Problem: `server sync` is redundant, and should be removed to avoid confusion

[robinbb]

'mina-indexer server sync' is the same as 'mina-indexer server start' if you have a database, but fails if you don't have a database. So, it is useless, and confusing. Remove 'sync'.

[robinbb]

... with the understanding that the functionality of 'sync' is mostly preserved by way of 'start'.

[jhult]

Adding comments from internal Slack:

start initializes the Indexer from a directory of blocks/staking ledgers without the presence of a database; maybe a "better" name is new??

More below:

This is how @Isaac-DeFrain thinks about the 3 initialization modes:

• start

  • In the absence of a database, ingest blocks and staking ledgers with canonical chain discovery
  • In the presence of a database, start defaults to sync

• sync

  • Requires existing database/snapshot
  • This is our "quick sync" feature which rebuilds the in-memory witness tree from the supplied database/snapshot; if you trust the source of the snapshot, this is what you would use as it's more performant than replay

• replay

  • requires existing database/snapshot
  • This "replays" all database events to rebuild the witness and does some data consistency checks in the process

[jhult]

Suggestion for redesigned CLI semantics:

• [ ] start
  • Rebuilds the in-memory witness tree from a (required) pre-populated database directory
  • Watches for new blocks & staking ledgers
  • Optional flags:
    • [x] --config: Use a config file instead of CLI args
    • [ ] --self-check: This "replays" all database events to rebuild the witness and does some data consistency checks in the process; by default, invisible to users
• [ ] create-database
  • Ingest blocks and staking ledgers with canonical chain discovery and then call create-snapshot
  • Required flag --output-path

[jhult]

The create-database command should solve #1203

[robinbb]

Assigning @Isaac-DeFrain because he has #1203 and solving that one will resolve this one.