You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-issues@hadoop.apache.org by "Sean Mackrory (JIRA)" <ji...@apache.org> on 2017/03/07 00:26:32 UTC

[jira] [Updated] (HADOOP-14027) Implicitly creating DynamoDB table ignores endpoint config

     [ https://issues.apache.org/jira/browse/HADOOP-14027?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Sean Mackrory updated HADOOP-14027:
-----------------------------------
    Attachment: HADOOP-14027-HADOOP-13345.001.patch

Attaching a patch that will allow a configured region to override the S3 region, even when the metadata store is initialized with an S3 bucket.

Several tests started failing after I did this. Some of them have failed for a while because of any time the configured table name differs from the bucket name (so only kinda related to this change in that it's a bigger deal when testing this specific kind of setup, but a pre-existing issue). Others don't really make sense why they're suddenly failing after my change (but I've confirmed they reliably passed before) - but the changes required to make them work aren't surprising at all - they all seem like the way the test should have been all along.

Ran all tests with and without S3Guard, split between a bucket in us-west-1 and tables in us-west-2.

> Implicitly creating DynamoDB table ignores endpoint config
> ----------------------------------------------------------
>
>                 Key: HADOOP-14027
>                 URL: https://issues.apache.org/jira/browse/HADOOP-14027
>             Project: Hadoop Common
>          Issue Type: Sub-task
>          Components: fs/s3
>            Reporter: Sean Mackrory
>            Assignee: Sean Mackrory
>         Attachments: HADOOP-14027-HADOOP-13345.001.patch
>
>
> When you're using the 'bin/hadoop s3a init' command, it correctly uses the endpoint provided on the command-line (if provided), it will then use the endpoint in the config (if provided), and failing that it will default to the same region as the bucket.
> However if you just set fs.s3a.s3guard.ddb.table.create to true and create a directory for a new bucket / table, it will always use the same region as the bucket, even if another endpoint is configured.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

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