4.7 KiB
Running a Calckey instance with Docker
It is simple to run an instance of calckey with docker and docker-compose - you can be running your own fedi server in no time.
Pre-built docker container
docker-compose
You can find a docker-compose.yml
file in the same folder as this README
, along with a folder called config
containing two example files needed to get the instance running:
- config/docker.env (db config settings)
- config/default.yml (calckey instance settings)
configuring calckey
Rename the files:
mv config/default_example.yml default.yml
mv config/docker_example.env docker.env
then edit them according to your environment.
You can configure docker.env
with anything you like, but you will have to pay attention to the default.yml
file:
url
should be set to the URL you will be hosting the web interface for the instance at.host
,db
,user
,pass
will have to be configured in the PostgreSQL configuration section -host
is the name of the postgres container (eg: calckey_db_1), and the others should match yourdocker.env
.host
will need to be configured in the Redis configuration section - it is the name of the redis container (eg: calckey_redis_1)
Everything else can be left as-is; unless you are setting up elasticsearch. More on this below.
Running docker-compose
This will take some time. The prebuilt container for calckey is fairly large, and may take a few minutes to download and extract using docker.
Copy docker-compose.yml
and the config/
to the directory from which you intend to run your instance, and then from that directory run the docker-compose command:
docker-compose up -d
.
This will take some time to come fully online, even after download and extracting the container images, and it may emit some error messages before completing successfully. Specifically, the db
container needs to initialize and so isn't available to the web
container right away. Only once the db
container comes online does the web
container start building and initializing the calckey tables.
Once the instance is up you can use a web browser to access the web interface at http://serverip:3000
(where serverip
is the IP of the server you are running the calckey instance on).
Securing your instance with a reverse proxy
On its own calckey serves itself with HTTP, and does not support SSL. In order to support encrypted connections via HTTPS - an absolute necessity if you intend to host an instance accessible from the public internet - you need to add a reverse proxy to your setup.
Adding reverse proxy container
Only swag has been tested, but a similar setup should be possible with caddy, or traefik or a bare metal nginx
install.
- Follow the instructions at https://github.com/linuxserver/docker-swag to set up a separate
swag
container. - Add a second shared external docker network for the
calckey
andswag
containers to talk to each other, ensure both are connected to it via their respectivedocker-compose.yml
files. - configure
<swag-root>/config/nginx/proxy-confs/calckey.subdomain.conf
with the following:
server {
listen 443 ssl;
listen [::]:443 ssl;
server_name yourdomain.whatever;
include /config/nginx/ssl.conf;
client_max_body_size 0;
location / {
include /config/nginx/proxy.conf;
include /config/nginx/resolver.conf;
set $upstream_app calckey_web_1;
set $upstream_port 3000;
set $upstream_proto http;
proxy_pass $upstream_proto://$upstream_app:$upstream_port;
}
}
Note: Make sure you set server_name
in the above to your domain, and $upstream_app
to the name for your calckey container.
Once you bring up both the swag
reverse proxy container with the above configuration, and the calckey
containers, you should be serving your instance and be able to browse to it at https://yourdomain.whatever
.
Enabling elasticsearch
If you decide to enable elasticsearch
, you can uncomment the configuration lines for it in the docker-compose.yml
, but you will also need to configure the Elasticsearch configuration
section of the config/default.yml
.
This is disabled by default, as it uses a great deal of server memory and can increase the load for your instance significantly. It is unlikely that you will need to enable this for a small server, and if you are running a giant server then you probably do not need this README
to begin with.