hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Arpit Agarwal (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-1994) Expose YARN/MR endpoints on multiple interfaces
Date Wed, 16 Jul 2014 03:26:04 GMT

    [ https://issues.apache.org/jira/browse/YARN-1994?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14063064#comment-14063064
] 

Arpit Agarwal commented on YARN-1994:
-------------------------------------

[~mipoto], I was unable to apply your v4 patch to trunk. Could you please rebase it to trunk
and fix the whitespace errors?. I’ll review it once it is fixed.

[~cwelch], Thank you for picking up this Jira. I reviewed your v3 patch and it looks great,
just a few comments below.

*AdminService.java:*
- Now that you have the code to construct the correct connectAddress (lines 180-192), {{RPCUtil.updateConnectAddr}}
looks a little inaccurate. It undoes the work of constructing the connectAddress since it
always gets the hostname from the property. I think instead of the call to {{RPCUtil.updateConnectAddr}},
we can just have _conf.updateConnectAddr(YarnConfiguration.RM_ADMIN_ADDRESS, connectAddress)_.
It is probably not a big deal since any real-world configuration will define this property.

*ApplicationMasterService.java, ClientRMService.java, HistoryClientService.java, ResourceLocalizationService.java,
ResourceTrackerService.java:*
- Do we need a similar logic to compute connectAddress in {{serviceStart}}?

*WebAppUtils.java:*
- {{getNMWebAppBindURLWithoutScheme}} can be simplified by using {{RPCUtil.getAddressAsString}}.

*YarnConfiguration.java:*
- {{TIMELINE_SERVICE_BIND_HOST}} is unused. Fix the timeline service or just drop this definition?

I don't understand YARN well enough to be sure that RM webapp endpoints are being correctly
handled, [~vinodkv] could you please review that?

Thanks, Arpit.


> Expose YARN/MR endpoints on multiple interfaces
> -----------------------------------------------
>
>                 Key: YARN-1994
>                 URL: https://issues.apache.org/jira/browse/YARN-1994
>             Project: Hadoop YARN
>          Issue Type: Improvement
>          Components: nodemanager, resourcemanager, webapp
>    Affects Versions: 2.4.0
>            Reporter: Arpit Agarwal
>            Assignee: Craig Welch
>         Attachments: YARN-1994.0.patch, YARN-1994.1.patch, YARN-1994.2.patch, YARN-1994.3.patch,
YARN-1994.4.patch
>
>
> YARN and MapReduce daemons currently do not support specifying a wildcard address for
the server endpoints. This prevents the endpoints from being accessible from all interfaces
on a multihomed machine.
> Note that if we do specify INADDR_ANY for any of the options, it will break clients as
they will attempt to connect to 0.0.0.0. We need a solution that allows specifying a hostname
or IP-address for clients while requesting wildcard bind for the servers.
> (List of endpoints is in a comment below)



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message