[
https://issues.apache.org/jira/browse/HADOOP-10401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13939538#comment-13939538
]
Colin Patrick McCabe commented on HADOOP-10401:
-----------------------------------------------
Hi Akira,
Thanks for looking at this!
My concern with the first patch is that you are now spawning two external
processes instead of one. It seems like this could regress performance
somewhat. There can be a large overhead to spawning an external process from
the JVM due to the need to copy page tables during fork().
I think maybe we should consider executing a shell, which then executes both
"id -gn" and "id -Gn". Perhaps something like 'sh -c id -gn ; id -Gn'
> ShellBasedUnixGroupsMapping#getGroups does not always return primary group
> first
> --------------------------------------------------------------------------------
>
> Key: HADOOP-10401
> URL: https://issues.apache.org/jira/browse/HADOOP-10401
> Project: Hadoop Common
> Issue Type: Bug
> Components: util
> Affects Versions: 2.4.0
> Reporter: Colin Patrick McCabe
> Assignee: Akira AJISAKA
> Attachments: HADOOP-10401.patch
>
>
> {{ShellBasedUnixGroupsMapping#getGroups}} does not always return the primary
> group first. It should do this so that clients who expect it don't get the
> wrong result. We should also document that the primary group is returned
> first in the API. Note that {{JniBasedUnixGroupsMapping}} does return the
> primary group first.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
