[jira] [Commented] (HADOOP-14109) improvements to S3GuardTool destroy command

["Steve Loughran (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HADOOP-14109?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15880301#comment-15880301 ] 

Steve Loughran commented on HADOOP-14109:
-----------------------------------------

Options

* Destroy command clears create-table flag to stop table creation
* adds special config option to indicate destroy in progress. Bad as this is is now a secret parameter
* add a new arg, {{boolean aboutToDestroy}}.
* bypass the init code in destroy, instead the destroy command passes in the config and is free to do whatever init logic before destroying the table.

I like the latter the best; the metadata store gets to do what it wants. the key thing is that the init for use is not the same init which init-for-destroy wants or needs

> improvements to S3GuardTool destroy command
> -------------------------------------------
>
>                 Key: HADOOP-14109
>                 URL: https://issues.apache.org/jira/browse/HADOOP-14109
>             Project: Hadoop Common
>          Issue Type: Sub-task
>          Components: fs/s3
>    Affects Versions: HADOOP-13345
>            Reporter: Steve Loughran
>            Priority: Minor
>
> The S3GuardTool destroy operation initializes dynamoDB, and in doing so has some issues
> # if the version of the table is incompatible, init fails, so table isn't deleteable
> # if the system is configured to create the table on demand, then whenever destroy is called for a table that doesn't exist, it gets created and then destroyed.



--
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