[
https://issues.apache.org/jira/browse/HADOOP-15774?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16638706#comment-16638706
]
Íñigo Goiri commented on HADOOP-15774:
--------------------------------------
I've been checking and it looks to me like the best solution is to move from
"Hadoop YARN Registry" to "Hadoop Registry".
The code in the package itself has very few dependencies to YARN
(YARNConfiguration and a couple exception handlers).
The only issues I see is what [~eyang] mentioned about the dynamically
ZooKeeper deployments.
However, for the cases of the RM and the NN, I think is fairly reasonable to
assume a well-known ZK addresses for HA.
[~eyang], from your comments I'm not sure if this assumption is fine or not;
can you clarify?
If this is fine, I can start a JIRA to move from "Hadoop YARN Registry" to
"Hadoop Registry", and use this.
Meanwhile, I'll update the design doc to make everything related to the new
Hadoop Registry.
BTW, it looks like there is an FS implementation of the Hadoop YARN Registry
already: FSRegistryOperationsService.
Anyway, eventually, I would also like to discover ZooKeeper through this
service somehow.
> 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]
