[BEAM-3217] Add a performance test for HadoopInputFormatIO

• Type: Test
• Status: Resolved
• Priority: P2
• Resolution: Done
• Fix Version/s: 2.4.0
• Component/s: io-java-hadoop-format
• Assignee: Lukasz Gajowy
• Reporter: Chamikara Madhusanka Jayalath
• Created: 17/Nov/17 09:58
• Updated: 16/May/20 13:57
• Resolved: 12/Jun/18 09:54

Description

We should add a large scale performance test for HadoopInputFormatIO. We should use PerfKitBenchmarker based performance testing framework [1] to manage Kubernetes based muti-node data store and to publish benchmark results.

Example input format implementation to use: DBInputFormat to connect to a Postgres instance.
https://github.com/hanborq/hadoop/blob/master/src/mapred/org/apache/hadoop/mapreduce/lib/db/DBInputFormat.java

Example docker image to use: https://hub.docker.com/_/postgres/

[1] https://beam.apache.org/documentation/io/testing/

Issue Links

• links to: GitHub Pull Request #4332
• links to: GitHub Pull Request #4758