Executor options

Executor options for the dsbulk command

Specify executor options for the dsbulk command.

The options can be used in short form (-k keyspace_name) or long form (--schema.keyspace keyspace_name).

--executor.maxPerSecond number

The maximum number of concurrent operations per second. This acts as a safeguard to prevent more requests than the cluster can handle. Batch statements are counted by the number of statements included. Reduce this setting when the latencies get too high and a remote cluster cannot keep up with throughput, as dsbulk requests will eventually time out. Extend the functionality to loading by making it take into account the number of received rows, not the number of requests sent, and applies to rate limiting for both loading and unloading. Setting this option to any negative value will disable it.

Default: -1

--executor.continuousPaging.enabled ( true | false )

Enable or disable continuous paging. If the target cluster does not support continuous paging, or if driver.query.consistency is not ONE or LOCAL_ONE, traditional paging will be used regardless of this setting. Used for unloading only.

Default: true

--executor.continuousPaging.maxPages number

The maximum number of pages to retrieve. Setting this value to zero retrieves all pages available.

Default: 0

--executor.continuousPaging.maxPagesPerSecond number

The maximum number of pages per second. Setting this value to zero indicates no limit.

Default: 0

--executor.continuousPaging.pageSize number

The size of the page. The unit to use is determined by the pageUnit setting.

Default: 5000

--executor.continuousPaging.pageUnit ( ROWS | BYTES )

The unit to use for the pageSize setting.

Default: ROWS

--executor.maxInFlight number

The maximum number of "in-flight" requests, or maximum number of concurrent requests waiting for a response from the server. This acts as a safeguard to prevent more requests than the cluster can handle. Batch statements count as one request. Reduce this value when the throughput for reads and writes cannot match the throughput of mappers; this is usually a sign that the workflow engine is not well calibrated and will eventually run out of memory. Setting this option to any negative value will disable it.

Default: 1024