I am using hdp 220.127.116.11-227. I am trying to enable phoenix tracing to monitor the queries and to analyze performance. I followed steps outlined here - https://phoenix.apache.org/tracing.html
i placed the hadoop-metrics2-hbase.properties in /etc/hbase/conf
hadoop-metrics2-phoenix.properties in /usr/hdp/18.104.22.168-227/phoenix/bin in all regions servers
Aslo added following properties to hbase.site
<name>phoenix.trace.frequency</name> <value>always</value>. After this
I am not clear where to place the ddl for SYSTEM.TRACING_STATS. Also i could not see ./bin/traceserver.py to start
@Ankit Singhal: i was trying to explore, phoenix tracing if there is a way to track the long running queries(scans) currently and try to kill them...
I checked https://issues.apache.org/jira/browse/PHOENIX-2640 which is still open.
Also can this tracing be kept on in production always if there is no impact to running queries or to be enabled only during an issue.