[ 
https://issues.apache.org/jira/browse/HDFS-5856?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13887324#comment-13887324
 ] 

Suresh Srinivas edited comment on HDFS-5856 at 1/31/14 1:08 AM:
----------------------------------------------------------------

[~elserj], thanks for the patch.

Minor nits:
# Instead of repeating e.getMessage() many times, you could use String msg = 
e.getMessage(). Might make it is easy to format the code.
# One of the lines goes beyond 80 chars. It might be more readable with a line 
limited to a single exception.

Other than that it looks good. 


was (Author: sureshms):
[~elserj], thanks for the patch.

Minor nits:
# Instead of repeating e.getMessage() many times, you could use String msg = 
e.getMessage(). Might make it is easy to format the code.
# One of the lines goes beyond 80 chars. It might be more readable with a line 
limited to a single exception.

Other than it looks good. 

> DataNode.checkDiskError might throw NPE
> ---------------------------------------
>
>                 Key: HDFS-5856
>                 URL: https://issues.apache.org/jira/browse/HDFS-5856
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: datanode
>    Affects Versions: 2.2.0
>            Reporter: Josh Elser
>            Assignee: Josh Elser
>            Priority: Minor
>         Attachments: HDFS-5856.diff
>
>
> Running a small 2.2.0 cluster with a heavy workload. I noticed in the 
> DataNode log that when faced with network exceptions, I would sometimes get a 
> NullPointerException from checkDiskError.
> Looking at the code, if the Exception that is handed to checkDiskError 
> contains a null message, the checks that try to identify if the Exception is 
> network-related will throw a NPE.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)

Reply via email to