[
https://issues.apache.org/jira/browse/MAPREDUCE-3999?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13227938#comment-13227938
]
Robert Joseph Evans commented on MAPREDUCE-3999:
------------------------------------------------
A few comments:
* In the patch if the path is N/A and the status is SUCCEEDED the proxy will
still try to load "N/A".
* For some reason the proxy which is part of yarn, and not MR is returning an
error message that is MR specific.
Because we will always return a 404 error in this case, I would prefer to see
something like
{code}
if(applicationReport.getOriginalTrackingUrl().equals("N/A")) {
String message;
switch(applicationReport.getFinalApplicationStatus()) {
case FAILED:
case KILLED:
case SUCCEEDED:
message = "The requested application exited before setting a
tracking URL.";
break;
case UNDEFINED:
message = "The requested application does not appear to be running
yet, and has not set a tracking URL.";
break;
default:
//This should never happen, but just to be safe
message = "The requested application has not set a tracking URL.";
break;
}
noFound(resp, message);
return;
}
{code}
If you would prefer to throw an exception in the default case that would be
good too.
> Tracking link gives an error if the AppMaster hasn't started yet
> ----------------------------------------------------------------
>
> Key: MAPREDUCE-3999
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-3999
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv2, webapps
> Affects Versions: 0.23.1
> Reporter: Ravi Prakash
> Assignee: Ravi Prakash
> Attachments: MAPREDUCE-3999.patch, MAPREDUCE-3999.patch
>
>
> Courtesy [~sseth]
> {quote}
> "The MRAppMaster died before writing anything."
> Steps to generate the error:
> 1. Setup a queue with 1 max active application per user
> 2. Submit a long running job to this queue.
> 3. Submit another job to the queue as the same user. Access the tracking URL
> for job 2 directly or via Oozie (not via the RM link - which is rewritten once
> the app starts).
> This would exist in situations where the queue doesn't have enough capacity -
> or for the small period of time between app submission and AM start.
> {quote}
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
