Closed der-eismann closed 4 years ago
You can remove ‘:8000’ in ccnet.conf and run the 'docker-compose restart' command to restart.
This is a special configuration, so you need to change the configuration according to the actual situation. Sorry for some impact on your use.
Yes that's what I did, but it's not a special configuration. This docker stack is supposed to run on port 80 & 443, so the SERVICE_URL
should be set by the bootstrapper to the same value as FILE_SERVER_ROOT
(without the seafhttp part).
The idea of containers is to pass everything inside with environment variables and have it running without adjusting configuration by hand.
However, some users will use ports other than 80 and 443, and this will be the same.
So instead of setting a usable default for some users everyone has to adjust it.
Why was this issue closed ? I had the same issue and did the same workaround. But its not only the port that's causing troubles. If you are using SSL, the browser will warn you about other content not using SSL.
SERVICE_URL
should be an environment variable.
If its not present then set the value to http://the-domain-here:8000
or what ever the default value is.
Just realised that the http issue also affects the FILE_SERVER_ROOT
variable in seahub_settings.py.
What does the SEAFILE_SERVER_LETSENCRYPT
environment variable actually do? Does it try and get a certificate from LE or does it only change http
to https
in the configs?
I just created a new Seafile instance with the
seafileltd/seafile-mc:7.0.5
image and I noticed a bug which I couldn't fix myself.After the service is up and running the small avatar in the upper right corner can't be loaded because it uses the wrong URL. My envvars are set as follows:
However the request for the small avatar is pointing at
Notice the port
8000
? That should be 80. The only value with 8000 in the config is theSERVICE_URL
in theccnet.conf
and I have no idea where this gets its value from.It would be great if the bootstrapper could set the correct value here.