hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jean-Daniel Cryans (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-7531) [replication] NPE in SequenceFileLogReader because ReplicationSource doesn't nullify the reader
Date Thu, 10 Jan 2013 22:34:13 GMT

    [ https://issues.apache.org/jira/browse/HBASE-7531?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13550515#comment-13550515

Jean-Daniel Cryans commented on HBASE-7531:

I was able to find one test failure caused by this:


The replication thread dies so truncating can't complete.

bq. The cause is the dubious semantics of openReader imho

Yeah I should probably fold in that reader somehow into ReplicationHLogReaderManager.
> [replication] NPE in SequenceFileLogReader because ReplicationSource doesn't nullify
the reader
> -----------------------------------------------------------------------------------------------
>                 Key: HBASE-7531
>                 URL: https://issues.apache.org/jira/browse/HBASE-7531
>             Project: HBase
>          Issue Type: Bug
>            Reporter: Jean-Daniel Cryans
>            Assignee: Jean-Daniel Cryans
>         Attachments: HBASE-7531.patch
> Here's a NPE I get half the time I run TestReplication:
> {noformat}
> 2012-12-20 08:59:17,259 ERROR [RegionServer:1;,49168,1356011734418-EventThread.replicationSource,2]
regionserver.ReplicationSource$1(727): Unexpected exception in ReplicationSource, currentPath=hdfs://localhost:65533/user/jdcryans/hbase/.logs/,49168,1356011734418/
> java.lang.NullPointerException
>         at org.apache.hadoop.hbase.regionserver.wal.SequenceFileLogReader.seek(SequenceFileLogReader.java:261)
>         at org.apache.hadoop.hbase.replication.regionserver.ReplicationHLogReaderManager.seek(ReplicationHLogReaderManager.java:103)
>         at org.apache.hadoop.hbase.replication.regionserver.ReplicationSource.readAllEntriesToReplicateOrNextFile(ReplicationSource.java:414)
>         at org.apache.hadoop.hbase.replication.regionserver.ReplicationSource.run(ReplicationSource.java:332)
> {noformat}
> The issue happens after an IOE was caught while opening the reader, the issue is that
it isn't set to null after that then the rest of the code assumes the reader is usable.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

View raw message