<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body>
    <br>
    <br>
    <div class="moz-cite-prefix">On 4/16/16 1:33 PM, Flavio Percoco
      wrote:<br>
    </div>
    <blockquote cite="mid:20160416173317.GL24746@redhat.com" type="cite">On
      15/04/16 11:42 -0400, Nikhil Komawar wrote:
      <br>
      <blockquote type="cite">comment inline
        <br>
        <br>
        On 4/15/16 11:08 AM, Sean Dague wrote:
        <br>
        <blockquote type="cite">On 04/15/2016 10:42 AM, Jay Pipes wrote:
          <br>
          <blockquote type="cite">On 04/01/2016 06:45 AM, Sean Dague
            wrote:
            <br>
            <blockquote type="cite">#2 - move discover major version
              back to glanceclient -
              <br>
<a class="moz-txt-link-freetext" href="https://github.com/openstack/nova/blob/3cdaa30566c17a2add5d9163a0693c97dc1d065b/nova/image/glance.py#L108">https://github.com/openstack/nova/blob/3cdaa30566c17a2add5d9163a0693c97dc1d065b/nova/image/glance.py#L108</a>
              <br>
              <br>
              <br>
              I don't understand why this was ever in nova. This really
              should be
              <br>
              <br>
              glanceclient.discover... something. It uses internal
              methods from
              <br>
              glanceclient and internal structures of the content
              returned.
              <br>
              <br>
              Catching, if desired, should also be on the glanceclient
              side.
              <br>
              glanceclient.reset_version() could exist to clear any
              caching.
              <br>
            </blockquote>
            This is exactly what I said in the original review in Mitaka
            :(
            <br>
            <br>
            <a class="moz-txt-link-freetext" href="https://review.openstack.org/#/c/222150/">https://review.openstack.org/#/c/222150/</a>
            <br>
            <br>
            Note that on PS10 I wrote:
            <br>
            <br>
            This code belongs in glanceclient, not Nova, IMHO...
            <br>
            <br>
            Line 169: Most of the above code should actually be in the
            <br>
            python-glanceclient package, not here. The Nova code should
            be able to
            <br>
            call glanceclient with a URI and get the latest supported
            Glance API
            <br>
            version.
            <br>
            <br>
            and then later I said:
            <br>
            <br>
            To be a little clearer... the Nova code should be able to do
            something
            <br>
            like this:
            <br>
            <br>
             glance_uris = CONF.glance_api_servers
            <br>
             glance_uri_version_map = {glance_uri:
            glanceclient.get_latest_version(uri)
            <br>
                        for uri in glance_uris}
            <br>
            <br>
            So... pretty much in line with what you say above.
            <br>
            <br>
            Flavio then responded to me:
            <br>
            <br>
            "While I agree with you, I believe we should let this in
            "for now"
            <br>
            whilst it's added to glanceclient. One of the things that
            blocked the
            <br>
            previous work during the Kilo cycle were things being added
            to
            <br>
            glanceclient and the fact that they weren't available right
            away.
            <br>
            <br>
            Can we agree on removing this code as soon as there's a
            glanceclient
            <br>
            release with it? Happy to have a bug filed against
            glanceclient. The
            <br>
            glance team will take care of this."
            <br>
            <br>
            and I wrote:
            <br>
            <br>
            "OK, if you promise to remove some of this code when
            glanceclient gets
            <br>
            this functionality, then I suppose I'm good with this going
            in Nova for
            <br>
            now."
            <br>
            <br>
            So, there's your answer to "why this was ever in Nova".
            <br>
          </blockquote>
          My expectation is that the turn around time on something like
          this would
          <br>
          have been weeks, not months. I feel like the delays getting
          things into
          <br>
          glanceclient makes me a pretty firm -2 on "let's just hack it
          into Nova
          <br>
          for now". Because for now seems to equal for ever. :(
          <br>
        </blockquote>
        <br>
        Sure, let's avoid more delay.
        <br>
        <br>
        <blockquote type="cite">Honestly, staring at all this again this
          morning I think that version
          <br>
          discovery in Nova is probably just complexity we don't need.
          Especially
          <br>
          because it tends to lead to code that goes and leans on
          version
          <br>
          discovery at weird deep layers, and it turns out you are using
          v1 for a
          <br>
          piece of a flow and v2 for a different piece.
          <br>
        </blockquote>
        <br>
        I can figure out a way to get this done in glanceclient
        soon-ish.
        <br>
      </blockquote>
      <br>
      FWIW, We had figure out a way to do *ALL* this in glanceclient
      during Mitaka but
      <br>
      then nova patches were blocked and the plans we had agreed on were
      changed.
      <br>
      Therefore, The glance team decided to dedicate resources on other
      priorities
      <br>
      that were actually going to make it.
      <br>
      <br>
      So, forgive me if I jump in with a defensive tone but I disagree
      with the
      <br>
      feeling this would've taken forever. The Glance team was already
      acting towards
      <br>
      this but then *Nova* happened.
      <br>
      <br>
    </blockquote>
    <br>
    Thanks for the clarification Flavio. I guess now we can move on with
    the alternate plan proposed and keep the momentum on deprecating
    Glance v1 intact. I'm hoping that the cross track (nova-glance)
    summit session (Thanks Matt!!!!) will help relieve from some of such
    issues and perspective gaps that different people have.<br>
    <br>
    <blockquote cite="mid:20160416173317.GL24746@redhat.com" type="cite">
      <blockquote type="cite">
        <blockquote type="cite">I've proposed new addendum to the spec
          to just make this a config, with
          <br>
          a flow of how we'd get rid of it -
          <a class="moz-txt-link-freetext" href="https://review.openstack.org/#/c/306447/">https://review.openstack.org/#/c/306447/</a>
          <br>
        </blockquote>
        <br>
        Having said all of the above, I'm good with your proposal so as
        to keep
        <br>
        the current traction on the work. (Guessing that was the reason
        to delay
        <br>
        glanceclient work before)
        <br>
        <br>
      </blockquote>
      <br>
      I'm happy to see this moving forward. TBH, anything that would let
      glance (and
      <br>
      Nova) move on from Glance's v1 is fine.
      <br>
      <br>
      Flavio
      <br>
      <br>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <br>
      <pre wrap="">__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: <a class="moz-txt-link-abbreviated" href="mailto:OpenStack-dev-request@lists.openstack.org?subject:unsubscribe">OpenStack-dev-request@lists.openstack.org?subject:unsubscribe</a>
<a class="moz-txt-link-freetext" href="http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev">http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev</a>
</pre>
    </blockquote>
    <br>
    <pre class="moz-signature" cols="72">-- 

Thanks,
Nikhil</pre>
  </body>
</html>