Snapshots properties
Generated from the connector's ConfigDef by ConfigDocsGeneratorTest; do not edit by hand.
| Property | Type | Default | Importance | Description |
|---|---|---|---|---|
cdc.snapshot.mode | string | initial | high | initial reads every captured table's existing rows on the connector's first start, while streaming, then streams on; none only streams; snapshot_only reads the tables and then stays idle without streaming; on_signal starts no snapshot by itself. A snapshot that a stored offset records as unfinished resumes in every mode. |
cdc.snapshot.threads | int | 4 | medium | Chunks of a table read in parallel, each on a connection of its own. |
cdc.snapshot.chunk.rows | int | 100000 | medium | Target rows per chunk. Each chunk is one flashback query, so smaller chunks need less undo; up to cdc.snapshot.max.pending.chunks chunks are held in memory. |
cdc.snapshot.chunk.retries | int | 5 | low | Times a failed chunk is read again with a fresh SCN before the task stops. |
cdc.snapshot.fetch.size | int | 5000 | low | JDBC fetch size of chunk reads. |
cdc.snapshot.max.pending.chunks | int | 8 | low | Chunks read but not yet published before reads pause; chunks wait until streaming has passed their SCN. |
cdc.snapshot.tables.order | list | empty | low | Tables to read first, as PDB.OWNER.TABLE; the others follow in name order. A filter for one table's snapshot goes in cdc.snapshot.select.override.PDB.OWNER.TABLE as a SQL condition, for example cdc.snapshot.select.override.FREEPDB1.APP.ORDERS=STATUS <> 'ARCHIVED'. |