Xuan Gong commented on YARN-1900:

[~ozawa] Do we really need this fix ? 
Probably, we could add some commandline/api to allow user to dynamically set 
the yarn.log.server.url on the run time even when the NM is running ? In that 
case, even when we change the log server, we could still get the logs without 
restart the NM. 
Thoughts ? 

> yarn.log.server.url needs to specify an http:// prefix
> ------------------------------------------------------
>                 Key: YARN-1900
>                 URL: https://issues.apache.org/jira/browse/YARN-1900
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: nodemanager
>    Affects Versions: 2.2.0, 2.3.0, 2.4.0
>            Reporter: Robert On
>            Assignee: Tsuyoshi Ozawa
>            Priority: Minor
>         Attachments: YARN-1900.1.patch
> If yarn.log.server.url in yarn-site.xml is specified with an http:// prefix 
> and log-aggregation is turned on, the meta-refresh redirect for non-mapreduce 
> jobs will append the url to the end of the current url rather than replace 
> the entire url with the yarn.log.server.url.  This results in a broken 
> redirect for FINISHED yarn jobs.  

This message was sent by Atlassian JIRA

Reply via email to