[jira] [Created] (SPARK-10187) Sometimes Web UI reports Application history not found

["Rares Vernica (JIRA)" <ji...@apache.org>]

Rares Vernica created SPARK-10187:
-------------------------------------

             Summary: Sometimes Web UI reports Application history not found
                 Key: SPARK-10187
                 URL: https://issues.apache.org/jira/browse/SPARK-10187
             Project: Spark
          Issue Type: Bug
          Components: Web UI
    Affects Versions: 1.4.1
         Environment: CentOS Linux release 7.1.1503 (Core)
Linux 3.10.0-229.4.2.el7.x86_64 #1 SMP Wed May 13 10:06:09 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux
            Reporter: Rares Vernica
            Priority: Minor


On the Web UI home page {{http://localhost:8080/}} in the list of Completed Applications clicking on the Name of a job shows the history of that job. This works for most of the time. For some jobs, Spark returns:

{quote}
Application history not found (app-20150824104953-0018)
No event logs found for application bid-00001-ph-1 in file:///data/disk1/spark/events. Did you specify the correct logging directory?
{quote}

All these jobs have the FINISHED state. {{spark-defaults.conf}} contains

{quote}
spark.eventLog.dir               file:///data/disk1/spark/events
spark.history.fs.logDirectory    file:///data/disk1/spark/events
{quote}

{{/data/disk1/spark/events}} is a normal directory on the disk. This works fine for some jobs but not others (even jobs on the same Spark session).

The Spark History UI shows correctly the history of the jobs that the Web UI cannot find the logs for. So, I believe the log directory is fine.

Here is the HTML snippet from the Web UI home page, under the Completed Applications section. Notice how the UI finds the history for some jobs, but not others:

{quote}
...
       <a href="app?appId=app-20150824110804-0019">app-20150824110804-0019</a>
       ...
        <a href="/history/app-20150824110804-0019">bid-00001-ph-2</a>
...
        <a href="app?appId=app-20150824104953-0018">app-20150824104953-0018</a>
        ...
        <a href="/history/not-found?msg=No+event+logs+found+for+application+bid-00001-ph-1+in+file%3A%2F%2F%2Fdata%2Fdisk1%2Fspark%2Fevents.+Did+you+specify+the+correct+logging+directory%3F&amp;title=Application history not found (app-20150824104953-0018)">bid-00001-ph-1</a>
...
        <a href="app?appId=app-20150824104907-0017">app-20150824104907-0017</a>
        ...
        <a href="/history/not-found?msg=No+event+logs+found+for+application+bid-00001-ph-2+in+file%3A%2F%2F%2Fdata%2Fdisk1%2Fspark%2Fevents.+Did+you+specify+the+correct+logging+directory%3F&amp;title=Application history not found (app-20150824104907-0017)">bid-00001-ph-2</a>
...
        <a href="app?appId=app-20150824103131-0016">app-20150824103131-0016</a>
        ...
        <a href="/history/app-20150824103131-0016">bid-00001-ph-1</a>
{quote}

These are sequential jobs and all the necessary files and directories in {{/data/disk1/spark/events}} and present and accessible and Spark History UI shows the history for all these jobs.

I looked around JIRA for similar issues, I think this is related to SPARK-6107 and SPARK-6950 but not a duplicate. The jobs for which these manifests are all FINISHED for some time.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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