Hot Standby Node Stream Distribution Enhancements | Teradata DSA - 17.20 - Hot Standby Node Stream Distribution Enhancements - BAR - Data Stream Architecture

Teradata® DSA User Guide

Product
BAR
Data Stream Architecture
Release Number
17.20
Release Date
October 2021
Content Type
User Guide
Publication ID
B035-3150-101K
Language
English (United States)

Prior to DSA 16.20.50, DSA used an internal ID to distribute the AMP stream. When the system configuration changed, for example, HSN becomes active, DSA did not guarantee the data from the same AMP would go to the same target device.

Beginning with DSA 16.20.50 or later, a new flag, backup.sortByAmp, is defined in the $DSA_CONFIG_DIR/dsc.properties file. It changes how the AMP stream distributes to third party media. With the default value of true, DSA uses the order of the AMP numbers to distribute the stream. This makes sure that DSA always uses the same stream distribution.

After you upgrade to DSA 16.20.50, backup job AMP distribution is different from previous job execution. After a new full backup is run, future AMP streams distributed to third party media are guaranteed as long as the AMP count and target group configuration remains the same, and media servers defined in the target group are all online.

For DDBoost after the upgrade to DSA 16.20.50:
  • When backup.sortByAmp=true (the default), gen0 is generated for all existing customer backup jobs until the first full backup is run. Teradata recommends that you run a full backup as soon as possible. After the full backup is run, the HSN or down node becoming active will not trigger a gen0 backup.
  • When backup.sortByAmp=false, gen0 is generated every time HSN becomes active and again after the down node becomes active. Avoid this by identifying when a gen0 backup may occur and setting the flag to true and running a full backup.