Deployment with public address for supervisor

[heikkilv]

• Added environmental variables in order to be able send deploy commands with a publicly available URL
• Also, changed the MongoDB connection environmental variables to be more general (and better supprot running orchestrator with a separated Mongo instance)
• Added a separate docker-compose file for running the orchestrator in a separate network than the device

The default values for the environmental variables should allow same behavior as before these changes but it is possible that some development environment breaks or needs adjusting.

Note: When testing the deployed Wasm modules in supervisor the following error was repeatedly given: "RuntimeError: underrun while parsing Wasm binary". It is unknown to me whether this is caused by my changes or something else.

Closes #26

[trkks]

Note: When testing the deployed Wasm modules in supervisor the following error was repeatedly given: "RuntimeError: underrun while parsing Wasm binary". It is unknown to me whether this is caused by my changes or something else.

This error was caused by the module download-URL sent from orchestrator being malformed (two '/' at beginning of path) and supervisor not being able to recognize this as the orchestrator responded with HTTP 200 (apparently the postfixed status call did not set the desired code) on bad paths.

The error message when running Wasm on supervisor was then caused by the module file containing not Wasm binary but the response string "Bad URL".

[heikkilv]

The error message when running Wasm on supervisor was then caused by the module file containing not Wasm binary but the response string "Bad URL".

Nice catch, should have checked that the deployed file was actually the Wasm file and not just that the file existed.