<html>
  <head>
    <meta content="text/html; charset=windows-1252"
      http-equiv="Content-Type">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <br>
    <div class="moz-cite-prefix">On 11/23/2014 04:59 AM, Nipun Arora
      wrote:<br>
    </div>
    <blockquote
cite="mid:CAA9cJVV9E2g=vfcbh8QwRF2HGahrMHunf2os3whbufzqF7UH-Q@mail.gmail.com"
      type="cite">
      <meta http-equiv="Content-Type" content="text/html;
        charset=windows-1252">
      <div dir="ltr">Hi Kir,
        <div><br>
        </div>
        <div>Thanks for the response, I'll update it, and tell you about
          the results.</div>
        <div><br>
        </div>
        <div>1. A follow up question... I found that the write I/O speed
          of 500-1Mbps increased the suspend time  to several
          minutes.(mostly pcopy stage)</div>
        <div>This seems extremely high for a relatively low I/O
          workload, which is why I was wondering if there are any
          special things I need to take care of.</div>
        <div>(I ran fio (flexible i/o writer) with fixed throughput
          while doing live migration)</div>
      </div>
    </blockquote>
    <br>
    Please retry with vzctl 4.8 and ploop 1.12.1 (make sure they are on
    both sides).<br>
    There was a 5 second wait for the remote side to finish syncing<br>
    copied ploop data. It helped a case with not much I/O activity in
    container, but<br>
    ruined the case you are talking about.<br>
    <br>
    Newer ploop and vzctl implement a feedback channel for ploop copy
    that eliminates<br>
    that wait time. <br>
    <br>
    <a class="moz-txt-link-freetext" href="http://git.openvz.org/?p=ploop;a=commit;h=20d754c91079165b">http://git.openvz.org/?p=ploop;a=commit;h=20d754c91079165b</a><br>
    <a class="moz-txt-link-freetext" href="http://git.openvz.org/?p=vzctl;a=commit;h=374b759dec45255d4">http://git.openvz.org/?p=vzctl;a=commit;h=374b759dec45255d4</a><br>
    <br>
    There are some other major improvements as well, such as async send
    for ploop.<br>
    <br>
    <a class="moz-txt-link-freetext" href="http://git.openvz.org/?p=ploop;a=commit;h=a55e26e9606e0b">http://git.openvz.org/?p=ploop;a=commit;h=a55e26e9606e0b</a><br>
    <blockquote
cite="mid:CAA9cJVV9E2g=vfcbh8QwRF2HGahrMHunf2os3whbufzqF7UH-Q@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div><br>
        </div>
        <div>2. <span
            style="font-family:arial,sans-serif;font-size:13px">For my
            purposes, I have modified the live migration script to allow
            me to do cloning... i.e. I start both the containers instead
            of deleting the original. I need to do this "cloning" from
            time to time for the same target container... </span></div>
        <div style="font-family:arial,sans-serif;font-size:13px"><br>
        </div>
        <div style="font-family:arial,sans-serif;font-size:13px">     
           a. Which means that lets say we cloned container C1 to
          container C2, and let both execute at time t0, this works with
          no apparent loss of service.</div>
        <div style="font-family:arial,sans-serif;font-size:13px">       </div>
        <div style="font-family:arial,sans-serif;font-size:13px">       
          b. Now at time t1 I would like to again clone C1 to C2, and
          would like to optimize the rsync process as most of the ploop
          file for C1 and C2 should still be the same (i.e. less time to
          sync). Can anyone suggest what would be the best way to
          realize the second point? <br>
        </div>
      </div>
    </blockquote>
    <br>
    You can create a ploop snapshot and use shared base image for both
    containers<br>
    (instead of copying the base delta, hardlink it). This is not
    supported by tools<br>
    (for example, since base delta is now shared you can't merge down to
    it, but the<br>
    tools are not aware) so you need to figure it out by yourself and be
    accurate<br>
    but it should work.<br>
    <br>
    <blockquote
cite="mid:CAA9cJVV9E2g=vfcbh8QwRF2HGahrMHunf2os3whbufzqF7UH-Q@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div><br>
        </div>
        <div><br>
        </div>
        <div>Thanks</div>
        <div>Nipun</div>
      </div>
      <div class="gmail_extra"><br>
        <div class="gmail_quote">On Sun, Nov 23, 2014 at 12:56 AM, Kir
          Kolyshkin <span dir="ltr">&lt;<a moz-do-not-send="true"
              href="mailto:kir@openvz.org" target="_blank">kir@openvz.org</a>&gt;</span>
          wrote:<br>
          <blockquote class="gmail_quote" style="margin:0 0 0
            .8ex;border-left:1px #ccc solid;padding-left:1ex">
            <div text="#000000" bgcolor="#FFFFFF"><span class=""> <br>
                <div>On 11/22/2014 09:09 AM, Nipun Arora wrote:<br>
                </div>
                <blockquote type="cite">
                  <div dir="ltr">Hi All,
                    <div><br>
                    </div>
                    <div>I was wondering if anyone can suggest what is
                      the most optimal way to do the following</div>
                    <div><br>
                    </div>
                    <div>1. Can anyone clarify if ploop is the best
                      layout for minimum suspend time during live
                      migration?</div>
                  </div>
                </blockquote>
                <br>
              </span> Yes (due to ploop copy which only copies the
              modified blocks).<span class=""><br>
                <br>
                <blockquote type="cite">
                  <div dir="ltr">
                    <div><br>
                    </div>
                    <div>2. I tried migrating a ploop device where I
                      increased the --diskspace to 5G,<br>
                      and found that the suspend time taken by live
                      migration increased to 57 seconds<br>
                      (mainly undump and restore increased)...<br>
                      whereas a 2G diskspace was taking 2-3 seconds
                      suspend time... Is this expected?</div>
                    <div><br>
                    </div>
                  </div>
                </blockquote>
                <br>
              </span> No. Undump and restore times depends mostly on
              amount of RAM used by a container.<br>
              <br>
              Having said that, live migration stages influence each
              other, although it's less so<br>
              in the latest vzctl release (I won't go into details here
              if you allow me -- just make sure<br>
              you test with vzctl 4.8).<span class=""><br>
                <br>
                <br>
                <blockquote type="cite">
                  <div dir="ltr">
                    <div>3. I tried running a write intensive workload,
                      and found that beyond 100-150Kbps,<br>
                      the suspend time during live migration rapidly
                      increased? Is this an expected trend?</div>
                  </div>
                </blockquote>
                <br>
              </span> Sure. With increased writing speed, the amount of
              data that needs to be copied after CT<br>
              is suspended increases.<span class=""><br>
                <br>
                <blockquote type="cite">
                  <div dir="ltr">
                    <div><br>
                    </div>
                    <div>I am using vzctl 4.7, and ploop 1.11 in centos
                      6.5</div>
                  </div>
                </blockquote>
                <br>
              </span> You need to update vzctl and ploop and rerun your
              tests, there should be<br>
              some improvement (in particular with respect to issue #3).<span
                class=""><br>
                <br>
                <blockquote type="cite">
                  <div dir="ltr">
                    <div><br>
                    </div>
                    <div>Thanks</div>
                    <div>Nipun</div>
                  </div>
                  <br>
                  <fieldset></fieldset>
                  <br>
                  <pre>_______________________________________________
Users mailing list
<a moz-do-not-send="true" href="mailto:Users@openvz.org" target="_blank">Users@openvz.org</a>
<a moz-do-not-send="true" href="https://lists.openvz.org/mailman/listinfo/users" target="_blank">https://lists.openvz.org/mailman/listinfo/users</a>
</pre>
                </blockquote>
                <br>
              </span></div>
            <br>
            _______________________________________________<br>
            Users mailing list<br>
            <a moz-do-not-send="true" href="mailto:Users@openvz.org">Users@openvz.org</a><br>
            <a moz-do-not-send="true"
              href="https://lists.openvz.org/mailman/listinfo/users"
              target="_blank">https://lists.openvz.org/mailman/listinfo/users</a><br>
            <br>
          </blockquote>
        </div>
        <br>
      </div>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <br>
      <pre wrap="">_______________________________________________
Users mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Users@openvz.org">Users@openvz.org</a>
<a class="moz-txt-link-freetext" href="https://lists.openvz.org/mailman/listinfo/users">https://lists.openvz.org/mailman/listinfo/users</a>
</pre>
    </blockquote>
    <br>
  </body>
</html>