[
https://issues.apache.org/jira/browse/MAPREDUCE-6090?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14134710#comment-14134710
]
Robert Kanter commented on MAPREDUCE-6090:
------------------------------------------
By default, only core-default and core-site are loaded unless you end up
calling something else that loads other site/default files. In this call, we
try to get the address from the Configuration before something else
(UserGroupInformation) causes mapred-default/site to be loaded, so we get the
default address, which is 0.0.0.0:10033. In the case where we don't pass any
user argument, it determines the current user by calling a method in
UserGroupInformation, which causes mapred-default/site to be loaded; and this
happens before we try to get the address -- so it works in this case.
> mapred hsadmin getGroups fails to connect in some cases
> -------------------------------------------------------
>
> Key: MAPREDUCE-6090
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-6090
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: client
> Affects Versions: 2.5.1
> Reporter: Robert Kanter
> Assignee: Robert Kanter
>
> If you do {{mapred hsadmin -getGroups}} it works fine (assuming
> {{mapreduce.jobhistory.admin.address}} is set properly in mapred-site.xml).
> But if you do {{mapred hsadmin -getGroups foo_user}}, it will keep retrying
> to connect to localhost:
> {noformat}
> INFO ipc.Client: Retrying connect to server: 0.0.0.0/0.0.0.0:10033. Already
> tried 0 time(s); retry policy is
> RetryUpToMaximumCountWithFixedSleep(maxRetries=10, sleepTime=1000
> MILLISECONDS)
> ...
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
