[
https://issues.apache.org/jira/browse/HADOOP-14023?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15838463#comment-15838463
]
Sean Mackrory commented on HADOOP-14023:
----------------------------------------
So my guess would appear to be right. Do note that it appears if you go through
the code path in which the incorrect region gets picked up, it's only ever used
in log and exception messages. So not the end of the world. But I'm not
immediately seeing a more correct way to get the region you have connected to...
> S3Guard: DynamoDBMetadataStore logs nonsense region
> ---------------------------------------------------
>
> Key: HADOOP-14023
> URL: https://issues.apache.org/jira/browse/HADOOP-14023
> Project: Hadoop Common
> Issue Type: Sub-task
> Reporter: Sean Mackrory
>
> When you specify a DynamoDB endpoint instead of it using the same one as the
> S3 bucket, it thinks the region is "dynamodb":
> {code}
> > bin/hadoop s3a init -m dynamodb://region-test
> 2017-01-25 11:41:28,006 INFO s3guard.S3GuardTool: create metadata store:
> dynamodb://region-test scheme: dynamodb
> 2017-01-25 11:41:28,116 WARN util.NativeCodeLoader: Unable to load
> native-hadoop library for your platform... using builtin-java classes where
> applicable
> 2017-01-25 11:41:29,421 INFO s3guard.DynamoDBMetadataStore: Creating
> non-existent DynamoDB table region-test in region dynamodb
> 2017-01-25 11:41:34,637 INFO s3guard.S3GuardTool: Metadata store
> DynamoDBMetadataStore{region=dynamodb, tableName=region-test} is initialized.
> {code}
> My guess is it's because of DynamoDBMetadataStore.initialize will call
> getEndpointPrefix, and all the published endpoints for DynamoDB actually
> start with dynamodb.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
