You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@storm.apache.org by "Boyang Jerry Peng (JIRA)" <ji...@apache.org> on 2015/11/04 00:09:27 UTC

[jira] [Created] (STORM-1165) normalize the scales of CPU/Mem/Net when choosing the best node for Resource Aware Scheduler

Boyang Jerry Peng created STORM-1165:
----------------------------------------

             Summary: normalize the scales of CPU/Mem/Net when choosing the best node for Resource Aware Scheduler
                 Key: STORM-1165
                 URL: https://issues.apache.org/jira/browse/STORM-1165
             Project: Apache Storm
          Issue Type: Improvement
            Reporter: Boyang Jerry Peng


1. Describe observed behavior.
Memory has a default capacity of 2000.0, CPU has 400.0, network has a distance value of either 1 or 2 (different rack); and the weights of them are the same (1). 
As a result, when choosing the best node for scheduling, the network plays a minimal role in the calculation of the geometric square of the three dimensions.
2. What is the expected behavior?
Three resources place comparable roles in the node choosing algorithm.
3. Outline the steps to reproduce the problem.
Need to assign appropriate or normalized scales to balance among the three dimensions.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)