencing issues with the Discourse forum when logged in as an admin on their production server. They suspect it may be related to changing the DISCOURSE_S3_CDN_URL value in the app.yml file on their staging server, which was temporarily set to the same as the production server. The browser console shows missing assets, but the cause is unclear.

After encountering the issue again when rebuilding Discourse on their standby server, Octoberon believes it may be related to their S3 bucket and CloudFront configuration. They were able to resolve the issue by rebuilding the production app, but are seeking guidance to understand the root cause of the problem.

1

I’m having a problem when logged in as the admin on my production server. I think I may have caused it myself when I changed the app.yml value for DISCOURSE_S3_CDN_URL on the staging server to the same as the production server. I’ve since changed it back and the staging server is working fine just pointing to the S3 bucket. Looking at the browser console, it appears that some assets are not being found but I’m not sure how, why or even if this is the underlying issue.

Here’s an extract from the console when I try to access main forum URL:

Screenshot 2024-01-16 at 12.08.20
Screenshot 2024-01-16 at 12.08.201993×500 175 KB

I’ve tried safe mode but that doersn’t halp.

I’m still very much on the learning curve with AWS. I’ll continue to try and work this problem out for myself but any advice or guidance would be appreciated.

2

I encountered this problem again today after I’d copied over the production app.yml to the standby server and rebuilt Discourse on that server. I presume it’s related to the S3 bucket and Cloudfront somehow but I don’t understand why there’s a conflict just from the rebuild.

It happened during testing of a script I’ve written to switch over the forum to a standby server and set that one as production. Usually I’d complete the switchover by re-pointing the reserved IP to the other server but I didn’t test that part in this run. I rebuilt the production app and things are working normally again but I’d like to understand the problem if anyone can help.