CODE Server reachable but not working

[GioDima]

Hello,
Thank you in advance for providing the app (and hopefully help) and sorry for any language error (english non-native). I'm currently trying to install the Built-in CODE server on a Nextcloud Hub 5 (version 27.0.2). The Nextcloud server is running in Apache and on the same server (different domain, but same local and public IP) is also running a jellyfin istance behind reverse proxy also configured in Apache.
Both servers are NOT running in a docker (definetly not a very smart decision). Both servers have valid SSL certificates generated by certbot and both servers are redirecting http request to https

The CODE is shown as "reachable" from the Nextcloud Office web page, (image below)

I tried creating some documents but they were not opening, so i checked the logs and found lots of GuzzleHttp\Exception\ClientException with the following explanation

http://SERVER-IP/apps/richdocumentscode/proxy.php?req=/hosting/capabilities resulted in a  404 Not Found response

Given that the server was redirecting all http to https, and having seen someone else on the nextcloud forum/github having issues with the SSL I tried disabling the redirect function of the server to see if it was causing issues. Also to ensure I was not blocking something I didn't know of at this point UFW was disabled on the system.

trying to navigate via a browser to the URL results in a redirect to the following URL:
http://MEDIASERVER-DOMAIN/apps/richdocumentscode/proxy.php?req=/hosting/capabilities

So it was obvious that the Jellyfin instance was doing something weird. I've tried to disable it so that at least i could rule it as the causing issue, but it didn't work. Navigating to the http://SERVER-IP/apps/richdocumentscode/proxy.php?req=/hosting/capabilities with jellyfin disabled resulted in this page

I've also tried to run the Collabora_Online.AppImage manually from the terminal, however, that resulted in another error

Failed to listen on Server port(s) (9983-9983). Exiting

I've verified that libfuse and glibc are installed.
I don't know what else to try; at this point, I wonder if it's easier to just configure another server instance to run the CODE server on.

richdocumentscode version: 23.5.301

[itagili]

I have the same problem and I haven't been able to solve it so far.

[GioDima]

The "fix" for me was to install a separate instance of collabora server and run it on the same machine using apache VirtualHost and a Reverse proxy, documentation is pretty good however should be noted that if you use the same machine for nextcloud/collabora server, in the "allow list for WOPI request" the loopback IP address should be added (usually it's 127.0.0.1)

[vitorsemeano]

I also have the same problem since version 23.05.5.1.1 (docker hub).

Reverting back to 23.05.4.2.1 works for me.

Also tried the latest version, the problem persists.

What i did was just pulling the new image and down up the docker container and this problem occurs. Nextcloud can reach it but can't load any document from it.

[GioDima]

At this point, I'm not using anymore the Built-in CODE server and I switched to a "normal" Collabora server, there's a very good guide on how to setup and it's working (knock on wood) somewhat without issue with two updates (both on Nextcloud and Collabora server), still I really hope they improve the reliability of the built-in CODE server because it would be a very simple and "fast" deployment. I'm going to keep the issue open if someone else has things to add

Link to the guide I used to setup the Collabora server

[vitorsemeano]

Actually this was happening to me on a normal collabora server.

I found a solution to the problem, by simply removing the dot escapes from "server_name". For some reason i was escaping the dots, and nextcloud couldn't resolve the generated collabora url. Perhaps this changed since 23.05.5.1.1.

[EmJotGeh]

Same Problem here (Build In)

[vasyugan]

I see the same both with the builtin code server as well as with collabora installed as native packages on a remote server. The only option that works for me is collabora in docker. This affects only one of my nextcloud instances, the other, which run on different servers are unaffected. I today moved the affected instance to a different computer, expecting that this would make the error go away, yet it did not. I still am trying to understand how this one instance is different from the others.