[YARN-6870] Fix floating point inaccuracies in resource availability check in AllocationBasedResourceUtilizationTracker

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.9.0, 3.0.0-beta1
• Component/s: api, nodemanager
• Target Version/s: 2.9.0, 3.0.0-beta1
• Assignee: Brook Zhou
• Reporter: Brook Zhou
• Created: 25/Jul/17 20:59
• Updated: 28/Jul/17 23:58
• Resolved: 28/Jul/17 23:36

Description

We have seen issues on our clusters where the current way of computing CPU usage is having float-arithmetic inaccuracies (the bug is still there in trunk)

Simple program to illustrate:

Bar.java

      public static void main(String[] args) throws Exception {
            float result = 0.0f;
            for (int i = 0; i < 7; i++) {
                  if (i == 6) {
                        result += (float) 4 / (float)18;
                  } else {
                        result += (float) 2 / (float)18;
                  }
            }
            for (int i = 0; i < 7; i++) {
                  if (i == 6) {
                        result -= (float) 4 / (float)18;
                  } else {
                        result -= (float) 2 / (float)18;
                  }                 
            }
            System.out.println(result);
      }

// Printed
4.4703484E-8

2017-04-12 05:43:24,014 INFO org.apache.hadoop.yarn.server.nodemanager.containermanager.monitor.ContainersMonitorImpl: Not enough cpu for [container_e3295_1491978508342_0467_01_000030], Current CPU Allocation: [0.8888891], Requested CPU Allocation: [0.11111111]

There are a few places with this issue:
1. ResourceUtilization.java - set/getCPU both use float. When ContainerScheduler calls ContainersMonitor.increase/decreaseResourceUtilization, this may lead to issues.

2. AllocationBasedResourceUtilizationTracker.java - hasResourcesAvailable uses float as well for CPU computation.