[jira] [Commented] (HADOOP-14288) TestDynamoDBMetadataStore is broken unless we can fail faster without a table version

["Sean Mackrory (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HADOOP-14288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15961076#comment-15961076 ] 

Sean Mackrory commented on HADOOP-14288:
----------------------------------------

DynamoDBMetadataStore.getVersionMarkerItem is not incrementing retryCount. That's the biggest problem - the test will actually get the failure it expects in about 2 minutes when this times out but only if you increment retryCount.

I hadn't realized this was sharing the retry policy with batch writes. Which is fine as this should be a rare / short-lived issue, and the instance is immutable once configured. The configs for it are already pretty generic. Will reuse the same property to crank the max retries way down and we should be good to go.

> TestDynamoDBMetadataStore is broken unless we can fail faster without a table version
> -------------------------------------------------------------------------------------
>
>                 Key: HADOOP-14288
>                 URL: https://issues.apache.org/jira/browse/HADOOP-14288
>             Project: Hadoop Common
>          Issue Type: Bug
>            Reporter: Sean Mackrory
>            Priority: Critical
>
> HADOOP-14215 causes TestDynamoDBMetadataStore to fail because not having a version meta-entry in the table no longer fails immediately, but must timeout all the retries.
> Let's make those configurable and have the test just turn them way down low.



--
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