Calibre, blank login screen after upgrade when proxied

[Xyn]

After upgrading to Calibre version 5 and making the necessary changes, I get the login screen if I access the container via IP:PORT, but if I hit the proxied URL calibre.domain.com, I get a blank screen.

Proxy config should be identical to the sample, except I added a couple blocks to allow me to access the content server at calibre.domain.com/server - which incidentally works just fine:

## Version 2021/05/18
# make sure that your dns has a cname set for calibre

server {
    listen 443 ssl;
    listen [::]:443 ssl;

    server_name calibre.*;

    include /config/nginx/ssl.conf;

    client_max_body_size 0;

    # enable for ldap auth, fill in ldap details in ldap.conf
    #include /config/nginx/ldap.conf;

    # enable for Authelia
    #include /config/nginx/authelia-server.conf;

    location / {
        # enable the next two lines for http auth
        #auth_basic "Restricted";
        #auth_basic_user_file /config/nginx/.htpasswd;

        # enable the next two lines for ldap auth
        #auth_request /auth;
        #error_page 401 =200 /ldaplogin;

        # enable for Authelia
        #include /config/nginx/authelia-location.conf;

        include /config/nginx/proxy.conf;
        include /config/nginx/resolver.conf;
        set $upstream_app calibre;
        set $upstream_port 8080;
        set $upstream_proto http;
        proxy_pass $upstream_proto://$upstream_app:$upstream_port;

        proxy_buffering off;
    }

    location /server/ {
        # enable the next two lines for http auth
        #auth_basic "Restricted";
        #auth_basic_user_file /config/nginx/.htpasswd;

        # enable the next two lines for ldap auth
        #auth_request /auth;
        #error_page 401 =200 /ldaplogin;

        # enable for Authelia
        #include /config/nginx/authelia-location.conf;

        include /config/nginx/proxy.conf;
        include /config/nginx/resolver.conf;
        set $upstream_app calibre;
        set $upstream_port 8081;
        set $upstream_proto http;
        proxy_pass $upstream_proto://$upstream_app:$upstream_port;

        proxy_buffering off;
    }

    location /server {
        rewrite /server /server/ permanent;
    }
}

Here is the microservice config while I’m at it.

calibre:
    image: ghcr.io/linuxserver/calibre
    container_name: calibre
    environment:
      - PUID=${PUID}
      - PGID=${PGID}
      - TZ=${TZ}
      - PASSWORD=${CALIBRE_PASS}
      - CALIBRE_USE_DARK_PALETTE=1
    volumes:
      - /volume1/docker/calibre:/config
      - /volume1/media/Books:/books
    ports:
      - 8080:8080
      - 8081:8081
    restart: unless-stopped

Aside from that, the IP:PORT access only works in Firefox, whereas before the upgrade I was able to hit the proxied URL in Brave without issues. I am guessing this is just Brave’s extra security though, as my normal MO is using Brave, and if a website doesn’t work I try it in Firefox.

[aptalca]

Try the stock proxy conf, and make sure other confs are updated. The logs should tell you which ones need to be updated as long as you’re on a recent SWAG image.

[Xyn]

I assume just having the latest docker image pulled is enough to update the .sample proxy confs?
Reverting to stock proxy conf made no difference - I’m just not sure it’s the latest one. A header comment says it was updated 05/18.

[aptalca]

I meant the other nginx confs. Post a full log and we’ll see.

[Xyn]

I have only ever edited some of the proxy confs (mainly the one for Calibre and Plex).
The nginx confs do look old though by last modified:

I can try deleting those files and restarting the container, I think that’s how you update them?

For the access log, I tried putting it in a code block but it froze up the browser. I tried uploading it but it seems like it only allows image MIME types. I copied it to my Documents to share via Nextcloud, and replaced my WAN IP with a placeholder:

access.log

Let me know if that works.

[Xyn]

Deleting all the conf files in swag/nginx and restarting the container did seem to fix the issue in Firefox, but it’s still not working in Brave. That’s a bummer since it was working fine in Brave before the container was redesigned for v5, but I can live with it, and it’s probably Brave’s fault.