[jira] Commented: (HADOOP-430) http server should die if the Datanode fails.

["Wendy Chien (JIRA)" <ji...@apache.org>]

    [ http://issues.apache.org/jira/browse/HADOOP-430?page=comments#action_12454224 ] 
            
Wendy Chien commented on HADOOP-430:
------------------------------------

I should add that we intentionally did not switch the order of first starting up the http server and then registering the datanode.  

It also looks like the http server is stopped when a datanode fails (in general), so no changes were added for that.  

> http server should die if the Datanode fails.
> ---------------------------------------------
>
>                 Key: HADOOP-430
>                 URL: http://issues.apache.org/jira/browse/HADOOP-430
>             Project: Hadoop
>          Issue Type: Bug
>          Components: dfs
>    Affects Versions: 0.5.0
>            Reporter: Konstantin Shvachko
>         Assigned To: Wendy Chien
>         Attachments: hadoop-430.patch
>
>
> HADOOP-375 changed the order in which a data node registers and starts its http server.
> Current version first starts the http server, then registers. As a result if the registration fails,
> which can happen e.g. if the datanode has illegal storage id (=does not belong to the cluster),
> the http server keeps running, which makes the data node seem running too.
> I think 2 things need to be done.
> 1) Datanode registration should precede the http server startup.
> 2) The http server should shutdown when the datanode fails (in general).
> I am attaching a patch for the first item.
> Does anybody know how the latter can be done in the version of jetty we are on now?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira