[ 
https://issues.apache.org/jira/browse/HADOOP-15774?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16622622#comment-16622622
 ] 

Íñigo Goiri commented on HADOOP-15774:
--------------------------------------

[[email protected]], I've been reading the documentation and the code for the 
YARN registry and I think it covers the storage part (including the API for 
registering and discoverying) but it would need to be integrated with the 
components.

I think it would make sense to still modify the following components:
* Namenode: register itself in the registry at startup time.
* Datanode: use the YARN Registry API for discovery and setup the blockpool 
based on this.
* Resource Manager: register itself in the registry at startup time.
* Node Manager: use the YARN Registry API for discovery and tune the client to 
connect to the proper RMs.

I think it still makes sense to have an intermediate class to switch between 
the current XML Configuration approach and this new service.

A couple limitation I see in the current registry:
* Would it make sense to make the registry a standalone service instead of 
running in the RM?
* Being part of the YARN may bring some weird dependencies (HDFS->YARN), any 
ideas here? Should it be move to a separate place? Maybe parts?

We should also rename this effort to something like "Leveraging YARN register 
in Hadoop services".
Thoughts?
We can refocus the work in that direction if so.

> Discovery of HA servers
> -----------------------
>
>                 Key: HADOOP-15774
>                 URL: https://issues.apache.org/jira/browse/HADOOP-15774
>             Project: Hadoop Common
>          Issue Type: New Feature
>            Reporter: Íñigo Goiri
>            Priority: Major
>         Attachments: Discovery Service.pdf
>
>
> Currently, Hadoop relies on configuration files to specify the servers.
> This requires maintaining these configuration files and propagating the 
> changes.
> Hadoop should have a framework to provide discovery.
> For example, in HDFS, we could define the Namenodes in a shared location and 
> the DNs would use the framework to find the Namenodes.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to