[AMBARI-20071] Hadoop metrics sink prints lots of logs if collector is unavailable

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 2.5.0
• Fix Version/s: 2.5.0
• Component/s: ambari-metrics
• Assignee: Aravindan Vijayan
• Reporter: Aravindan Vijayan
• Created: 17/Feb/17 19:55
• Updated: 21/Feb/17 19:14
• Resolved: 21/Feb/17 18:29

Description

The metrics sink prints lots of such messages in Hadoop daemons log every second, which makes logs rotates and purge fast.

2017-02-16 19:05:48,896 INFO  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:refreshCollectorsFromConfigured(419)) - Collector ctr-e129-1487033772569-2546-01-000004.hwx.site is not longer live. Removing it from list of know live collector hosts : []
2017-02-16 19:05:48,896 WARN  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:findPreferredCollectHost(399)) - Couldn't find any live collectors. Returning null
2017-02-16 19:05:48,896 WARN  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:emitMetrics(227)) - No live collector to send metrics to. Metrics to be sent will be discarded.
2017-02-16 19:05:50,901 WARN  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:findLiveCollectorHostsFromKnownCollector(476)) - Unable to connect to collector to find live nodes.
2017-02-16 19:05:50,901 INFO  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:refreshCollectorsFromConfigured(419)) - Collector ctr-e129-1487033772569-2546-01-000004.hwx.site is not longer live. Removing it from list of know live collector hosts : []
2017-02-16 19:05:50,901 WARN  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:findPreferredCollectHost(399)) - Couldn't find any live collectors. Returning null
2017-02-16 19:05:50,902 WARN  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:emitMetrics(227)) - No live collector to send metrics to. Metrics to be sent will be discarded.
2017-02-16 19:06:48,896 WARN  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:findLiveCollectorHostsFromKnownCollector(476)) - Unable to connect to collector to find live nodes.
2017-02-16 19:06:48,897 INFO  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:refreshCollectorsFromConfigured(419)) - Collector ctr-e129-1487033772569-2546-01-000004.hwx.site is not longer live. Removing it from list of know live collector hosts : []
2017-02-16 19:06:48,897 INFO  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:findPreferredCollectHost(359)) - No live collectors from configuration. Requesting zookeeper...
2017-02-16 19:06:48,988 INFO  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:findPreferredCollectHost(369)) - No new collector was found from Zookeeper. Will not request zookeeper for 120000 millis
2017-02-16 19:06:48,989 INFO  availability.MetricSinkWriteShardHostnameHashingStrategy (MetricSinkWriteShardHostnameHashingStrategy.java:findCollectorShard(42)) - Calculated collector shard ctr-e129-1487033772569-2546-01-000004.hwx.site based on hostname: ctr-e129-1487033772569-2546-01-000003.hwx.site
2017-02-16 19:06:59,004 INFO  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:emitMetrics(217)) - Removing collector ctr-e129-1487033772569-2546-01-000004.hwx.site from allKnownLiveCollectors.
2017-02-16 19:07:01,009 WARN  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:findLiveCollectorHostsFromKnownCollector(476)) - Unable to connect to collector to find live nodes.
2017-02-16 19:07:01,010 INFO  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:refreshCollectorsFromConfigured(419)) - Collector ctr-e129-1487033772569-2546-01-000004.hwx.site is not longer live. Removing it from list of know live collector hosts : []
2017-02-16 19:07:01,010 WARN  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:findPreferredCollectHost(399)) - Couldn't find any live collectors. Returning null
2017-02-16 19:07:01,010 WARN  timeline.HadoopTimelineMetricsSink (AbstractTimelineMetricsSink.java:emitMetrics(227)) - No live collector to send metrics to. Metrics to be sent will be discarded.

Issue Links

• links to: Review Board