[
https://issues.apache.org/jira/browse/HADOOP-17318?focusedWorklogId=509771&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-509771
]
ASF GitHub Bot logged work on HADOOP-17318:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 10/Nov/20 15:45
Start Date: 10/Nov/20 15:45
Worklog Time Spent: 10m
Work Description: steveloughran commented on pull request #2399:
URL: https://github.com/apache/hadoop/pull/2399#issuecomment-724786879
some more detail for the watchers from my testing (hadoop-trunk + CDP spark
2.4). I could not get spark master and hadoop trunk to build together this week.
* RDD.saveAs needs to pass down the setting too
[https://issues.apache.org/jira/browse/SPARK-33402](https://issues.apache.org/jira/browse/SPARK-33402)
* I'm getting errors with FileSystem instantiation in Hive and the isolated
classloader
[https://issues.apache.org/jira/browse/HADOOP-17372](https://issues.apache.org/jira/browse/HADOOP-17372).
I'm not going near that other than to add a para in troubleshooting.md
saying "you're in classloader hell". Will need to be testing against spark
master before worrying about WTF is going on there
I'm also now worried that if anyone does >1 job with the same dest dir and
overwrite=true, then there's a risk that you get the same duplicate app attempt
ID race condition. It's tempting just to do something ambitious like use a
random number to generate a timestamp for the cluster launch, or some
random(year-month-day)+ seconds-of-day, so that this problem goes away almost
completely
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 509771)
Time Spent: 3h 10m (was: 3h)
> S3A committer to support concurrent jobs with same app attempt ID & dest dir
> ----------------------------------------------------------------------------
>
> Key: HADOOP-17318
> URL: https://issues.apache.org/jira/browse/HADOOP-17318
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/s3
> Affects Versions: 3.3.0
> Reporter: Steve Loughran
> Assignee: Steve Loughran
> Priority: Minor
> Labels: pull-request-available
> Time Spent: 3h 10m
> Remaining Estimate: 0h
>
> Reported failure of magic committer block uploads as pending upload ID is
> unknown. Likely cause: it's been aborted by another job
> # Make it possible to turn off cleanup of pending uploads in magic committer
> # log more about uploads being deleted in committers
> # and upload ID in the S3aBlockOutputStream errors
> There are other concurrency issues when you look close, see SPARK-33230
> * magic committer uses app attempt ID as path under __magic; if there are
> duplicate then they will conflict
> * staging committer local temp dir uses app attempt id
> Fix will be to have a job UUID which for spark will be picked up from the
> SPARK-33230 changes, (option to self-generate in job setup for hadoop 3.3.1+
> older spark builds); fall back to app-attempt *unless that fallback has been
> disabled*
> MR: configure to use app attempt ID
> Spark: configure to fail job setup if app attempt ID is the source of a job
> uuid
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
