[
https://issues.apache.org/jira/browse/AMBARI-20403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15905887#comment-15905887
]
Aravindan Vijayan commented on AMBARI-20403:
--------------------------------------------
PROBLEM
The collector URI used for yarn container metrics is initially seeded and not
refreshed based on collector availability (HA changes). A couple of other
places where the Container metrics URL was wrongly invoked.
FIX
Changed container metrics path to use collector host HA changes.
> Yarn Container metrics are not being collected by AMS
> ------------------------------------------------------
>
> Key: AMBARI-20403
> URL: https://issues.apache.org/jira/browse/AMBARI-20403
> Project: Ambari
> Issue Type: Bug
> Components: ambari-metrics
> Affects Versions: 2.5.0
> Reporter: Aravindan Vijayan
> Assignee: Aravindan Vijayan
> Priority: Blocker
> Fix For: 2.5.0
>
> Attachments: AMBARI-20403.patch
>
>
> Exception in NM logs :
> {code:java}
> 2017-03-07 19:14:06,143 INFO timeline.HadoopTimelineMetricsSink
> (AbstractTimelineMetricsSink.java:emitMetricsJson(197)) - Unable to connect
> to collector, null
> This exceptions will be ignored for next 100 times
>
> 2017-03-07 19:14:06,144 WARN timeline.HadoopTimelineMetricsSink
> (HadoopTimelineMetricsSink.java:putMetrics(344)) - Unable to send metrics to
> collector by address:null
> {code}
> 0jdbc:phoenix:localhost:2181:/ams-hbase-uns> select * from CONTAINER_METRICS
> order by START_TIME desc limit 10;
> +------------+--------------+------------+-------------+------------+------------+------------+-----------------------+-----------------+------------------+-------------------+
> | APP_ID | CONTAINER_ID | START_TIME | FINISH_TIME | DURATION |
> HOSTNAME | EXIT_CODE | LOCALIZATION_DURATION | LAUNCH_DURATION |
> MEM_REQUESTED_GB | MEM_REQUESTED_GB_ |
> +------------+--------------+------------+-------------+------------+------------+------------+-----------------------+-----------------+------------------+-------------------+
> +------------+--------------+------------+-------------+------------+------------+------------+-----------------------+-----------------+------------------+-------------------+
> No rows selected (0.505 seconds)
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
