New self-hosted instance tries to connect to api.m...
# support
c
I'm trying to deploy a new self-hosted instance for our production environment. Everything seems to deploy fine, but it's not clear that variable replacement works for the application's static files. When I open the app URL in the browser, error messages appear in the console that the app is trying to connect to api.medplum.com, which fails the content security policy. Obviously, it should be instead connecting to the hosted API (which is running just fine, and returns an OK).
A bunch of messages like this - I'm not including the CSP because it correctly references our actual API server
I'm being forced to use the 2.1.1 CLI, because anything later results in this same error message from a previous open thread: https://discord.com/channels/905144809105260605/1161375729804656640
Pretty strange - basically the
baseUrl
from the config doesn't seem to be getting used, even though it is set correctly there. I would be happy to jump on a call to screen share if that's easiest.
Shoot, our development environment app is now broken with the same error after updating
r
Hi @codingtom - we're very sorry about the error. There was a bug in our deploy process for 2.1.5 that caused the ap to always point to
api.medplum.com
. This has been fixed here (https://github.com/medplum/medplum/pull/3070) and will be released in 2.1.6. For now, we recommend reverting to 2.1.4
c
Ahh gotcha - I see that the cli always grabs and uploads assets from the latest version. I can probably wait for 2.1.6 release, thanks!
Will make sure I check issues next time 😅 sorry for the noise!
r
No need to apologize, this was our fault
We should be able to release 2.1.6 by EOD today
173 Views