search

Hubs-Foundation / hubs-cloud

Resources for self hosted Hubs Cloud instances
Mozilla Public License 2.0
152 stars 96 forks source link

Removing use Cloudflare workers results in 'Content Security Policy' violation #87

Open robinkwilson opened 4 years ago

robinkwilson commented 4 years ago

Needs repro. Unchecked use Cloudflare workers in the admin panel. After refresh my hubs cloud page is white with attached errors about "because it violates the following Content Security Policy"

No settings about proxies exist in my aws parameter store to delete or add.

Domain Recipe: Personal 3 domains Screen Shot 2020-07-20 at 11 03 56 AM

***.workers.dev is the cloudflare workers domain apparently still trying to connect despite the setting being not selected.

Steps to repro:

  1. Have stack with working cloudflare workers selected, then Update to Offline mode.
  2. Update stack from Offline mode to Online mode
  3. Wait til it just finishes spinning up
  4. Unselect the use Cloudlfare workers in Admin Panel > Content CDN menu
  5. Refresh and the page should be blank

Backup and Restore is the only option to get it working again currently.

┆Issue is synchronized with this Jira Task

smakhtin commented 4 years ago

I also stuck with Cloudflare — they just blocked our workers (TOS violation by caching videos) and now I can't log in to the admin dashboard. Any way to fix this?

robinkwilson commented 3 years ago

Relates to: https://github.com/mozilla/hubs-cloud/issues/94

antpb commented 3 years ago

I'm hitting this on an Enterprise instance after setting up the CF workers. My instance is with workers active, all scene files (image skybox, and videos) are showing CSP issues as described in this issue. Is there any workaround for this? Its pretty essential to the success of an upcoming event and I'm nervous about turning off the workers and getting locked out.

antpb commented 3 years ago

ack, the reason mine were bad was due to a scene published before setting up the workers. Resetting the media items in spoke and republishing fixed them up!

Ponderjaunt commented 3 years ago

Hit the same isssue exactly.

Ponderjaunt commented 3 years ago

Hit the same isssue exactly.

Fixed by re-initializing cloud stack - after deleting both IPs DNS records and DBs.