Created 10-03-2016 03:03 PM
Hi,
I'm doing POC on HBase and Phoenix Query Server for next project. We would like to tune both the services for the best outcome for operations on 100million records. Right now we are using most of the default parameters which come with Hortonworks installation.
Can anyone provide the best set of tuning properties operations on 100million records?
Any help is highly appreciated and thanks in advance.
Created 10-04-2016 06:13 AM
For HBase tuning you can refer following links:
http://www.slideshare.net/lhofhansl/h-base-tuninghbasecon2015ok
For Phoenix tuning following link helps:
http://phoenix.apache.org/tuning.html
Use http://phoenix.apache.org/update_statistics.html for more parallelization and better performance.
For more phoenix level optimizations You can refer Optimization sections in
http://www.slideshare.net/je2451/apache-phoenix-and-apache-hbase-an-enterprise-grade-data-warehouse
Created 10-04-2016 06:13 AM
For HBase tuning you can refer following links:
http://www.slideshare.net/lhofhansl/h-base-tuninghbasecon2015ok
For Phoenix tuning following link helps:
http://phoenix.apache.org/tuning.html
Use http://phoenix.apache.org/update_statistics.html for more parallelization and better performance.
For more phoenix level optimizations You can refer Optimization sections in
http://www.slideshare.net/je2451/apache-phoenix-and-apache-hbase-an-enterprise-grade-data-warehouse
Created 10-04-2016 05:34 PM
Some extra thoughts on top of Rajeshbabu's reply:
1. Increase the heapsize of the Phoenix Query Server via the PHOENIX_QUERYSERVER_OPTS variable hbase-env.sh
2. For writing data, make sure the addBatch() and executeBatch() API calls are used for the best performance