KineticaDB (deprecated)
Supported pipeline types:
|
When you configure the KineticaDB destination, you specify the URL for the Kinetica head node, the credentials for the connection, and the table name. You specify the batch size for the bulk inserter and whether to compress the data before passing it to Kinetica.
When necessary, you can disable multihead ingest, and you can specify a regular expression to filter the IP addresses that the bulk inserter uses.
Multihead Ingest
By default, the KineticaDB destination writes to Kinetica using multihead ingest when possible.
When using multihead ingest, the destination can send data directly to the appropriate shard manager. When writing to a replicated table, the destination passes the data only to the head node, which then replicates the data as expected.
You can configure the Kinetica DB destination to send data only to the Kinetica head node instead. You might need to disable multihead ingest, for example, when the Kinetica worker nodes reside behind a firewall.
To disable multihead ingest, select the Disable Multihead Ingest property on the Connection tab. For more information about multihead ingestion, see the Kinetica documentation.
Inserts and Updates
By default when writing to Kinetica, the KineticaDB destination inserts all new records. If the destination finds an existing record in the table with the same primary key, it leaves the existing record as is, and discards the new record.
You can configure the destination to replace the existing record instead. To replace existing records with the same primary key, select the Update on Existing PK property on the Table tab.