[TEZ-4137] Input/Output/Processor should merge payload to local conf

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Assignee: Mustafa İman
• Reporter: Mustafa İman
• Created: 01/Apr/20 06:53
• Updated: 12/Jun/20 05:11
• Resolved: 27/May/20 05:40

Description

This patch introduces config merging to various Input and Output processors. As described in https://issues.apache.org/jira/browse/TEZ-4073 , we need to reduce the size of the configuration objects transferred over the wire. There are two improvements we are planning to do regarding to that:

1. Skip sending default configs and configuration coming from xml files in payload
2. Send dag, vertex and session configurations in layers instead of sending dag + vertex + session configs all together three times.

In order to achieve these,

• We need to expose local config on Task side through TaskContext.
• Input/Output/Processors must merge the config from user payload to local config in their TaskContext

Since runtime components did not have access to local config before, tez clients sent all config required at runtime in user payload. After this change, tez clients can reduce their payload size.

Issue Links

• is depended upon by: HIVE-23175 Skip serializing hadoop and tez config on HS side