SQLite schema

The full schema lives under internal/harmonysqlite/schema-curio-core/NNNN_*.sql in the repo. The daemon applies each migration in numeric order on startup and records applied migrations in harmony_schema_migrations.

Key tables

Table	Purpose
harmony_task	Live task queue
harmony_task_history	Completed/failed task records
harmony_task_singletons	Singleton task state (PDPv0_PaySettle, ParkComplete, etc.)
harmony_config	TOML configuration bundle (one row, "base")
harmony_schema_migrations	Tracker of applied migrations
eth_keys	Secp256k1 private keys + role labels
message_sends_eth	Pending + sent FEVM transactions
message_waits_eth	Tx receipt tracking (pending → confirmed)
pdp_data_sets	Active client datasets
pdp_data_set_pieces	Pieces within datasets
pdp_payment_rails	Discovered FilecoinPay rails for this SP (curio-core-specific)
pdp_piece_pulls	Pull-from-URL upload sessions
pdp_piece_pull_items	Per-piece state within a pull session
pdp_piece_streaming_uploads	Streaming-upload sessions
parked_pieces	Local pieces backing dataset entries
parked_piece_refs	URL-to-bytes references for parked pieces
pdp_prove_tasks	Active prove-task bookkeeping
pdp_services	Registered services (FWSS endpoints)
alerts	Operator alerts feed
ipni_head	IPNI announcement tracker (placeholder)

Inspection

sqlite3 /var/lib/curio-core/state.sqlite '.tables'
sqlite3 /var/lib/curio-core/state.sqlite '.schema pdp_data_sets'

Or use the dashboard's Terminal to run config show and other read-only diagnostics.

Direct queries

Curio Core stores all metadata in one SQLite file, which makes it easy to inspect:

# Count active datasets
sqlite3 state.sqlite "SELECT COUNT(*) FROM pdp_data_sets WHERE COALESCE(terminated_at_epoch, 0) = 0"

# Show last 10 prove tasks
sqlite3 -header -column state.sqlite "
  SELECT id, name, substr(work_end, 1, 19) AS ended, result, substr(COALESCE(err, ''), 1, 60) AS err
  FROM harmony_task_history
  WHERE name = 'PDPv0_Prove'
  ORDER BY id DESC LIMIT 10
"

# Total USDFC accumulated (read from chain, not from DB - this is a stand-in)
sqlite3 -header -column state.sqlite "
  SELECT rail_id, last_settled_epoch, last_settle_tx_hash
  FROM pdp_payment_rails
  WHERE terminated = 0
  ORDER BY rail_id
"

Migrations

To add a new migration:

1. Create internal/harmonysqlite/schema-curio-core/NNNN_<name>.sql where NNNN is the next sequential number.
2. Write the SQL. Forward-only — no transactions wrapping multiple statements (SQLite bails on the second-DDL-after-DML pattern that Postgres handles transparently).
3. Add a corresponding entry to PORT-STATUS.md documenting which upstream migration this mirrors (if any).

On the next curio-core run start, the migration runner picks it up and records it in harmony_schema_migrations. There is no downgrade/ SQL by design (see Upgrading for the rollback model).