3.5 KiB
Mare Synchronos Docker Setup
This is primarily aimed at developers who want to spin up their own local server for development purposes without having to spin up a VM. Obligatory requires Docker to be installed on the machine.
There are two directories: build and run
1. build images
There is two ways to build the necessary docker images which are differentiated by the folders -local and -git
- -local will run the image build against the current locally present sources
- -git will run the image build against the latest git main commit
It is possible to build all required images at once by runningdocker-build.bat/sh(Server, Servies, StaticFilesServer) or all 3 separately withdocker-build-<whatever>.bat/sh
2. Configure ports + token
You should set up 2 environment variables that hold server specific configuration and open up ports.
The default ports used through the provided configuration are 6000 for the main server and 6200 as well as 6201 for the files downloads.
Both ports should be open to your computer through your router if you wish to test this with clients.
Furthermore there are two environment variables DEV_MARE_CDNURL and DEV_MARE_DISCORDTOKEN which you are required to set.
DEV_MARE_CDNURL should point to http://<yourip or dyndns>:6200/cache/ and DEV_MARE_DISCORDTOKEN is an oauth token from a bot you need to create through the Discord bot portal.
You should also set DEV_MARE_CDNURL2 to http://<yourip or dyndns>:6201/cache/
It is enough to set them as User variables. The compose files refer to those environment variables to overwrite configuration settings for the Server and Services to set those respective values.
It is also possible to set those values in the configuration.json files themselves.
Without a valid Discord bot you will not be able to register accounts without fumbling around in the PostgreSQL database.
3. Run Mare Server
The run folder contains two major Mare configurations which is standalone and sharded.
Both configurations default to port 6000 for the main server connection and 6200 for the files downloads. Sharded configuration additionally uses 6201 for downloads. No HTTPS.
All appsettings.json configurations provided are extensive at the point of writing, note the differences between the shard configurations and the main servers respectively.
They can be used as examples if you want to spin up your own servers otherwise.
The scripts to start the respective services are divided by name, the daemon-start/stop files use compose up -d to run it in the background and to be able to stop the containers as well.
The respective docker-compose files lie in the compose folder. I would not recommend editing them unless you know what you are doing.
All data (postgresql and files uploads) will be thrown into the data folder after startup.
All logs from the mare services will be thrown into logs, divided by shard, where applicable.
The standalone configuration features PostgeSQL, Mare Server, Mare StaticFilesServer and Mare Services.
The sharded configuration features PostgreSQL, Redis, HAProxy, Mare Server Main, 2 Mare Server Shards, Mare Services, Mare StaticFilesServer Main and 2 Mare StaticFilesServer Shards.
Haproxy is set up that it takes the same ports as the standalone configuration and distributes the connections between the shards.
In theory it should be possible to switch between the standalone and sharded configuration by shutting down one composition container and starting up the other. They share the same Database.