FangYongs opened a new pull request, #78: URL: https://github.com/apache/flink-benchmarks/pull/78
This PR aims to add high availability service micro benchmark for olap. It will compare the throughputs for none and zookeeper ha service in dispatcher to show that, the current high availability service should be divided into cluster and job high availability services, which have been discussed in https://issues.apache.org/jira/browse/FLINK-32667. Then we can configure none ha service for job when cluster ha service is needed in flink cluster to improve the performance for olap jobs. The main work of this PR include: 1. Create a `MiniCluster` with `ZOOKEEPER` and `NONE` high availability services. 2. Create a job with one no-op vertex 3. Count the throughput of submitting jobs per second The results show that the throughput with none HA is more than twice that of zookeeper HA. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
