You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@beam.apache.org by "ASF GitHub Bot (JIRA)" <ji...@apache.org> on 2016/10/25 03:21:58 UTC
[jira] [Commented] (BEAM-814) Improve performance when staging
files
[ https://issues.apache.org/jira/browse/BEAM-814?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15604035#comment-15604035 ]
ASF GitHub Bot commented on BEAM-814:
-------------------------------------
GitHub user dhalperi opened a pull request:
https://github.com/apache/incubator-beam/pull/1184
[BEAM-814] Dataflow: parallelize hashing/compressing of files to stage
R: @lukecwik
This is a WIP for the linked issue. Early feedback appreciated.
Decided to parallelize just hashing/compressing separately from uploading, because this should use a very tiny buffer per file and so can be parallelized nearly arbitrarily. In contrast, uploading uses a very large buffer and cannot be parallelized simply.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/dhalperi/incubator-beam dataflow-runner-speedup
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/incubator-beam/pull/1184.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #1184
----
commit 04c7e1cc8596a7be8496e8f7e1e186e727366409
Author: Dan Halperin <dh...@google.com>
Date: 2016-10-25T00:27:23Z
PackageUtil: remove deprecated, unused function
commit b625e840b78a8dd5686be66bf949b8e42dba42db
Author: Dan Halperin <dh...@google.com>
Date: 2016-10-25T02:20:53Z
PackageUtil: parallelize hashing of files
----
> Improve performance when staging files
> --------------------------------------
>
> Key: BEAM-814
> URL: https://issues.apache.org/jira/browse/BEAM-814
> Project: Beam
> Issue Type: Improvement
> Components: runner-dataflow
> Reporter: Daniel Halperin
> Assignee: Daniel Halperin
> Priority: Minor
>
> Right now, {{PackageUtil}} hashes/compresses/sizes files, checks if they have already been staged, and uploads them if not, completely serially.
> There is some low-hanging fruit to parallelize this.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)