[
https://issues.apache.org/jira/browse/HDFS-14035?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16667660#comment-16667660
]
Chen Liang commented on HDFS-14035:
-----------------------------------
Two options were considered:
# make HAServiceProtocol to leverage delegation token.
# add the getServiceStatus into ClientNameNodeProtocol, and change to use
ClientNameNodeProtocol#getServiceStatus instead
The issue with 1 is that HAServiceProtocol is in Hadoop-common but delegation
token is in HDFS, there may be some way to make this work but I'm almost
certain it won't be clean. More importantly, associating delegation token to
HAServiceProtocol seems violating the abstraction we have. Also, I was under
impression that HAServiceProtocol is mainly for human admins to leverage, not
by YARN worker nodes.
The issue with 2 though, is that we will have getServiceStatus method in both
HAServiceProtocol and ClientNameNodeProtocol that does exactly the same thing.
NN side will have two methods with probably exactly same code, only
implementing two different protocols..which is not clean either.
[~shv], [~csun], [~xkrogen], [~zero45] any comments?
> NN status discovery does not leverage delegation token
> ------------------------------------------------------
>
> Key: HDFS-14035
> URL: https://issues.apache.org/jira/browse/HDFS-14035
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Chen Liang
> Assignee: Chen Liang
> Priority: Major
>
> Currently ObserverReadProxyProvider uses
> {{HAServiceProtocol#getServiceStatus}} to get the status of each NN.
> {{HAServiceProtocol}} does not leverage delegation token. So when YARN node
> manager makes this call, token authentication will fail, causing the
> application to fail.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
