[
https://issues.apache.org/jira/browse/HIVE-16352?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17263987#comment-17263987
]
Zoltan Haindrich commented on HIVE-16352:
-----------------------------------------
Is it really not avoidable to write things correctly
Are these incorrect files created when the writer is being shut down
incorrectly - or hive is possibly reading an incomplete file?
Although I think the best would be to have a writer which could give better
consistency guarantees - I'm not against this change: because it's small and is
off by default.
Any strong opinion against merging it?
> Ability to skip or repair out of sync blocks with HIVE at runtime
> -----------------------------------------------------------------
>
> Key: HIVE-16352
> URL: https://issues.apache.org/jira/browse/HIVE-16352
> Project: Hive
> Issue Type: New Feature
> Components: Avro, File Formats, Reader
> Affects Versions: 3.1.2
> Reporter: Navdeep Poonia
> Assignee: gabrywu
> Priority: Major
> Labels: pull-request-available
> Time Spent: 1h 50m
> Remaining Estimate: 0h
>
> When a file is corrupted it raises the error java.io.IOException: Invalid
> sync! with hive.
> Can we have some functionality to skip or repair such blocks at runtime to
> make avro more error resilient in case of data corruption.
> Error: java.io.IOException: java.io.IOException: java.io.IOException: While
> processing file
> s3n://<bucket>/navdeepp/warehouse/avro_test/354dc34474404f4bbc0d8013fc8e6e4b_000042.
> java.io.IOException: Invalid sync!
> at
> org.apache.hadoop.hive.io.HiveIOExceptionHandlerChain.handleRecordReaderNextException(HiveIOExceptionHandlerChain.java:121)
> at
> org.apache.hadoop.hive.io.HiveIOExceptionHandlerUtil.handleRecordReaderNextException(HiveIOExceptionHandlerUtil.java:77)
> at
> org.apache.hadoop.hive.shims.HadoopShimsSecure$CombineFileRecordReader.doNextWithExceptionHandler(HadoopShimsSecure.java:334)
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
