[
https://issues.apache.org/jira/browse/YARN-3629?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14539495#comment-14539495
]
nijel commented on YARN-3629:
-----------------------------
Moving the logs message is bit tricky since it logs some parameters which is
not available in serviceStart. So keeping this log as it is
Adding a new log message to print the nodeid for information purpose
Any different thought ?
> NodeID is always printed as "null" in node manager initialization log.
> ----------------------------------------------------------------------
>
> Key: YARN-3629
> URL: https://issues.apache.org/jira/browse/YARN-3629
> Project: Hadoop YARN
> Issue Type: Bug
> Reporter: nijel
> Assignee: nijel
> Attachments: YARN-3629-1.patch
>
>
> In Node manager log during startup the following logs is printed
> 2015-05-12 11:20:02,347 INFO
> org.apache.hadoop.yarn.server.nodemanager.NodeStatusUpdaterImpl: Initialized
> nodemanager for *null* : physical-memory=4096 virtual-memory=8602
> virtual-cores=8
> This line is printed from NodeStatusUpdaterImpl.serviceInit.
> But the nodeid assignment is happening only in
> NodeStatusUpdaterImpl.serviceStart
> {code}
> protected void serviceStart() throws Exception {
> // NodeManager is the last service to start, so NodeId is available.
> this.nodeId = this.context.getNodeId();
> {code}
> Assigning the node id in serviceinit is not feasible since it is generated by
> ContainerManagerImpl.serviceStart.
> The log can be moved to service start to give right information to user.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
