apache / apache/beam

Slow GCS uploads for large Java staging jars

Open
#19,949 0 comments 0 reactions 0 assignees View on GitHub
dataflow improvement P3 runners
Dominant language
Java
Stars
8.7k
Forks
4.7k
Avg merge
1d 20h
Merged PRs (30d)
196

Description

The default and max upload chunk size is 1M, in 

[https://github.com/apache/beam/blob/master/runners/google-cloud-dataflow-java/src/main/java/org/apache/beam/runners/dataflow/util/GcsStager.java#L96](https://github.com/apache/beam/blob/master/runners/google-cloud-dataflow-java/src/main/java/org/apache/beam/runners/dataflow/util/GcsStager.java#L96)

This should be at least 8M to improve performance.

8m is the default python write chunk size

[https://github.com/apache/beam/blob/master/sdks/python/apache_beam/io/gcp/gcsio.py#L86](https://github.com/apache/beam/blob/master/sdks/python/apache_beam/io/gcp/gcsio.py#L86)

 

Imported from Jira [BEAM-8832](https://issues.apache.org/jira/browse/BEAM-8832). Original Jira may contain additional context.
Reported by: amyrvold.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.