ambari-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hudson (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (AMBARI-17834) HDFS copy thread in OutputHDFSFile is not removing files that are processed
Date Sun, 24 Jul 2016 03:59:20 GMT

    [ https://issues.apache.org/jira/browse/AMBARI-17834?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15390930#comment-15390930
] 

Hudson commented on AMBARI-17834:
---------------------------------

FAILURE: Integrated in Ambari-trunk-Commit #5375 (See [https://builds.apache.org/job/Ambari-trunk-Commit/5375/])
AMBARI-17834. Logfeeder: Removed file that is processed by (oleewere: [http://git-wip-us.apache.org/repos/asf?p=ambari.git&a=commit&h=842d382efc98bd1264b618cd42cad3c5f5b424c0])
* ambari-logsearch/ambari-logsearch-logfeeder/src/main/java/org/apache/ambari/logfeeder/output/OutputHDFSFile.java


> HDFS copy thread in OutputHDFSFile is not removing files that are processed 
> ----------------------------------------------------------------------------
>
>                 Key: AMBARI-17834
>                 URL: https://issues.apache.org/jira/browse/AMBARI-17834
>             Project: Ambari
>          Issue Type: Bug
>          Components: ambari-logsearch
>    Affects Versions: 2.4.0
>            Reporter: Hemanth Yamijala
>            Assignee: Hayat Behlim
>             Fix For: 2.4.0
>
>         Attachments: AMBARI-17834.patch
>
>
> Consider the HDFS copy thread here: https://github.com/apache/ambari/blob/trunk/ambari-logsearch/ambari-logsearch-logfeeder/src/main/java/org/apache/ambari/logfeeder/output/OutputHDFSFile.java#L177
> It uses a data structure localReadyFiles that has the list of files to copy to HDFS.
But AFAICS, nowhere are processed files removed from this data structure. This could result
in multiple incorrect behaviours:
> * Over a period of time, this structure could keep growing indefinitely. While the check
for existence of files will allow progress, it will take longer and longer to run.
> *  The check to wait when there are no files to process will never pass because the size
of the data structure will never be 0. This could lead to tight looped checks and some utilization
issues.



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

Mime
View raw message