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-15210) Handle FNFE from S3Guard.getMetadataStore() in S3A initialize()
Date Fri, 16 Feb 2018 00:04:10 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-15210?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16366439#comment-16366439

Aaron Fabbri commented on HADOOP-15210:

Seems like some customers *would* want the fail-fast-and-hard behavior we have today (don't
do anything without S3Guard).. especially since successful S3Guard usage requires all bucket
modifying processes to participate.

If we agree on that, the question becomes: Should we have a "run degraded on S3Guard failure"
config option?

> Handle FNFE from S3Guard.getMetadataStore() in S3A initialize()
> ---------------------------------------------------------------
>                 Key: HADOOP-15210
>                 URL: https://issues.apache.org/jira/browse/HADOOP-15210
>             Project: Hadoop Common
>          Issue Type: Sub-task
>          Components: fs/s3
>    Affects Versions: 3.0.0
>            Reporter: Steve Loughran
>            Priority: Minor
> {{S3Guard.getMetadataStore()}} throws FileNotFoundExceptions up, as the comments say
" rely on callers to catch and treat specially"
> S3A Filesystem doesn't do that, instead it will just fail FileSystem.initialize; the
FNFE  is generated by DynamoDBMetadataStore.
> Are we happy with this? 
> Downgrading has some appeal: if you don't have the table, it will keep going. But failures
could be a sign of bad config, so maybe silent recovery is bad.

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