hadoop-common-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Loughran (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HADOOP-13059) S3a over-reacts to potentially transient network problems in its init() logic
Date Tue, 12 Sep 2017 19:40:00 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-13059?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16163533#comment-16163533
] 

Steve Loughran commented on HADOOP-13059:
-----------------------------------------

the new lambda-operator retry logic will retry in init(), but it'll give up there if things
aren't working. that way, you're allowed a brief bit of failure with the same retry logic
as other idempotent calls get.

> S3a over-reacts to potentially transient network problems in its init() logic
> -----------------------------------------------------------------------------
>
>                 Key: HADOOP-13059
>                 URL: https://issues.apache.org/jira/browse/HADOOP-13059
>             Project: Hadoop Common
>          Issue Type: Sub-task
>          Components: fs/s3
>    Affects Versions: 2.8.0
>            Reporter: Steve Loughran
>            Assignee: Steve Loughran
>            Priority: Minor
>         Attachments: HADOOP-13059-001.patch
>
>
> If there's a reason for s3a not being able to connect to AWS, then the constructor fails,
even if this is a potentially transient event.
> This happens because the code to check for a bucket existing will relay the exceptions.
> The constructor should catch IOEs against the remote FS, downgrade to warn and let the
code continue; it may fail later, but it may also recover.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

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


Mime
View raw message