hadoop-common-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Aaron Fabbri (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HADOOP-14971) Merge S3A committers into trunk
Date Mon, 30 Oct 2017 20:17:00 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-14971?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16225666#comment-16225666

Aaron Fabbri commented on HADOOP-14971:

Thanks for the notes [~stevel@apache.org].  Looks like you got the major points.  We also

* Compatibility. Looks like existing OutputCommitter APIs and configurations are still supported.
 We didn't think any breaks in compatibility are introduced by the patch.
* Retries: The policy is all in one place in the {{S3ARetryPolicy}} constructor. Review of
current implementation: looks good for now. 
* "idempotent" delete: There doesn't seem to be an obvious best answer here. Current patch
does not make the flag {{DELETE_CONSIDERED_IDEMPOTENT}} configurable; it is hard-coded to
{{true}}.  While delete is not idempotent, treating it like it is for retries makes sense
given that S3 is not strongly consistent, and thus multi-client delete/modify scenarios are
already racy. This allows us to retry on transient S3 service issues and get more robust job
completion under throttling, etc..  I feel like making the idempotent flag configurable would
be safest, but I understand the counterpoint of "too much configurability".  I'm going to
defer to [~stevel@apache.org] on this, so I'm +1 on this part of the current patch.

I think this stuff is ready to go in once we (1) restore separate SDK/S3A retry config knobs
(2) finish any testing we want to do.

> Merge S3A committers into trunk
> -------------------------------
>                 Key: HADOOP-14971
>                 URL: https://issues.apache.org/jira/browse/HADOOP-14971
>             Project: Hadoop Common
>          Issue Type: Sub-task
>          Components: fs/s3
>    Affects Versions: 3.0.0
>            Reporter: Steve Loughran
>            Assignee: Steve Loughran
> Merge the HADOOP-13786 committer into trunk. This branch is being set up as a github
PR for review there & to keep it out the mailboxes of the watchers on the main JIRA

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org

View raw message