[jira] [Assigned] (YARN-5312) Parameter 'size' in the webservices "/containerlogs/$containerid/$filename" and in AHSWebServices is semantically confusing

["Vinod Kumar Vavilapalli (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/YARN-5312?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Vinod Kumar Vavilapalli reassigned YARN-5312:
---------------------------------------------

    Assignee: Vinod Kumar Vavilapalli

> Parameter 'size' in the webservices "/containerlogs/$containerid/$filename" and in AHSWebServices is semantically confusing
> ---------------------------------------------------------------------------------------------------------------------------
>
>                 Key: YARN-5312
>                 URL: https://issues.apache.org/jira/browse/YARN-5312
>             Project: Hadoop YARN
>          Issue Type: Bug
>            Reporter: Vinod Kumar Vavilapalli
>            Assignee: Vinod Kumar Vavilapalli
>
> This got added in YARN-5088 and I found this while reviewing YARN-5224.
> bq. Also, the parameter 'size' in the API "/containerlogs/$containerid/$filename" and similarly in AHSWebServices is confusing with semantics. I think we are better off with an offset and size.
> An offset (in bytes, +ve to indicate from the start and -ve to indicate from the end) together with a size (in bytes) indicating how much to read from the offset are a better combination - this is how most file-system APIs look like, for comparison.
> I can also imagine number of lines as a better unit than bytes for offset and size - perhaps yet another ticket.
> /cc [~vvasudev].



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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