Replicante bootstrapping procedure

Allow one-command start of a Replicante instance that can be used to configure all datastores and dependences for Replicante itself, then generate a configuration file to be used for production instances.

May be just a guide but will likely require code changes/extra features. Most likely, in-memory version of the dependencies should be made available so the process can start with no additional complexity or impairment.

  • Consider the all-in-one process like jaeger does.
  • Probably introduce some all_in_one cargo feature for specialied “backends” that can be excluded from release builds.
  • Coordinator: in-memory mutexes, elections are all primaries.
  • Store: document-store version of sqlite so I don’t need to implement my thing?
  • Tasks: use crossbeam_channels to fake queues, drop on skip (and retry?).
  • Streams: same as tasks.