<html><head>
    <meta content="text/html; charset=utf-8" http-equiv="Content-Type" />
  
  </head><body bgcolor="#FFFFFF" text="#000000">On the other hand,  tracking that stat between versions with a known test sequence may be valuable for watching for performance issues or improvements. <br><br><div class="gmail_quote">On May 17, 2017 10:03:28 PM PDT, Ravishankar N &lt;ravishankar@redhat.com&gt; wrote:<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">

  
    <div class="moz-cite-prefix">On 05/17/2017 11:07 PM, Pranith Kumar
      Karampuri wrote:<br />
    </div>
    <blockquote cite="mid:CAOgeEnYt67XOab_tOtxF_ovddKE4d=e3zPK9W0tSq+27uOdZqA@mail.gmail.com" type="cite">
      <div dir="ltr">+ gluster-devel<br />
      </div>
      <div class="gmail_extra"><br />
        <div class="gmail_quote">On Wed, May 17, 2017 at 10:50 PM, mabi
          <span dir="ltr">&lt;<a moz-do-not-send="true" href="mailto:mabi@protonmail.ch" target="_blank">mabi@protonmail.ch</a>&gt;</span>
          wrote:<br />
          <blockquote class="gmail_quote" style="margin:0 0 0              .8ex;border-left:1px #ccc solid;padding-left:1ex">
            <div>I don't know exactly what kind of context-switches it
              was but what I know is that it is the &quot;cs&quot; number under
              &quot;system&quot; when you run vmstat.<br />
            </div>
          </blockquote>
        </div>
      </div>
    </blockquote>
    Okay, that could be due to the&nbsp; syscalls themselves or pre-emptive
    multitasking in case there aren't enough cpu cores. I think the
    spike in numbers is due to more users accessing the files at the
    same time like you observed, translating into more syscalls.&nbsp; You
    can try capturing the gluster volume profile info the next time it
    occurs and co-relate with the cs count. If you don't see any
    negative performance impact, I think you don't need to be bothered
    much by the numbers.<br />
    <br />
    HTH,<br />
    Ravi<br />
    <blockquote cite="mid:CAOgeEnYt67XOab_tOtxF_ovddKE4d=e3zPK9W0tSq+27uOdZqA@mail.gmail.com" type="cite">
      <div class="gmail_extra">
        <div class="gmail_quote">
          <blockquote class="gmail_quote" style="margin:0 0 0              .8ex;border-left:1px #ccc solid;padding-left:1ex">
            <div><br />
            </div>
            <div>Also I use the percona linux monitoring template for
              cacti (<a moz-do-not-send="true" href="https://www.percona.com/doc/percona-monitoring-plugins/LATEST/cacti/linux-templates.html" target="_blank">https://www.percona.com/doc/<wbr />percona-monitoring-plugins/<wbr />LATEST/cacti/linux-templates.<wbr />html</a>)
              which monitors context switches too. If that's of any use
              interrupts where also quite high during that time with
              peaks up to 50k interrupts.<br />
            </div>
            <div class="HOEnZb">
              <div class="h5">
                <div><br />
                </div>
                <div class="m_-9093338394098711715protonmail_signature_block  m_-9093338394098711715protonmail_signature_block-empty">
                  <div class="m_-9093338394098711715protonmail_signature_block-proton  m_-9093338394098711715protonmail_signature_block-empty"><br />
                  </div>
                </div>
                <div><br />
                </div>
                <blockquote class="m_-9093338394098711715protonmail_quote" type="cite">
                  <div>-------- Original Message --------<br />
                  </div>
                  <div>Subject: Re: [Gluster-users] 120k context
                    switches on GlsuterFS nodes<br />
                  </div>
                  <div>Local Time: May 17, 2017 2:37 AM<br />
                  </div>
                  <div>UTC Time: May 17, 2017 12:37 AM<br />
                  </div>
                  <div>From: <a moz-do-not-send="true" href="mailto:ravishankar@redhat.com" target="_blank">ravishankar@redhat.com</a><br />
                  </div>
                  <div>To: mabi &lt;<a moz-do-not-send="true" href="mailto:mabi@protonmail.ch" target="_blank">mabi@protonmail.ch</a>&gt;,
                    Gluster Users &lt;<a moz-do-not-send="true" href="mailto:gluster-users@gluster.org" target="_blank">gluster-users@gluster.org</a>&gt;<br />
                  </div>
                  <div><br />
                  </div>
                  <div> <br />
                  </div>
                  <div class="m_-9093338394098711715moz-cite-prefix">On
                    05/16/2017 11:13 PM, mabi wrote:<br />
                  </div>
                  <blockquote type="cite">
                    <div>Today I even saw up to 400k context switches
                      for around 30 minutes on my two nodes replica...
                      Does anyone else have so high context switches on
                      their GlusterFS nodes?<br />
                    </div>
                    <div><br />
                    </div>
                    <div>I am wondering what is &quot;normal&quot; and if I should
                      be worried...<br />
                    </div>
                    <div><br />
                    </div>
                    <div class="m_-9093338394098711715protonmail_signature_block  m_-9093338394098711715protonmail_signature_block-empty">
                      <div class="m_-9093338394098711715protonmail_signature_block-user  m_-9093338394098711715protonmail_signature_block-empty">
                        <div><br />
                        </div>
                      </div>
                      <div class="m_-9093338394098711715protonmail_signature_block-proton  m_-9093338394098711715protonmail_signature_block-empty"><br />
                      </div>
                    </div>
                    <div><br />
                    </div>
                    <blockquote class="m_-9093338394098711715protonmail_quote" type="cite">
                      <div>-------- Original Message --------<br />
                      </div>
                      <div>Subject: 120k context switches on GlsuterFS
                        nodes<br />
                      </div>
                      <div>Local Time: May 11, 2017 9:18 PM<br />
                      </div>
                      <div>UTC Time: May 11, 2017 7:18 PM<br />
                      </div>
                      <div>From: <a moz-do-not-send="true" rel="noreferrer nofollow noopener" href="mailto:mabi@protonmail.ch" class="m_-9093338394098711715moz-txt-link-abbreviated" target="_blank">mabi@protonmail.ch</a><br />
                      </div>
                      <div>To: Gluster Users <a moz-do-not-send="true" rel="noreferrer nofollow noopener" href="mailto:gluster-users@gluster.org" class="m_-9093338394098711715moz-txt-link-rfc2396E" target="_blank">&lt;gluster-users@gluster.org&gt;</a><br />
                      </div>
                      <div><br />
                      </div>
                      <div>Hi,<br />
                      </div>
                      <div><br />
                      </div>
                      <div>Today I noticed that for around 50 minutes my
                        two GlusterFS 3.8.11 nodes had a very high
                        amount of context switches, around 120k. Usually
                        the average is more around 1k-2k. So I checked
                        what was happening and there where just more
                        users accessing (downloading) their files at the
                        same time. These are directories with typical
                        cloud files, which means files of any sizes
                        ranging from a few kB to MB and a lot of course.<br />
                      </div>
                      <div><br />
                      </div>
                      <div>Now I never saw such a high number in context
                        switches in my entire life so I wanted to ask if
                        this is normal or to be expected? I do not find
                        any signs of errors or warnings in any log
                        files.<br />
                      </div>
                      <div><br />
                      </div>
                    </blockquote>
                  </blockquote>
                  <div class="m_-9093338394098711715protonmail_signature_block  m_-9093338394098711715protonmail_signature_block-empty">
                    <div class="m_-9093338394098711715protonmail_signature_block-user  m_-9093338394098711715protonmail_signature_block-empty">
                      <div><br />
                      </div>
                    </div>
                  </div>
                  <div>What context switch are you referring to
                    (syscalls context-switch on the bricks?) ? How did
                    you measure this?<br />
                  </div>
                  <div> -Ravi<br />
                  </div>
                  <div> <br />
                  </div>
                  <blockquote type="cite">
                    <blockquote class="m_-9093338394098711715protonmail_quote" type="cite">
                      <div>My volume is a replicated volume on two nodes
                        with ZFS as filesystem behind and the volume is
                        mounted using FUSE on the client (the cloud
                        server). On that cloud server the glusterfs
                        process was using quite a lot of system CPU but
                        that server (VM) only has 2 vCPUs so maybe I
                        should increase the number of vCPUs...<br />
                      </div>
                      <div><br />
                      </div>
                      <div>Any ideas or recommendations?<br />
                      </div>
                      <div><br />
                      </div>
                      <div class="m_-9093338394098711715protonmail_signature_block  m_-9093338394098711715protonmail_signature_block-empty">
                        <div class="m_-9093338394098711715protonmail_signature_block-user  m_-9093338394098711715protonmail_signature_block-empty">
                          <div><br />
                          </div>
                        </div>
                        <div class="m_-9093338394098711715protonmail_signature_block-proton  m_-9093338394098711715protonmail_signature_block-empty"><br />
                        </div>
                      </div>
                      <div>Regards,<br />
                      </div>
                      <div>M.<br />
                      </div>
                    </blockquote>
                    <div><br />
                    </div>
                    <div><br />
                    </div>
                    <div><br />
                    </div>
                    <pre>______________________________<wbr />_________________
Gluster-users mailing list
<a moz-do-not-send="true" rel="noreferrer nofollow noopener" href="mailto:Gluster-users@gluster.org" class="m_-9093338394098711715moz-txt-link-abbreviated" target="_blank">Gluster-users@gluster.org</a>
<a moz-do-not-send="true" rel="noreferrer nofollow noopener" href="http://lists.gluster.org/mailman/listinfo/gluster-users" class="m_-9093338394098711715moz-txt-link-freetext" target="_blank">http://lists.gluster.org/<wbr />mailman/listinfo/gluster-users</a>
</pre></blockquote><p>
</p></blockquote><div>
</div></div></div>
______________________________<wbr />_________________

Gluster-users mailing list

<a moz-do-not-send="true" href="mailto:Gluster-users@gluster.org">Gluster-users@gluster.org</a>

<a moz-do-not-send="true" href="http://lists.gluster.org/mailman/listinfo/gluster-users" rel="noreferrer" target="_blank">http://lists.gluster.org/<wbr />mailman/listinfo/gluster-users</a>
</blockquote></div>


-- 
<div class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr">Pranith
</div></div>
</div>



</blockquote><p>
</p></blockquote></div><br>
-- <br>
Sent from my Android device with K-9 Mail. Please excuse my brevity.</body></html>