[ 
https://issues.apache.org/jira/browse/HADOOP-2404?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Devaraj Das updated HADOOP-2404:
--------------------------------

    Status: Open  (was: Patch Available)

There is one problem that HADOOP-2185 introduced (without knowledge I am sure) 
in the tasktracker's reinitialization (after the jobtracker declares the 
tasktracker as lost, and upon the next connection to the jobtracker, the 
tasktracker is asked to reinitialize). Post HADOOP-2185, the tasktracker's 
initialize method sets "mapred.task.tracker.report.bindAddress" to the listener 
address. This would be in the format _/<name>:<port>_ (due to the way 
InetAddress.toString() is implemented). Now if the tasktracker were lost and 
were to reinitialize, it would call the initialize method again, and this time 
the call to createSocketAddress would fail because of the leading slash in the 
argument, and the tasktracker would not be able to recover from this. So maybe, 
we should do a check for '/' in the tasktracker initialization before passing 
it to createSocketAddress? Not sure whether such problems exist elsewhere in 
the framework code too ....
Pre-HADOOP-2185, the "mapred.task.tracker.report.bindAddress" would not be 
reset to the listener address, and, hence the above problem is not there..

This issue needs a resolution (and hence I am cancelling the patch).

> HADOOP-2185 breaks compatibility with hadoop-0.15.0
> ---------------------------------------------------
>
>                 Key: HADOOP-2404
>                 URL: https://issues.apache.org/jira/browse/HADOOP-2404
>             Project: Hadoop
>          Issue Type: Bug
>          Components: conf
>    Affects Versions: 0.16.0
>            Reporter: Arun C Murthy
>            Assignee: Konstantin Shvachko
>            Priority: Blocker
>             Fix For: 0.16.0
>
>         Attachments: ConfigConvert.patch, ConfigurationConverter.patch
>
>
> HADOOP-2185 removed the following configuration parameters:
> {noformat}
> dfs.secondary.info.port
> dfs.datanode.port
> dfs.info.port
> mapred.job.tracker.info.port
> tasktracker.http.port
> {noformat}
> and changed the following configuration parameters:
> {noformat}
> dfs.secondary.info.bindAddress
> dfs.datanode.bindAddress
> dfs.info.bindAddress
> mapred.job.tracker.info.bindAddress
> mapred.task.tracker.report.bindAddress
> tasktracker.http.bindAddress
> {noformat}
> without a backward-compatibility story.
> Lots are applications/cluster-configurations are prone to fail hence, we need 
> a way to keep things working as-is for 0.16.0 and remove them for 0.17.0.

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

Reply via email to