[
https://issues.apache.org/jira/browse/HBASE-4269?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13230972#comment-13230972
]
Jan Lukavsky commented on HBASE-4269:
-------------------------------------
Hi,
I think patch to this issue changed semantics for mapreduce API. In HBASE-4196
there was no change in semantics in
org.apache.hadoop.hbase.mapreduce.TableRecordReaderImpl, the only change was in
org.apache.hadoop.hbase.mapred.TableRecordReaderImpl (where the catch of
UnknownScannerException was changed to IOException). Now the semantics of
mapreduce API is different of the one before HBASE-4196, and I think this
should be reverted. Is there any reason why to have different semantics for the
two APIs? Wouldn't it be better to accept the change of semantics in
HBASE-4196? Are there any negative side-effects of this change? I don't see any
discussion of the type "do we need to change the semantics back"?
Thanks for reply :)
Jan
> Add tests and restore semantics to TableInputFormat/TableRecordReader
> ---------------------------------------------------------------------
>
> Key: HBASE-4269
> URL: https://issues.apache.org/jira/browse/HBASE-4269
> Project: HBase
> Issue Type: Improvement
> Components: mapred, mapreduce, test
> Affects Versions: 0.90.5, 0.92.0
> Reporter: Jonathan Hsieh
> Assignee: Jonathan Hsieh
> Fix For: 0.90.5
>
> Attachments:
> 0001-HBASE-4269-Add-tests-and-restore-semantics-to-TableI.patch,
> 0001-HBASE-4269-Add-tests-and-restore-semantics-to-TableI.patch
>
>
> HBASE-4196 Modified the semantics of failures in
> TableImportFormat/TableRecordReader, and had no tests cases. This patch
> restores semantics to rethrow when a DoNotRetryIOException is triggered and
> adds test cases.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
