[
https://issues.apache.org/jira/browse/YARN-1666?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13910849#comment-13910849
]
Hitesh Shah commented on YARN-1666:
-----------------------------------
{code}
- if (!(this.configurationProvider instanceof LocalConfigurationProvider)) {
- // load yarn-site.xml
- this.conf =
- this.configurationProvider.getConfiguration(this.conf,
- YarnConfiguration.YARN_SITE_XML_FILE);
- // load core-site.xml
- this.conf =
- this.configurationProvider.getConfiguration(this.conf,
- YarnConfiguration.CORE_SITE_CONFIGURATION_FILE);
- // Do refreshUserToGroupsMappings with loaded core-site.xml
- Groups.getUserToGroupsMappingServiceWithLoadedConfiguration(this.conf)
- .refresh();
- }
+
+ // load yarn-site.xml
+ this.conf.addResource(this.configurationProvider
+ .getConfigurationInputStream(this.conf,
+ YarnConfiguration.YARN_SITE_CONFIGURATION_FILE));
+ // load core-site.xml
+ this.conf.addResource(this.configurationProvider
+ .getConfigurationInputStream(this.conf,
+ YarnConfiguration.CORE_SITE_CONFIGURATION_FILE));
+ // Do refreshUserToGroupsMappings with loaded core-site.xml
+ Groups.getUserToGroupsMappingServiceWithLoadedConfiguration(this.conf)
+ .refresh();
{code}
The above code seems to be breaking MiniClusters. Is the expectation now that
anyone using a MiniCluster has to create the appropriate config files and add
them into the unit test class path?
Stack trace below:
{code}
Exception: null
java.lang.NullPointerException
at
org.apache.hadoop.conf.Configuration$Resource.<init>(Configuration.java:182)
at org.apache.hadoop.conf.Configuration.addResource(Configuration.java:751)
at
org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.serviceInit(ResourceManager.java:193)
at org.apache.hadoop.service.AbstractService.init(AbstractService.java:163)
at
org.apache.hadoop.yarn.server.MiniYARNCluster.initResourceManager(MiniYARNCluster.java:268)
at
org.apache.hadoop.yarn.server.MiniYARNCluster.access$400(MiniYARNCluster.java:90)
at
org.apache.hadoop.yarn.server.MiniYARNCluster$ResourceManagerWrapper.serviceInit(MiniYARNCluster.java:419)
{code}
> Make admin refreshNodes work across RM failover
> -----------------------------------------------
>
> Key: YARN-1666
> URL: https://issues.apache.org/jira/browse/YARN-1666
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Xuan Gong
> Assignee: Xuan Gong
> Fix For: 2.4.0
>
> Attachments: YARN-1666.1.patch, YARN-1666.2.patch, YARN-1666.2.patch,
> YARN-1666.3.patch, YARN-1666.4.patch, YARN-1666.4.patch, YARN-1666.5.patch,
> YARN-1666.6.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
