Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Date: Fri, 3 Oct 2014 19:21:34 +0000 (UTC)
From: "Enis Soztutar (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Message-ID: <JIRA.12731001.1406787774000.185800.1412364094152@Atlassian.JIRA>
In-Reply-To: <JIRA.12731001.1406787774000@Atlassian.JIRA>
References: <JIRA.12731001.1406787774000@Atlassian.JIRA>
 <JIRA.12731001.1406787774912@arcas>
Subject: [jira] [Commented] (HBASE-11625) Reading datablock throws "Invalid
 HFile block magic" and can not switch to hdfs checksum
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HBASE-11625?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14158395#comment-14158395 ] 

Enis Soztutar commented on HBASE-11625:
---------------------------------------

Nick pointed me out to this issue. I have been trying to nail down a test failure on windows (TestHFileBlock#testConcurrentReading) which fails with the same stack trace. I can repro the failure only on windows, but with jdk6, jdk7u45 and jdk7u67 as well, and hadoop versions 2.2.0, 2.4.0, 2.5.0 and 2.6.0-SNAPSHOT. 

The test does write a file containing random HFileBlocks and does concurrent reads from multiple threads. Once in a while, whatever we seek() + read() does not match what is there in the file (I've verified multiple times from offsets to the actual file). I think there is a rare edge case that we are hitting. 

The other interesting bit is that the test only starts failing after 0.98.3. I was not able to get previous versions to fail. But also I was not able to bisect the commit because reproducing the failure is not easy. 

> Reading datablock throws "Invalid HFile block magic" and can not switch to hdfs checksum 
> -----------------------------------------------------------------------------------------
>
>                 Key: HBASE-11625
>                 URL: https://issues.apache.org/jira/browse/HBASE-11625
>             Project: HBase
>          Issue Type: Bug
>          Components: HFile
>    Affects Versions: 0.94.21, 0.98.4, 0.98.5
>            Reporter: qian wang
>         Attachments: 2711de1fdf73419d9f8afc6a8b86ce64.gz
>
>
> when using hbase checksum,call readBlockDataInternal() in hfileblock.java, it could happen file corruption but it only can switch to hdfs checksum inputstream till validateBlockChecksum(). If the datablock's header corrupted when b = new HFileBlock(),it throws the exception "Invalid HFile block magic" and the rpc call fail


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