hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "stack (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-12981) FSHLog: UNEXPECTED java.lang.ArrayIndexOutOfBoundsException: -4
Date Fri, 06 Feb 2015 23:01:37 GMT

    [ https://issues.apache.org/jira/browse/HBASE-12981?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14310116#comment-14310116
] 

stack commented on HBASE-12981:
-------------------------------

Thanks [~enis] I think you are right.  I was looking at wrong 0.98 version.  This is a patched
0.98.1 (CDH5.1.3).  It doesn't have HBASE-11200.  Let me close this as a duplicate (Thanks
for reviews [~lhofhansl] and [~busbey] -- agree w/ Sean).

> FSHLog: UNEXPECTED java.lang.ArrayIndexOutOfBoundsException: -4
> ---------------------------------------------------------------
>
>                 Key: HBASE-12981
>                 URL: https://issues.apache.org/jira/browse/HBASE-12981
>             Project: HBase
>          Issue Type: Bug
>          Components: wal
>    Affects Versions: 0.98.10
>            Reporter: stack
>            Assignee: stack
>             Fix For: 0.98.11
>
>         Attachments: 12981.0.98.txt
>
>
> A user reported the below. It happens after the RS has been running a while.
> 015-01-20 22:33:23,031 ERROR org.apache.hadoop.hbase.regionserver.wal.FSHLog: UNEXPECTED
> java.lang.ArrayIndexOutOfBoundsException: -4
> at org.apache.hadoop.hbase.regionserver.wal.FSHLog$AsyncWriter.run(FSHLog.java:1149)
> at java.lang.Thread.run(Thread.java:745)
> 2015-01-20 22:33:23,035 INFO org.apache.hadoop.hbase.regionserver.wal.FSHLog: regionserver60020-WAL.AsyncWriter
exiting
> ##################
> Similarly on Node 23 - on 12-20-2014 05:13:
> 2014-12-20 05:13:40,715 ERROR org.apache.hadoop.hbase.regionserver.wal.FSHLog: UNEXPECTED
> java.lang.ArrayIndexOutOfBoundsException: -3
> at org.apache.hadoop.hbase.regionserver.wal.FSHLog$AsyncWriter.run(FSHLog.java:1149)
> at java.lang.Thread.run(Thread.java:745)
> ###################
> Looking in code, I can't see how this could come about other than our write seqid ran
over the top of a long (unlikely). I think this a 0.98 issue since 1.0+ is different here.
It does:
> int index = Math.abs(this.syncRunnerIndex++) % this.syncRunners.length;
> I'm going to add logging of the circumstance that produces a negative index and then
defense against our using negative indices; there could be more going on in here, more than
I can see.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message