Mailing-List: contact hdfs-issues-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: hdfs-issues@hadoop.apache.org
Date: Wed, 30 Jan 2013 04:59:14 +0000 (UTC)
From: "Suresh Srinivas (JIRA)" <jira@apache.org>
To: hdfs-issues@hadoop.apache.org
Message-ID: <JIRA.12629660.1359451296695.213481.1359521954971@arcas>
In-Reply-To: <JIRA.12629660.1359451296695@arcas>
References: <JIRA.12629660.1359451296695@arcas>
Subject: [jira] [Commented] (HDFS-4450) Duplicate data node on the name node
 after formatting data node
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable


    [ https://issues.apache.org/jira/browse/HDFS-4450?page=3Dcom.atlassian.=
jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D13566=
164#comment-13566164 ]=20

Suresh Srinivas commented on HDFS-4450:
---------------------------------------

bq. I view the other versions of the code will not have this problem, In ad=
dition practical use impossible to restart namenode when maintenance datano=
de

You are right about the behavior. I had forgotten the previous discussion r=
elated to this from the comment I had posted:
https://issues.apache.org/jira/browse/HDFS-3224?focusedCommentId=3D13472817=
&page=3Dcom.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#co=
mment-13472817

Looking at the description of this jira, the Host2NodesMap always has {{key=
 =3D ipAddress}} and {{value =3D DatanodeDescriptor}}. So add() method is d=
oing the correct thing. If you look at {{host2DatanodeMap.getDatanodeByXfer=
Addr()}}, it does look up the host by the ipAddress. It then compares the x=
ferAddress to that of the transfer address of the DatanodeDescriptor stored=
 as value.

Can you post the logs for this issue?

               =20
> Duplicate data node on the name node after formatting data node
> ---------------------------------------------------------------
>
>                 Key: HDFS-4450
>                 URL: https://issues.apache.org/jira/browse/HDFS-4450
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: namenode
>    Affects Versions: 2.0.2-alpha
>            Reporter: WenJin Ma
>   Original Estimate: 168h
>  Remaining Estimate: 168h
>
> Duplicate data node on the name node after formatting data node=E3=80=82
> When we registered data node=EF=BC=8Cuse nodeReg.getXferPort() to find Da=
tanodeDescriptor.
> {code}
>  DatanodeDescriptor nodeN =3D host2DatanodeMap.getDatanodeByXferAddr(
>         nodeReg.getIpAddr(), nodeReg.getXferPort());
> {code}
> but add data node use node.getIpAddr().
> {code}
>     /** add node to the map=20
>    * return true if the node is added; false otherwise.
>    */
>   boolean add(DatanodeDescriptor node) {
>     hostmapLock.writeLock().lock();
>     try {
>       if (node=3D=3Dnull || contains(node)) {
>         return false;
>       }
>      =20
>       String ipAddr =3D node.getIpAddr();
>       DatanodeDescriptor[] nodes =3D map.get(ipAddr);
>       DatanodeDescriptor[] newNodes;
>       if (nodes=3D=3Dnull) {
>         newNodes =3D new DatanodeDescriptor[1];
>         newNodes[0]=3Dnode;
>       } else { // rare case: more than one datanode on the host
>         newNodes =3D new DatanodeDescriptor[nodes.length+1];
>         System.arraycopy(nodes, 0, newNodes, 0, nodes.length);
>         newNodes[nodes.length] =3D node;
>       }
>       map.put(ipAddr, newNodes);
>       return true;
>     } finally {
>       hostmapLock.writeLock().unlock();
>     }
>   }
> {code}

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