[
https://issues.apache.org/jira/browse/YARN-4372?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15018418#comment-15018418
]
Naganarasimha G R commented on YARN-4372:
-----------------------------------------
hi [~vinodkv],
bq. Can you try my uploaded patch here together with YARN-4350 and see if you
still find issues?
took the latest code from trunk as YARN-2859 is commited, later just added the
below modification in
{{TestDistributedShell}}
{code}
@@ -80,6 +80,7 @@ protected void setupInternal(int numNodeManager) throws
Exception {
conf.setInt(YarnConfiguration.RM_SCHEDULER_MINIMUM_ALLOCATION_MB, 128);
conf.set("yarn.log.dir", "target");
conf.setBoolean(YarnConfiguration.TIMELINE_SERVICE_ENABLED, true);
+ conf.setBoolean(YarnConfiguration.RM_SYSTEM_METRICS_PUBLISHER_ENABLED,
true);
conf.set(YarnConfiguration.RM_SCHEDULER,
CapacityScheduler.class.getName());
conf.setBoolean(YarnConfiguration.NODE_LABELS_ENABLED, true);
conf.set("mapreduce.jobhistory.address",
{code}
Even after the patch {{TestDistributedShell.testDSShellWithoutDomain}} is
failing (test case passes but the in the console logs there were logs for
unreachable timlineserver for each smp events). Crossverified the {{resURI}} is
getting set properly in {{TimelineClientImpl}}. This is the same which i had
faced when i was last analyzing and i too had done the similar modifications in
MiniYarnCluster.
Few console logs
{code}
2015-11-20 23:17:15,867 ERROR [AsyncDispatcher event handler]
impl.TimelineClientImpl (TimelineClientImpl.java:doPosting(336)) - Failed to
get the response from the timeline server.
2015-11-20 23:17:15,868 ERROR [AsyncDispatcher event handler]
metrics.SystemMetricsPublisher (SystemMetricsPublisher.java:putEntity(485)) -
Error when publishing entity
[YARN_CONTAINER,container_1448041615020_0001_01_000002]
{code}
bq. The separate thread is not an issue as RM will only start after AHS
successfully starts.
Yes you are right, my mistake had just observed that it was started as part of
thread but we are further checking whether the service is started and throwing
exception if necessary.
bq. Even if that store is specified, TimelineDataManager will continue to be
instantiated in the server, so your assumption is wrong that levelDB is not
created. Agree?
Oops had wrongly mentioned *levelDBtimelinestore* but wanted to mention as
{{ApplicationHistoryManagerOnTimelineStore}}.
As any way we are starting Timelinestore so better to have
ApplicationHistoryManagerOnTimelineStore than the almost deprecated
{{yarn.timeline-service.generic-application-history.store-class}} interface .
(not a must fix but better to have)
> Cannot enable system-metrics-publisher inside MiniYARNCluster
> -------------------------------------------------------------
>
> Key: YARN-4372
> URL: https://issues.apache.org/jira/browse/YARN-4372
> Project: Hadoop YARN
> Issue Type: Bug
> Reporter: Vinod Kumar Vavilapalli
> Assignee: Vinod Kumar Vavilapalli
> Attachments: YARN-4372-20151119.1.txt
>
>
> [~Naganarasimha] found this at YARN-2859, see [this
> comment|https://issues.apache.org/jira/browse/YARN-2859?focusedCommentId=15005746&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15005746].
> The way daemons are started inside MiniYARNCluster, RM is not setup correctly
> to send information to TimelineService.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
