Mailing-List: contact core-dev-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: core-dev@hadoop.apache.org
Message-ID: <1791686391.1213054304985.JavaMail.jira@brutus>
Date: Mon, 9 Jun 2008 16:31:44 -0700 (PDT)
From: "Konstantin Shvachko (JIRA)" <jira@apache.org>
To: core-dev@hadoop.apache.org
Subject: [jira] Updated: (HADOOP-3520) Generation stamp upgrade fails
 TestDFSUpgradeFromImage
In-Reply-To: <950139835.1213053225042.JavaMail.jira@brutus>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


     [ https://issues.apache.org/jira/browse/HADOOP-3520?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Konstantin Shvachko updated HADOOP-3520:
----------------------------------------

    Attachment: failure.log

Attaching the log of the failed test and a link to the Hudson build that generated it:
http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/2621/testReport/org.apache.hadoop.dfs/TestDFSUpgradeFromImage/testUpgradeFromImage/

It's also worth noting that there is another NPE in the log, which may the cause of the previous one:
{code}
2008-06-07 02:37:00,428 ERROR dfs.DataNode (UpgradeObjectDatanode.java:run(97)) - java.lang.NullPointerException
	at org.apache.hadoop.dfs.GenerationStampUpgradeDatanode.doUpgrade(GenerationStampUpgrade.java:291)
	at org.apache.hadoop.dfs.UpgradeObjectDatanode.run(UpgradeObjectDatanode.java:95)
	at java.lang.Thread.run(Thread.java:595)
{code}

The NPE in UpgradeManagerDatanode.initializeUpgrade() should not happen, because with currentUpgrades == null 
the method should exit after super.initializeUpgrade().
I think this is a synchronization issue, because UpgradeManagerDatanode.initializeUpgrade() is not synchronized.
If this can happen only when using MiniDFSCluster then we shoulddowngrade this issue and fix it later.
Marking it for 0.18 for now.

> Generation stamp upgrade fails TestDFSUpgradeFromImage
> ------------------------------------------------------
>
>                 Key: HADOOP-3520
>                 URL: https://issues.apache.org/jira/browse/HADOOP-3520
>             Project: Hadoop Core
>          Issue Type: Bug
>          Components: dfs
>    Affects Versions: 0.18.0
>            Reporter: Konstantin Shvachko
>             Fix For: 0.18.0
>
>         Attachments: failure.log
>
>
> TestDFSUpgradeFromImage occasionally fails doing distributed upgrade related to generation stamps with the following NPE:
> {code}
> java.lang.NullPointerException
> 	at org.apache.hadoop.dfs.UpgradeManagerDatanode.initializeUpgrade(UpgradeManagerDatanode.java:54)
> 	at org.apache.hadoop.dfs.DataStorage.verifyDistributedUpgradeProgress(DataStorage.java:402)
> 	at org.apache.hadoop.dfs.DataStorage.doTransition(DataStorage.java:232)
> 	at org.apache.hadoop.dfs.DataStorage.recoverTransitionRead(DataStorage.java:139)
> 	at org.apache.hadoop.dfs.DataNode.startDataNode(DataNode.java:268)
> 	at org.apache.hadoop.dfs.DataNode.<init>(DataNode.java:185)
> 	at org.apache.hadoop.dfs.DataNode.makeInstance(DataNode.java:2952)
> 	at org.apache.hadoop.dfs.DataNode.instantiateDataNode(DataNode.java:2907)
> 	at org.apache.hadoop.dfs.MiniDFSCluster.startDataNodes(MiniDFSCluster.java:401)
> 	at org.apache.hadoop.dfs.MiniDFSCluster.<init>(MiniDFSCluster.java:267)
> 	at org.apache.hadoop.dfs.MiniDFSCluster.<init>(MiniDFSCluster.java:162)
> 	at org.apache.hadoop.dfs.TestDFSUpgradeFromImage.testUpgradeFromImage(TestDFSUpgradeFromImage.java:180)
> {code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.