[YARN-3375] NodeHealthScriptRunner.shouldRun() check is performing 3 times for starting NodeHealthScriptRunner

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 2.8.0, 3.0.0-alpha1
• Component/s: nodemanager
• Target Version/s: 2.8.0
• Hadoop Flags: Reviewed
• Assignee: Devaraj Kavali
• Reporter: Devaraj Kavali
• Created: 19/Mar/15 09:04
• Updated: 30/Aug/16 01:25
• Resolved: 04/May/15 23:04

Description

1. NodeHealthScriptRunner.shouldRun() check is happening 3 times for starting the NodeHealthScriptRunner.

NodeManager.java

    if(!NodeHealthScriptRunner.shouldRun(nodeHealthScript)) {
      LOG.info("Abey khali");
      return null;
    }

NodeHealthCheckerService.java

    if (NodeHealthScriptRunner.shouldRun(
        conf.get(YarnConfiguration.NM_HEALTH_CHECK_SCRIPT_PATH))) {
      addService(nodeHealthScriptRunner);
    }

NodeHealthScriptRunner.java

    if (!shouldRun(nodeHealthScript)) {
      LOG.info("Not starting node health monitor");
      return;
    }

2. If we don't configure node health script or configured health script doesn't execute permission, NM logs with the below message.

2015-03-19 19:55:45,713 INFO org.apache.hadoop.yarn.server.nodemanager.NodeManager: Abey khali

Issue Links

• duplicates: YARN-2832 Wrong Check Logic of NodeHealthCheckerService Causes Latent Errors
• is related to: YARN-2980 Move health check script related functionality to hadoop-common