<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>question Re: yarn.nodemanager.resource.cpu-vcores is by default coming up with value multiplied by 2 in Archives of Support Questions (Read Only)</title>
    <link>https://community.cloudera.com/t5/Archives-of-Support-Questions/yarn-nodemanager-resource-cpu-vcores-is-by-default-coming-up/m-p/208166#M76766</link>
    <description>&lt;P&gt;Hi &lt;A rel="user" href="https://community.cloudera.com/users/13950/dheervijjiraghu.html" nodeid="13950"&gt;@Venkata Sudheer Kumar M&lt;/A&gt; CPU is capable of running multiple containers if the jobs are not cpu intensive. &lt;/P&gt;&lt;P&gt;The stack advisor is only recommending to not go beyond "CPU(s) * 2". However, there is nothing stopping you from configuring higher.&lt;/P&gt;&lt;P&gt;if you observe your container concurrency metrics and CPU utilization, you can identify your threshold of vcores to 1 CPU and set it accordingly. Note: &lt;STRONG&gt;yarn.nodemanager.resource.cpu-vcores &lt;/STRONG&gt;would only be applicable if you &lt;STRONG&gt;enable CPU Scheduling&lt;/STRONG&gt;.&lt;/P&gt;</description>
    <pubDate>Wed, 04 Apr 2018 04:08:23 GMT</pubDate>
    <dc:creator>smayani</dc:creator>
    <dc:date>2018-04-04T04:08:23Z</dc:date>
    <item>
      <title>yarn.nodemanager.resource.cpu-vcores is by default coming up with value multiplied by 2</title>
      <link>https://community.cloudera.com/t5/Archives-of-Support-Questions/yarn-nodemanager-resource-cpu-vcores-is-by-default-coming-up/m-p/208165#M76765</link>
      <description>&lt;P&gt;
	We have cluster with below CPU configuration:&lt;/P&gt;
&lt;TABLE&gt;

&lt;TBODY&gt;
&lt;TR&gt;
	&lt;TD&gt;
		# lscpu
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		Architecture:  x86_64
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		CPU op-mode(s):  32-bit, 64-bit
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		Byte Order:  Little Endian
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		CPU(s):  56
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		On-line CPU(s) list:  0-55
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		Thread(s) per core:  2
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		Core(s) per socket:  14
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		Socket(s):  2
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		NUMA node(s):  2
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		Vendor ID:  GenuineIntel
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		CPU family:  6
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		Model:  79
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		Stepping:  1
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		CPU MHz:  2400.000
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		BogoMIPS:  4794.00
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		Virtualization:  VT-x
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		L1d cache:  32K
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		L1i cache:  32K
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		L2 cache:  256K
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		L3 cache:  35840K
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		NUMA node0 CPU(s):  0-13,28-41
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;TR&gt;
	&lt;TD&gt;
		NUMA node1 CPU(s):  14-27,42-55
	&lt;/TD&gt;
&lt;/TR&gt;
&lt;/TBODY&gt;
&lt;/TABLE&gt;&lt;P&gt;We have 2 Physical Cores, 14 CPU's each, with Hyper threading 2 (physical) * 14 (cpu's each) * 2 (hyder threading) =  56&lt;/P&gt;&lt;P&gt;But the YARN Configs from Ambari shows 112 cores for the property &lt;STRONG&gt;yarn.nodemanager.resource.cpu-vcores (56*2)&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;This is being done at the &lt;A href="https://github.com/apache/ambari/blob/trunk/ambari-server/src/main/resources/stacks/HDP/2.2/services/stack_advisor.py" target="_blank"&gt;Ambari Stack Advisor Code&lt;/A&gt;&lt;/P&gt;&lt;P&gt;&lt;A href="https://github.com/apache/ambari/blob/trunk/ambari-server/src/main/resources/stacks/HDP/2.2/services/stack_advisor.py" target="_blank"&gt;&lt;/A&gt;The question here is are we by default doing the multiplication by 2  by assuming that the Hyper Threading is not enabled or are we considering that the CPU is capable of holding multiple containers and leaving the scope to admins to tune the environment based on CPU  or I/O work loads.&lt;/P&gt;&lt;P&gt;Thanks&lt;/P&gt;&lt;P&gt;Venkat&lt;/P&gt;</description>
      <pubDate>Tue, 03 Apr 2018 12:42:49 GMT</pubDate>
      <guid>https://community.cloudera.com/t5/Archives-of-Support-Questions/yarn-nodemanager-resource-cpu-vcores-is-by-default-coming-up/m-p/208165#M76765</guid>
      <dc:creator>dheer_vijji_rag</dc:creator>
      <dc:date>2018-04-03T12:42:49Z</dc:date>
    </item>
    <item>
      <title>Re: yarn.nodemanager.resource.cpu-vcores is by default coming up with value multiplied by 2</title>
      <link>https://community.cloudera.com/t5/Archives-of-Support-Questions/yarn-nodemanager-resource-cpu-vcores-is-by-default-coming-up/m-p/208166#M76766</link>
      <description>&lt;P&gt;Hi &lt;A rel="user" href="https://community.cloudera.com/users/13950/dheervijjiraghu.html" nodeid="13950"&gt;@Venkata Sudheer Kumar M&lt;/A&gt; CPU is capable of running multiple containers if the jobs are not cpu intensive. &lt;/P&gt;&lt;P&gt;The stack advisor is only recommending to not go beyond "CPU(s) * 2". However, there is nothing stopping you from configuring higher.&lt;/P&gt;&lt;P&gt;if you observe your container concurrency metrics and CPU utilization, you can identify your threshold of vcores to 1 CPU and set it accordingly. Note: &lt;STRONG&gt;yarn.nodemanager.resource.cpu-vcores &lt;/STRONG&gt;would only be applicable if you &lt;STRONG&gt;enable CPU Scheduling&lt;/STRONG&gt;.&lt;/P&gt;</description>
      <pubDate>Wed, 04 Apr 2018 04:08:23 GMT</pubDate>
      <guid>https://community.cloudera.com/t5/Archives-of-Support-Questions/yarn-nodemanager-resource-cpu-vcores-is-by-default-coming-up/m-p/208166#M76766</guid>
      <dc:creator>smayani</dc:creator>
      <dc:date>2018-04-04T04:08:23Z</dc:date>
    </item>
    <item>
      <title>Re: yarn.nodemanager.resource.cpu-vcores is by default coming up with value multiplied by 2</title>
      <link>https://community.cloudera.com/t5/Archives-of-Support-Questions/yarn-nodemanager-resource-cpu-vcores-is-by-default-coming-up/m-p/208167#M76767</link>
      <description>&lt;P&gt;&lt;A rel="user" href="https://community.cloudera.com/users/220/smayani.html" nodeid="220"&gt;@Saumil Mayani&lt;/A&gt; Thanks a lot for the details. That makes it more clear.&lt;/P&gt;</description>
      <pubDate>Wed, 04 Apr 2018 15:27:37 GMT</pubDate>
      <guid>https://community.cloudera.com/t5/Archives-of-Support-Questions/yarn-nodemanager-resource-cpu-vcores-is-by-default-coming-up/m-p/208167#M76767</guid>
      <dc:creator>dheer_vijji_rag</dc:creator>
      <dc:date>2018-04-04T15:27:37Z</dc:date>
    </item>
  </channel>
</rss>

