<html>
  <head>
    <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <br>
    Hi,<br>
    <br>
    We redid the dd tests (this time using conv=sync oflag=sync to avoid
    caching questions).  The profile results are in<br>
    <br>
<a class="moz-txt-link-freetext" href="http://mseas.mit.edu/download/phaley/GlusterUsers/profile_gluster_fuse_test">http://mseas.mit.edu/download/phaley/GlusterUsers/profile_gluster_fuse_test</a><br>
    <br>
    <br>
    <div class="moz-cite-prefix">On 05/05/2017 12:47 PM, Ravishankar N
      wrote:<br>
    </div>
    <blockquote
      cite="mid:412870d9-1003-f8f1-aaf0-6a701b393b80@redhat.com"
      type="cite">
      <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
      <div class="moz-cite-prefix">On 05/05/2017 08:42 PM, Pat Haley
        wrote:<br>
      </div>
      <blockquote
        cite="mid:b1f4209e-a0e5-5df4-f954-399df8270b21@mit.edu"
        type="cite"> <br>
        Hi Pranith,<br>
        <br>
        I presume you are asking for some version of the profile data
        that just shows the dd test (or a repeat of the dd test).  If
        yes, how do I extract just that data?<br>
      </blockquote>
      Yes, that is what he is asking for. Just clear the existing
      profile info using `gluster volume profile volname clear` and run
      the dd test once. Then when you run profile info again, it should
      just give you the stats for the dd test.<br>
      <blockquote
        cite="mid:b1f4209e-a0e5-5df4-f954-399df8270b21@mit.edu"
        type="cite"> <br>
        Thanks<br>
        <br>
        Pat<br>
        <br>
        <br>
        <br>
        <div class="moz-cite-prefix">On 05/05/2017 10:58 AM, Pranith
          Kumar Karampuri wrote:<br>
        </div>
        <blockquote
cite="mid:CAOgeEnYKv+v=Dv3y0tC9WVJSfN=dH2U9QmrXqwOOvbTGZD=yWw@mail.gmail.com"
          type="cite">
          <div dir="ltr">
            <div>
              <div>
                <div>hi Pat,<br>
                </div>
                      Let us concentrate on the performance numbers part
                for now. We will look at the permissions one after this?<br>
                <br>
              </div>
              As per the profile info, only 2.6% of the work-load is
              writes. There are too many Lookups.<br>
              <br>
            </div>
            Would it be possible to get the data for just the dd test
            you were doing earlier?<br>
            <br>
          </div>
          <div class="gmail_extra"><br>
            <div class="gmail_quote">On Fri, May 5, 2017 at 8:14 PM, Pat
              Haley <span dir="ltr">&lt;<a moz-do-not-send="true"
                  href="mailto:phaley@mit.edu" target="_blank">phaley@mit.edu</a>&gt;</span>
              wrote:<br>
              <blockquote class="gmail_quote" style="margin:0 0 0
                .8ex;border-left:1px #ccc solid;padding-left:1ex">
                <div bgcolor="#FFFFFF" text="#000000"> <br>
                  Hi Pranith &amp; Ravi,<br>
                  <br>
                  A couple of quick questions<br>
                  <br>
                  We have profile turned on. Are there specific queries
                  we should make that would help debug our
                  configuration?  (The default profile info was
                  previously sent in <a moz-do-not-send="true"
                    class="m_-8630651799581052987moz-txt-link-freetext"
href="http://lists.gluster.org/pipermail/gluster-users/2017-May/030840.html"
                    target="_blank">http://lists.gluster.org/<wbr>pipermail/gluster-users/2017-<wbr>May/030840.html</a>
                  but I'm not sure if that is what you were looking
                  for.)<br>
                  <br>
                  We also started to do a test on serving gluster over
                  NFS.  We rediscovered an issue we previously reported
                  ( <a moz-do-not-send="true"
                    class="m_-8630651799581052987moz-txt-link-freetext"
href="http://lists.gluster.org/pipermail/gluster-users/2016-September/028289.html"
                    target="_blank">http://lists.gluster.org/<wbr>pipermail/gluster-users/2016-<wbr>September/028289.html</a>
                  ) in that the NFS mounted version was ignoring the
                  group write permissions.  What specific information
                  would be useful in debugging this?<br>
                  <br>
                  Thanks<span class="HOEnZb"><font color="#888888"><br>
                      <br>
                      Pat</font></span>
                  <div>
                    <div class="h5"><br>
                      <br>
                      <br>
                      <div class="m_-8630651799581052987moz-cite-prefix">On
                        04/14/2017 03:01 AM, Ravishankar N wrote:<br>
                      </div>
                      <blockquote type="cite">
                        <div
                          class="m_-8630651799581052987moz-cite-prefix">On
                          04/14/2017 12:20 PM, Pranith Kumar Karampuri
                          wrote:<br>
                        </div>
                        <blockquote type="cite">
                          <div dir="ltr"><br>
                            <div class="gmail_extra"><br>
                              <div class="gmail_quote">On Sat, Apr 8,
                                2017 at 10:28 AM, Ravishankar N <span
                                  dir="ltr">&lt;<a
                                    moz-do-not-send="true"
                                    href="mailto:ravishankar@redhat.com"
                                    target="_blank">ravishankar@redhat.com</a>&gt;</span>
                                wrote:<br>
                                <blockquote class="gmail_quote"
                                  style="margin:0px 0px 0px
                                  0.8ex;border-left:1px solid
                                  rgb(204,204,204);padding-left:1ex">
                                  <div bgcolor="#FFFFFF">
                                    <div
                                      class="m_-8630651799581052987gmail-m_1278894059907384689moz-cite-prefix">Hi
                                      Pat,<br>
                                      <br>
                                      I'm assuming you are using gluster
                                      native (fuse mount). If it helps,
                                      you could try mounting it via
                                      gluster NFS (gnfs) and then see if
                                      there is an improvement in speed.
                                      Fuse mounts are slower than gnfs
                                      mounts but you get the benefit of
                                      avoiding a single point of
                                      failure. Unlike fuse mounts, if
                                      the gluster node containing the
                                      gnfs server goes down, all mounts
                                      done using that node will fail).
                                      For fuse mounts, you could try
                                      tweaking the write-behind xlator
                                      settings to see if it helps. See
                                      the performance.write-behind and
                                      performance.write-behind-windo<wbr>w-size
                                      options in `gluster volume set
                                      help`. Of course, even for gnfs
                                      mounts, you can achieve fail-over
                                      by using CTDB.<br>
                                    </div>
                                  </div>
                                </blockquote>
                                <div><br>
                                </div>
                                <div>Ravi,<br>
                                </div>
                                <div>      Do you have any data that
                                  suggests fuse mounts are slower than
                                  gNFS servers? <br>
                                </div>
                              </div>
                            </div>
                          </div>
                        </blockquote>
                        I have heard anecdotal evidence time and again
                        on the ML and IRC, which is why I wanted to
                        compare it with NFS numbers on his setup. <br>
                        <blockquote type="cite">
                          <div dir="ltr">
                            <div class="gmail_extra">
                              <div class="gmail_quote">
                                <div><br>
                                </div>
                                <div>Pat,<br>
                                </div>
                                <div>      I see that I am late to the
                                  thread, but do you happen to have
                                  "profile info" of the workload?<br>
                                  <br>
                                </div>
                                <div>You can follow <a
                                    moz-do-not-send="true"
href="https://gluster.readthedocs.io/en/latest/Administrator%20Guide/Monitoring%20Workload/"
                                    target="_blank">https://gluster.readthedocs.<wbr>io/en/latest/Administrator%<wbr>20Guide/Monitoring%20Workload/</a>
                                  to get the information.<br>
                                </div>
                              </div>
                            </div>
                          </div>
                        </blockquote>
                        Yeah, Let's see if profile info shows up
                        anything interesting.<br>
                        -Ravi<br>
                        <blockquote type="cite">
                          <div dir="ltr">
                            <div class="gmail_extra">
                              <div class="gmail_quote">
                                <div> </div>
                                <blockquote class="gmail_quote"
                                  style="margin:0px 0px 0px
                                  0.8ex;border-left:1px solid
                                  rgb(204,204,204);padding-left:1ex">
                                  <div bgcolor="#FFFFFF">
                                    <div
                                      class="m_-8630651799581052987gmail-m_1278894059907384689moz-cite-prefix">
                                      <br>
                                      Thanks,<br>
                                      Ravi
                                      <div>
                                        <div
                                          class="m_-8630651799581052987gmail-h5"><br>
                                          <br>
                                          On 04/08/2017 12:07 AM, Pat
                                          Haley wrote:<br>
                                        </div>
                                      </div>
                                    </div>
                                    <blockquote type="cite">
                                      <div>
                                        <div
                                          class="m_-8630651799581052987gmail-h5">
                                          <br>
                                          Hi,<br>
                                          <br>
                                          We noticed a dramatic slowness
                                          when writing to a gluster disk
                                          when compared to writing to an
                                          NFS disk. Specifically when
                                          using dd (data duplicator) to
                                          write a 4.3 GB file of zeros:<br>
                                          <ul>
                                            <li>on NFS disk (/home): 9.5
                                              Gb/s</li>
                                            <li>on gluster disk
                                              (/gdata): 508 Mb/s<br>
                                            </li>
                                          </ul>
                                          The gluser disk is 2 bricks
                                          joined together, no
                                          replication or anything else.
                                          The hardware is (literally)
                                          the same:<br>
                                          <ul>
                                            <li>one server with 70 hard
                                              disks  and a hardware RAID
                                              card.</li>
                                            <li>4 disks in a RAID-6
                                              group (the NFS disk)</li>
                                            <li>32 disks in a RAID-6
                                              group (the max allowed by
                                              the card, /mnt/brick1)</li>
                                            <li>32 disks in another
                                              RAID-6 group (/mnt/brick2)</li>
                                            <li>2 hot spare<br>
                                            </li>
                                          </ul>
                                          <p>Some additional information
                                            and more tests results
                                            (after changing the log
                                            level):<br>
                                          </p>
                                          <p><span>glusterfs 3.7.11
                                              built on Apr 27 2016
                                              14:09:22</span><br>
                                            <span>CentOS release 6.8
                                              (Final)</span><br>
                                            RAID bus controller: LSI
                                            Logic / Symbios Logic
                                            MegaRAID SAS-3 3108
                                            [Invader] (rev 02)<br>
                                            <br>
                                            <br>
                                            <br>
                                            <b>Create the file to /gdata
                                              (gluster)</b><br>
                                            [root@mseas-data2 gdata]# dd
                                            if=/dev/zero of=/gdata/zero1
                                            bs=1M count=1000<br>
                                            1000+0 records in<br>
                                            1000+0 records out<br>
                                            1048576000 bytes (1.0 GB)
                                            copied, 1.91876 s, <b>546
                                              MB/s</b><br>
                                            <br>
                                            <b>Create the file to /home
                                              (ext4)</b><br>
                                            [root@mseas-data2 gdata]# dd
                                            if=/dev/zero of=/home/zero1
                                            bs=1M count=1000<br>
                                            1000+0 records in<br>
                                            1000+0 records out<br>
                                            1048576000 bytes (1.0 GB)
                                            copied, 0.686021 s, <b>1.5
                                              GB/s - </b>3 times as
                                            fast<b><br>
                                              <br>
                                              <br>
                                              Copy from /gdata to /gdata
                                              (gluster to gluster)<br>
                                            </b>[root@mseas-data2
                                            gdata]# dd if=/gdata/zero1
                                            of=/gdata/zero2<br>
                                            2048000+0 records in<br>
                                            2048000+0 records out<br>
                                            1048576000 bytes (1.0 GB)
                                            copied, 101.052 s, <b>10.4
                                              MB/s</b> - realllyyy
                                            slooowww<br>
                                            <br>
                                            <br>
                                            <b>Copy from /gdata to
                                              /gdata</b> <b>2nd time <b>(gluster
                                                to gluster)</b></b><br>
                                            [root@mseas-data2 gdata]# dd
                                            if=/gdata/zero1
                                            of=/gdata/zero2<br>
                                            2048000+0 records in<br>
                                            2048000+0 records out<br>
                                            1048576000 bytes (1.0 GB)
                                            copied, 92.4904 s, <b>11.3
                                              MB/s</b> <span>-
                                              realllyyy slooowww</span>
                                            again<br>
                                            <br>
                                            <br>
                                            <br>
                                            <b>Copy from /home to /home
                                              (ext4 to ext4)</b><br>
                                            [root@mseas-data2 gdata]# dd
                                            if=/home/zero1
                                            of=/home/zero2<br>
                                            2048000+0 records in<br>
                                            2048000+0 records out<br>
                                            1048576000 bytes (1.0 GB)
                                            copied, 3.53263 s, <b>297
                                              MB/s </b>30 times as fast<br>
                                            <br>
                                            <br>
                                            <b>Copy from /home to /home
                                              (ext4 to ext4)</b><br>
                                            [root@mseas-data2 gdata]# dd
                                            if=/home/zero1
                                            of=/home/zero3<br>
                                            2048000+0 records in<br>
                                            2048000+0 records out<br>
                                            1048576000 bytes (1.0 GB)
                                            copied, 4.1737 s, <b>251
                                              MB/s</b> <span>- 30 times
                                              as fast<br>
                                              <br>
                                              <br>
                                              As a test, can we copy
                                              data directly to the xfs
                                              mountpoint (/mnt/brick1)
                                              and bypass gluster?<br>
                                              <br>
                                              <br>
                                              Any help you could give us
                                              would be appreciated.<br>
                                              <br>
                                            </span>Thanks<br>
                                          </p>
                                          <pre class="m_-8630651799581052987gmail-m_1278894059907384689moz-signature" cols="72">-- 

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=<wbr>-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=<wbr>-=-=-
Pat Haley                          Email:  <a moz-do-not-send="true" class="m_-8630651799581052987gmail-m_1278894059907384689moz-txt-link-abbreviated" href="mailto:phaley@mit.edu" target="_blank">phaley@mit.edu</a>
Center for Ocean Engineering       Phone:  (617) 253-6824
Dept. of Mechanical Engineering    Fax:    (617) 253-8125
MIT, Room 5-213                    <a moz-do-not-send="true" class="m_-8630651799581052987gmail-m_1278894059907384689moz-txt-link-freetext" href="http://web.mit.edu/phaley/www/" target="_blank">http://web.mit.edu/phaley/www/</a>
77 Massachusetts Avenue
Cambridge, MA  02139-4301
</pre>
      

      <fieldset class="m_-8630651799581052987gmail-m_1278894059907384689mimeAttachmentHeader"></fieldset>
      

      </div></div><pre>______________________________<wbr>_________________
Gluster-users mailing list
<a moz-do-not-send="true" class="m_-8630651799581052987gmail-m_1278894059907384689moz-txt-link-abbreviated" href="mailto:Gluster-users@gluster.org" target="_blank">Gluster-users@gluster.org</a>
<a moz-do-not-send="true" class="m_-8630651799581052987gmail-m_1278894059907384689moz-txt-link-freetext" href="http://lists.gluster.org/mailman/listinfo/gluster-users" target="_blank">http://lists.gluster.org/mailm<wbr>an/listinfo/gluster-users</a></pre>
    </blockquote>
    <p>

    </p>
  </div>


______________________________<wbr>_________________

Gluster-users mailing list

<a moz-do-not-send="true" href="mailto:Gluster-users@gluster.org" target="_blank">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/mailm<wbr>an/listinfo/gluster-users</a>
</blockquote></div>


-- 
<div class="m_-8630651799581052987gmail_signature"><div dir="ltr">Pranith
</div></div>
</div></div>



</blockquote><p>
</p>


</blockquote>
<pre class="m_-8630651799581052987moz-signature" cols="72">-- 

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=<wbr>-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=<wbr>-=-=-
Pat Haley                          Email:  <a moz-do-not-send="true" class="m_-8630651799581052987moz-txt-link-abbreviated" href="mailto:phaley@mit.edu" target="_blank">phaley@mit.edu</a>
Center for Ocean Engineering       Phone:  (617) 253-6824
Dept. of Mechanical Engineering    Fax:    (617) 253-8125
MIT, Room 5-213                    <a moz-do-not-send="true" class="m_-8630651799581052987moz-txt-link-freetext" href="http://web.mit.edu/phaley/www/" target="_blank">http://web.mit.edu/phaley/www/</a>
77 Massachusetts Avenue
Cambridge, MA  02139-4301
</pre></div></div></div></blockquote></div>


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



</blockquote>
<pre class="moz-signature" cols="72">-- 

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Pat Haley                          Email:  <a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:phaley@mit.edu">phaley@mit.edu</a>
Center for Ocean Engineering       Phone:  (617) 253-6824
Dept. of Mechanical Engineering    Fax:    (617) 253-8125
MIT, Room 5-213                    <a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://web.mit.edu/phaley/www/">http://web.mit.edu/phaley/www/</a>
77 Massachusetts Avenue
Cambridge, MA  02139-4301
</pre>


</blockquote><p>
</p>


</blockquote>
<pre class="moz-signature" cols="72">-- 

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Pat Haley                          Email:  <a class="moz-txt-link-abbreviated" href="mailto:phaley@mit.edu">phaley@mit.edu</a>
Center for Ocean Engineering       Phone:  (617) 253-6824
Dept. of Mechanical Engineering    Fax:    (617) 253-8125
MIT, Room 5-213                    <a class="moz-txt-link-freetext" href="http://web.mit.edu/phaley/www/">http://web.mit.edu/phaley/www/</a>
77 Massachusetts Avenue
Cambridge, MA  02139-4301
</pre></body></html>