[jira] [Work logged] (BEAM-5036) Optimize FileBasedSink's WriteOperation.moveToOutput()

[ASF GitHub Bot (JIRA)]

     [ 
https://issues.apache.org/jira/browse/BEAM-5036?focusedWorklogId=159608&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-159608
 ]

ASF GitHub Bot logged work on BEAM-5036:
----------------------------------------

                Author: ASF GitHub Bot
            Created on: 27/Oct/18 11:36
            Start Date: 27/Oct/18 11:36
    Worklog Time Spent: 10m 
      Work Description: timrobertson100 edited a comment on issue #6843: 
[BEAM-5036] Optimize the FileBasedSink WriteOperation.moveToOutput()
URL: https://github.com/apache/beam/pull/6843#issuecomment-433612987
 
 
   The performance tests have almost 100% variation in runtime across various 
runs so seem mostly unhelpful.  Attached are 2 screenshots showing the result 
of this PR for a simple Avro file rewrite.
   
   Without this PR
   <img width="1479" alt="screenshot 2018-10-27 13 24 38" 
src="https://user-images.githubusercontent.com/237221/47603493-ddb5f980-d9ec-11e8-9cfd-4f25f5402291.png";>
   
   With this PR
   <img width="1480" alt="screenshot 2018-10-27 12 52 32" 
src="https://user-images.githubusercontent.com/237221/47603498-e5759e00-d9ec-11e8-9c6a-9e28f4ff62c1.png";>
   
   This is using [this 
code](https://github.com/gbif/beam-perf/tree/master/avro-to-avro) and executed 
as 
   ```
   spark2-submit --class org.gbif.beam.perf.avrotoavro.BeamAvroIOTest   --conf 
spark.dynamicAllocation.enabled=false   --master yarn --executor-memory 16G 
--executor-cores 5 --num-executors 10   avro-to-avro-0.1-SNAPSHOT-shaded.jar 
--runner=SparkRunner   
--source=hdfs:///user/hive/warehouse/tim.db/occurrence_avro/*   
--target=hdfs:///tmp/beam4
   ```

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


Issue Time Tracking
-------------------

    Worklog Id:     (was: 159608)
    Time Spent: 12h 50m  (was: 12h 40m)

> Optimize FileBasedSink's WriteOperation.moveToOutput()
> ------------------------------------------------------
>
>                 Key: BEAM-5036
>                 URL: https://issues.apache.org/jira/browse/BEAM-5036
>             Project: Beam
>          Issue Type: Improvement
>          Components: io-java-files
>    Affects Versions: 2.5.0
>            Reporter: Jozef Vilcek
>            Assignee: Tim Robertson
>            Priority: Major
>          Time Spent: 12h 50m
>  Remaining Estimate: 0h
>
> moveToOutput() methods in FileBasedSink.WriteOperation implements move by 
> copy+delete. It would be better to use a rename() which can be much more 
> effective for some filesystems.
> Filesystem must support cross-directory rename. BEAM-4861 is related to this 
> for the case of HDFS filesystem.
> Feature was discussed here:
> http://mail-archives.apache.org/mod_mbox/beam-dev/201807.mbox/%3CCAF9t7_4Mp54pQ+vRrJrBh9Vx0=uaknupzd_qdh_qdm9vxll...@mail.gmail.com%3E



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)