MySQL (`mysql`)¶

The mysql connector indexes table rows from a single MySQL database as searchable records. Each table also exposes schema.json; when [summary].enabled is on, that schema produces a searchable schema_summary. One connector covers one database; register another connector for another database.

How MFS sees it¶

The configured database is the connector scope, so tables sit directly under the alias:

mysql://prod-db/
├── tickets/
│   ├── rows.jsonl     table_rows    → one searchable chunk per row
│   └── schema.json    table_schema  → browsable schema; searchable with summary enabled
└── orders/
    ├── rows.jsonl
    └── schema.json

Rows are chunked per-row and need text_fields to become searchable (see [[objects]]).

Credentials¶

Five fields: host, port, database, user, password. Pull them from your app config, or create a dedicated read-only user:

CREATE USER 'mfs_reader'@'%' IDENTIFIED BY '<password>';
GRANT SELECT ON prod.* TO 'mfs_reader'@'%';

Use the server's egress host instead of % when you can. For managed MySQL, also allow the server's IP or security group before testing. Confirm connectivity from the machine that runs the server before handing credentials to MFS:

mysql -h <host> -P <port> -u <user> -p<pw> <database> -e "SHOW TABLES"

Configuration¶

host = "db.example.com"
port = 3306
database = "prod"
user = "mfs_reader"
password = "env:MYSQL_PASSWORD"
cursor_column = "updated_at"   # strengthens the object fingerprint
max_read_rows = 100000

[[objects]]
match = "/tickets"
text_fields = ["title", "description"]
locator_fields = ["id"]

Keep the password in the server environment, then probe and index:

export MYSQL_PASSWORD='<password>'
mfs connector probe mysql://prod-db --config ./mysql.toml
mfs add mysql://prod-db --config ./mysql.toml

Sync and freshness¶

With cursor_column set (usually updated_at), the connector includes max(cursor_column) in the table object's fingerprint. If the row count or cursor maximum changes, MFS re-reads and re-indexes that table's rows.jsonl object. Deletions are caught by full_scan. grep is a pushdown straight to MySQL.

Search and browse¶

mfs search "billing bug" mysql://prod-db/tickets/rows.jsonl
mfs search "email column" mysql://prod-db --kind schema_summary
mfs cat mysql://prod-db/tickets/schema.json
mfs cat mysql://prod-db/tickets/rows.jsonl --locator '{"id":12345}'

Pitfalls¶

One connector = one database.
No text_fields → browsable rows, but no row search.
If probe fails while a local mysql command works, test from the server host or container; connector credentials are resolved there.
Legacy utf8 (3-byte) collations can return mojibake for 4-byte characters; prefer utf8mb4.
Long scans can hit server timeouts; lower max_read_rows while testing.
schema_summary search requires [summary].enabled; schema.json is still browsable without it.

MySQL (mysql)¶