[jira] Commented: (HADOOP-6986) SequenceFile.Reader should distinguish between Network IOE and Parsing IOE

["Nicolas Spiegelberg (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HADOOP-6986?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12917795#action_12917795 ] 

Nicolas Spiegelberg commented on HADOOP-6986:
---------------------------------------------

To fix this issue, I kept all the existing error types & messages, but I added ParseException as the cause to all parsing-related IOEs.  None of the changed exceptions had an associated cause prior.  This will allow us to maintain 100% backwards compatibility (in case any users were doing deep inspection of the IOE text) while allowing new users and easy way to check:  if(ioe.getCause() instanceof ParseException)

> SequenceFile.Reader should distinguish between Network IOE and Parsing IOE
> --------------------------------------------------------------------------
>
>                 Key: HADOOP-6986
>                 URL: https://issues.apache.org/jira/browse/HADOOP-6986
>             Project: Hadoop Common
>          Issue Type: Bug
>          Components: io
>    Affects Versions: 0.21.1, 0.22.0, 0.20-append
>            Reporter: Nicolas Spiegelberg
>            Priority: Minor
>             Fix For: 0.21.1, 0.22.0, 0.20-append
>
>         Attachments: HADOOP-6986_0.21.patch, HADOOP-6986_20-append.patch
>
>
> The SequenceFile.Reader api should give the user an easy way to distinguish between a Network/Low-level IOE and a Parsing IOE.  The use case appeared recently in the HBase project:
> Originally, if a RegionServer got an IOE from HDFS while opening a region file, it would abort the open and let the HMaster reassign the region.  The assumption being that this is a network failure that will likely disappear at a later time or different partition of the network.  However, if HBase gets parsing exceptions, we want to log the problem and continue opening the region anyways, because parsing is an idempotent problem and retries won't fix this issue.
> Although this problem was found in HBase, it seems to be a generic problem of being able to more easily identify idempotent vs transient errors.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.