[
https://issues.apache.org/jira/browse/HBASE-3386?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jean-Daniel Cryans updated HBASE-3386:
--------------------------------------
Attachment: HBASE-3386.patch
We can't throw an IOE when we fail to set the HTable, so instead I do it in
TableInputFormatBase and try to give a nice message.
> NPE in TableRecordReaderImpl.restart
> ------------------------------------
>
> Key: HBASE-3386
> URL: https://issues.apache.org/jira/browse/HBASE-3386
> Project: HBase
> Issue Type: Bug
> Reporter: Jean-Daniel Cryans
> Assignee: Jean-Daniel Cryans
> Priority: Minor
> Fix For: 0.90.0
>
> Attachments: HBASE-3386.patch
>
>
> I was running some testing on a already busy cluster and my clients in the
> mappers sometimes aren't able to create an HTable because of
> ConnectionLoseException. The code logs that as an error but still continues
> with a null HTable then this happens:
> {noformat}
> java.lang.NullPointerException
> at
> org.apache.hadoop.hbase.mapreduce.TableRecordReaderImpl.restart(TableRecordReaderImpl.java:58)
> at
> org.apache.hadoop.hbase.mapreduce.TableRecordReaderImpl.init(TableRecordReaderImpl.java:67)
> at
> org.apache.hadoop.hbase.mapreduce.TableRecordReader.init(TableRecordReader.java:57)
> at
> org.apache.hadoop.hbase.mapreduce.TableInputFormatBase.createRecordReader(TableInputFormatBase.java:108)
> {noformat}
> Needs to be prettier.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
