hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Wangda Tan (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-4843) [Umbrella] Revisit YARN ProtocolBuffer int32 usages that need to upgrade to int64
Date Mon, 21 Mar 2016 17:06:25 GMT

    [ https://issues.apache.org/jira/browse/YARN-4843?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15204621#comment-15204621
] 

Wangda Tan commented on YARN-4843:
----------------------------------

[~kasha], actually we can do this in compatible way like what [~djp] suggested: add two int64
fields, deprecate int32 fields. User can set/get both fields, but int64 fields will be preferred.

I would suggest to keep compatibility even we ship 3.0.0, all YARN applications use the Resource
object frequently, it gonna be hard to ask everyone update lots of source code to use 3.0.0.

> [Umbrella] Revisit YARN ProtocolBuffer int32 usages that need to upgrade to int64
> ---------------------------------------------------------------------------------
>
>                 Key: YARN-4843
>                 URL: https://issues.apache.org/jira/browse/YARN-4843
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: api
>            Reporter: Wangda Tan
>
> This JIRA is to track all int32 usages in YARN's ProtocolBuffer APIs that we possibly
need to update to int64.
> One example is resource API. We use int32 for memory now, if a cluster has 10k nodes,
each node has 210G memory, we will get a negative total cluster memory.
> We may have other fields may need to upgrade from int32 to int64. 



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

Mime
View raw message