hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Daniel Templeton (JIRA)" <j...@apache.org>
Subject [jira] [Created] (YARN-6609) AdminService should use RMFatalEvent instead of calling transitionToStandby() directly
Date Tue, 16 May 2017 19:15:04 GMT
Daniel Templeton created YARN-6609:
--------------------------------------

             Summary: AdminService should use RMFatalEvent instead of calling transitionToStandby()
directly
                 Key: YARN-6609
                 URL: https://issues.apache.org/jira/browse/YARN-6609
             Project: Hadoop YARN
          Issue Type: Improvement
          Components: resourcemanager
    Affects Versions: 2.9.0
            Reporter: Daniel Templeton
            Assignee: Daniel Templeton
            Priority: Minor


In YARN-3742, we put the RM in charge of transitioning itself to standby.  Instead of having
lots of different components driving the transition from lots of different places, they now
throw an {{RMFatalEvent}} at the RM, and the RM decides what the right thing to do is.

{{AdminService}} still calls {{transitionToStandby()}} directly, without first checking whether
HA is enabled.  To be consistent and safer, {{AdminService}} should probably also use an {{RMFatalEvent}}
to initiate the transition.

This change might have some cascading API impact, making some methods no longer useful.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

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


Mime
View raw message