You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hdfs-issues@hadoop.apache.org by "Aaron T. Myers (JIRA)" <ji...@apache.org> on 2011/07/25 23:10:09 UTC
[jira] [Created] (HDFS-2193) The NN should determine its own
hostname using the same method as the DN
The NN should determine its own hostname using the same method as the DN
------------------------------------------------------------------------
Key: HDFS-2193
URL: https://issues.apache.org/jira/browse/HDFS-2193
Project: Hadoop HDFS
Issue Type: Bug
Components: name-node, security
Affects Versions: 0.23.0
Reporter: Aaron T. Myers
Assignee: Aaron T. Myers
When the "{{_HOST}}" replacement feature of Kerberos principal names is used, the various Hadoop daemons must determine their own canonical hostname in order to form their principal name.
Presently, the NN determines its own hostname based on the configured value of "{{fs.default.name}}", whereas the DNs determine their hostnames based on the result of "{{InetAddress.getLocalHost().getCanonicalHostName()}}". This asymmetry can cause confusion during configuration of a secure cluster.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (HDFS-2193) The NN should determine its own
hostname using the same method as the DN
Posted by "Aaron T. Myers (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-2193?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13070761#comment-13070761 ]
Aaron T. Myers commented on HDFS-2193:
--------------------------------------
I'm of the opinion that the NN's hostname determination should be changed to be the same as the way the DN presently does it, namely the result of "{{InetAddress.getLocalHost().getCanonicalHostName()}}".
It may be that there's some good reason for this asymmetry that I can't think of. (Distinct internal/external interfaces on the NN?) If so, the resolution of this JIRA should probably just be to clearly document how all the various daemons determine their own hostnames.
This same asymmetry exists between the JT and TT as well, but I figure we should just discuss the appropriate solution to this issue here first, and then I'll file a complementary MR JIRA once we figure out the correct course of action.
> The NN should determine its own hostname using the same method as the DN
> ------------------------------------------------------------------------
>
> Key: HDFS-2193
> URL: https://issues.apache.org/jira/browse/HDFS-2193
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: name-node, security
> Affects Versions: 0.23.0
> Reporter: Aaron T. Myers
> Assignee: Aaron T. Myers
>
> When the "{{_HOST}}" replacement feature of Kerberos principal names is used, the various Hadoop daemons must determine their own canonical hostname in order to form their principal name.
> Presently, the NN determines its own hostname based on the configured value of "{{fs.default.name}}", whereas the DNs determine their hostnames based on the result of "{{InetAddress.getLocalHost().getCanonicalHostName()}}". This asymmetry can cause confusion during configuration of a secure cluster.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira