[jira] [Commented] (HDFS-12358) Catch IOException when transferring edit log to Journal current dir through JN sync

["Arpit Agarwal (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HDFS-12358?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16142373#comment-16142373 ] 

Arpit Agarwal commented on HDFS-12358:
--------------------------------------

Thanks for this improvement [~hanishakoneru]. Minor comments:
# Initialize moveSuccess to false at declaration:
{code}
+    boolean moveSuccess;
{code}
# You can delete this line from the catch block after fixing 1.
{code}
+      moveSuccess = false;
{code}
# The log statement can be at INFO level so we see the failed move in the log file by default.
{code}
+      LOG.debug("Could not move %s to current directory.", tmpEditsFile);
{code}

+1 otherwise.

> Catch IOException when transferring edit log to Journal current dir through JN sync
> -----------------------------------------------------------------------------------
>
>                 Key: HDFS-12358
>                 URL: https://issues.apache.org/jira/browse/HDFS-12358
>             Project: Hadoop HDFS
>          Issue Type: Bug
>            Reporter: Hanisha Koneru
>            Assignee: Hanisha Koneru
>         Attachments: HDFS-12358.001.patch
>
>
> During JN sync, a missing edit log is first downloaded from a remote JN to a tmp dir and then moved to the current directory (protected by the Journal's synchronization). 
> Irrespective of whether the move succeeds or fails, we should delete the tmp file.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

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