hadoop-hdfs-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chuan Liu (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (HDFS-5227) Namenode.getAddress() should return namenode rpc-address
Date Sun, 22 Sep 2013 21:33:05 GMT

     [ https://issues.apache.org/jira/browse/HDFS-5227?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel

Chuan Liu resolved HDFS-5227.

    Resolution: Not A Problem

I noticed we have the following code in {{Namenode.initializeGenericKeys()}} method. This
method is executed before {{Namenode.initialize()}} method in the namenode constructor. Our
problem should be fixed with the following code. I think we run into failure previously because
"dfs.namenode.rpc-address" is not properly configured.
    // If the RPC address is set use it to (re-)configure the default FS
    if (conf.get(DFS_NAMENODE_RPC_ADDRESS_KEY) != null) {
      URI defaultUri = URI.create(HdfsConstants.HDFS_URI_SCHEME + "://"
          + conf.get(DFS_NAMENODE_RPC_ADDRESS_KEY));
      conf.set(FS_DEFAULT_NAME_KEY, defaultUri.toString());
      LOG.debug("Setting " + FS_DEFAULT_NAME_KEY + " to " + defaultUri.toString());
As for the unit test failure above, the federated namenode addresses of MiniDFSCluster will
be wrong in TestDeleteBlockPool if the {{getAddress()}} is modified according to the patch.

Resolve this one as "not a problem" because the behavior is the same as desired, i.e. getAddress()
will return "dfs.namenode.rpc-address" if namenode is initialized and "dfs.namenode.rpc-address"
is configured. 
> Namenode.getAddress() should return namenode rpc-address
> --------------------------------------------------------
>                 Key: HDFS-5227
>                 URL: https://issues.apache.org/jira/browse/HDFS-5227
>             Project: Hadoop HDFS
>          Issue Type: Bug
>    Affects Versions: 3.0.0, 2.1.1-beta
>            Reporter: Chuan Liu
>            Assignee: Chuan Liu
>         Attachments: HDFS-5227-trunk.patch
> Currently, {{Namenode.getAddress(Configuration conf)}} will return default file system
address as its result. The correct behavior should be returning config value of "dfs.namenode.rpc-address"
if it presents in the configurations. Otherwise namenode will fail to start if the default
file system is configured to another file system other than the one running in the cluster.
We have a similar issue in 1.0 code base. The JIRA is HDFS-4320. The previous JIRA is closed
and we cannot open it. Thus create a new one to track the issue in trunk.

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