Types of Checkpoints - Parallel Transporter

Teradata Parallel Transporter User Guide
Product
Parallel Transporter
Release Number
15.00
Language
English (United States)
Last Update
2018-09-27
dita:id
B035-2445
lifecycle
previous
Product Category
Teradata Tools and Utilities

Types of Checkpoints

The following table describes the various types of checkpoints taken in Teradata PT jobs.

 

Table 7: Checkpoint Types and Functions 

Checkpoint Type

Function

Basic (default)

If the tbuild command does not specify a checkpoint interval, the job automatically takes just two checkpoints during each job step that has consumer and producer operators:

  • a Start-of-Data checkpoint
  • an End-of-Data checkpoint

    • These two checkpoints allow Teradata PT jobs to restart automatically without requiring user intervention, if the interruption was caused by a Teradata server restart or deadlock situation. The default checkpoints can also be used for a manual restart. In either case, the job will restart after the last checkpoint written to the files.

  • If the End-of-Data checkpoint was taken, the work accomplished between these two checkpoints will not have to be repeated by the restarted job.
  • If the job failed before the End-of-Data checkpoint was taken whatever work was accomplished after the Start-of-Data checkpoint was taken will have to be repeated by the restarted job.

    • This default checkpoint protection against redoing work is quite minimal, so Teradata recommends the use of interval checkpointing.

    Interval Checkpointing

    When you specify a checkpoint interval for a Teradata PT job using the tbuild ‑z command, the job takes a checkpoint for each specified interval (in seconds), for each job step with producer and consumer operators.

    If a job with interval checkpointing fails to run to completion and is later restarted, the only work that will have to be performed over again is the work done after the last checkpoint was taken. This option offers increased fault tolerance for long running jobs containing a substantial amount of data to be loaded/exported.

    Direct Command

    A Teradata PT job can also be directed to take a checkpoint at any time through the twbcmd command, either explicitly with the JOB CHECKPOINT command option, or implicitly with the JOB PAUSE command option, which suspends job execution after the checkpoint is taken.

    For details, see “Teradata PT Utility Commands” in Teradata Parallel Transporter Reference.

    Operator Initiated

    The DataConnector operator automatically initiates a checkpoint after processing all the input files found during an interval-driven scan of a directory.