Use Multiple Agents for PT API Jobs | Teradata Data Mover - Using Multiple Agents for Jobs That Use Teradata PT API - Teradata Data Mover

Teradata® Data Mover User Guide - 20.01

Deployment
VantageCloud
VantageCore
Edition
Enterprise
IntelliFlex
Lake
VMware
Product
Teradata Data Mover
Release Number
20.01
Published
November 2023
Language
English (United States)
Last Update
2023-12-05
dita:mapPath
hlv1700545853003.ditamap
dita:ditavalPath
mpm1591127278842.ditaval
dita:id
don1467241476387
Product Category
Analytical Ecosystem

Multiple agents can be used to copy a single large database object such as a table or a database with the create or move command for jobs that use Teradata PT API by editing the XML file for the command.

  1. Add the max_agents_per_task element.
  2. Set the value of the max_agents_per_task element to an integer number greater than one, but less than or equal to the number of installed, online agents.
    This value corresponds to the maximum number of agents that are allocated for copying each database object in the job.
  3. Set the value of the data_streams element to an integer number greater than 1.
    Work is spread across multiple agents by assigning each agent one or more data streams.
  4. For Teradata PT API only, set the value of source_sessions and target_sessions to an integer number greater than or equal to the value of data_streams.
    Teradata PT API distributes sessions among the data streams. Because each Teradata PT API data stream requires at least one source and one target session, the number of source and target sessions must be greater than or equal to the number of data streams.
    The parameter values for which all conditions must be true for Teradata PT API are:
    Parameters Product
    max_agents_per_task > 1 Teradata PT API
    max_agents_per_task <= number-of-online-agents Teradata PT API
    data_streams > 1 Teradata PT API
    source_sessions >= data_streams Teradata PT API
    target_sessions >= data_streams Teradata PT API

    If you do not specify a value, Data Mover calculates max_agents_per_task dynamically for optimal performance.

    The following table describes what occurs with different values for the data_streams, max_agents_per_task, source_sessions, and target_sessions parameters.
    Value Description
    max_agents_per_task is greater than the number of online agents.

    Teradata PT API:

    Sets max_agents_per_task equal to the number of online agents.
    data_streams is greater than max_agents_per_task

    Teradata PT API:

    Optimizes the distribution of the available data streams evenly among the online agents.
    max_agents_per_task is greater than the value of data_streams. The number of agents used for running a task is equal to the value of data_streams.
    data_streams and max_agents_per_task are not provided.

    Teradata PT API:

    Optimal Data streams is first computed, then sets the value of max_agents_per_task equal to the computed data_streams.
    source_sessions or target_sessions is less than data_streams.

    Teradata PT API only:

    Sets source_sessions or target_sessions equal to the number of data streams.
    source_sessions or target_sessions are not provided.

    Teradata PT API:

    Sets source_sessions or target_sessions dynamically to provide optimal performance.
    data_streams is provided and max_agents_per_task is not provided.

    Teradata PT API:

    Sets the value of max_agents_per_task equal to the data_streams.
    data_streams is not provided and max_agents_per_task is provided.

    Teradata PT API:

    Optimal data streams is computed. If max_agents_per_task is specified and the value is less than or equal to the number of computed data streams, the specified value is used; otherwise, the max_agents_per_task value is set to match number of computed data streams.