[
https://issues.apache.org/jira/browse/TWILL-119?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15930901#comment-15930901
]
ASF GitHub Bot commented on TWILL-119:
--------------------------------------
Github user anew commented on a diff in the pull request:
https://github.com/apache/twill/pull/38#discussion_r106763105
--- Diff:
twill-yarn/src/main/hadoop20/org/apache/twill/internal/yarn/Hadoop20YarnAppClient.java
---
@@ -211,7 +218,6 @@ public YarnApplicationReport getReport() {
try {
return new
Hadoop20YarnApplicationReport(yarnClient.getApplicationReport(appId));
} catch (YarnRemoteException e) {
- LOG.error("Failed to get application report {}", appId, e);
--- End diff --
removing the logging here is good. However, the messages could add useful
context (here: the appId). If you wrap it into a RuntimeException, why not add
the message?
> Updates to work on a secure cluster with HA ResourceManager
> -----------------------------------------------------------
>
> Key: TWILL-119
> URL: https://issues.apache.org/jira/browse/TWILL-119
> Project: Apache Twill
> Issue Type: Improvement
> Components: yarn
> Reporter: Albert Shau
> Assignee: Terence Yim
> Fix For: 0.11.0
>
>
> I have not tried it yet, but it seems like we may need some extra work for
> Twill to run on a secure cluster that uses HA ResourceManager. The YarnUtils
> class gets the socket addr for the RM using YarnConfiguration.RM_ADDRESS, but
> in HA mode the conf contains YarnConfiguration.RM_HA_IDS, with each id having
> a corresponding hostname setting.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
