<div dir="auto"><div><br><br><div class="gmail_quote"><div dir="ltr">Le jeu. 13 sept. 2018 à 19:29, Naichuan Sun <<a href="mailto:naichuan.sun@citrix.com">naichuan.sun@citrix.com</a>> a écrit :<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">





<div lang="EN-US" link="blue" vlink="purple">
<div class="m_-5038233586662420261WordSection1">
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">Hi, Sylvain,<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"><u></u> <u></u></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">Thank you very much for the information. It is pity that I can’t attend the meeting.<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">I have a concern about reshaper in multi-type vgpu support.
<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">In the old vgpu support, we only have one vgpu inventory in root resource provider, which means we only support one vgpu type. When do reshape, placement will
 send allocations(which include just one vgpu resource allocation information) to the driver, if the host have more than one pgpu/pgpug(which support different vgpu type), how do we know which pgpu/pgpug own the allocation information? Do we need to communicate
 with hypervisor the confirm that?</span></p></div></div></blockquote></div></div><div dir="auto"><br></div><div dir="auto">The reshape will actually move the existing allocations for a VGPU resource class to the inventory for this class that is on the child resource provider now with the reshape.</div><div dir="auto"><br></div><div dir="auto">Since we agreed on keeping consistent naming, there is no need to guess which is which. That said, you raise a point that was discussed during the PTG and we all agreed there was an upgrade impact as multiple vGPUs shouldn't be allowed until the reshape is done.</div><div dir="auto"><br></div><div dir="auto">Accordingly, see my spec I reproposed for Stein which describes the upgrade impact <a href="https://review.openstack.org/#/c/602474/">https://review.openstack.org/#/c/602474/</a></div><div dir="auto"><br></div><div dir="auto">Since I'm at the PTG, we have huge time difference between you and me, but we can discuss on that point next week when I'm back (my mornings match then your afternoons)</div><div dir="auto"><br></div><div dir="auto">-Sylvain</div><div dir="auto"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div lang="EN-US" link="blue" vlink="purple"><div class="m_-5038233586662420261WordSection1"><p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"><u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"><u></u> <u></u></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">Thank you very much.<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"><u></u> <u></u></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">BR.<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">Naichuan Sun<u></u><u></u></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"><u></u> <u></u></span></p>
<p class="MsoNormal"><b><span style="font-size:11.0pt;font-family:"Calibri",sans-serif">From:</span></b><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"> Sylvain Bauza [mailto:<a href="mailto:sbauza@redhat.com" target="_blank" rel="noreferrer">sbauza@redhat.com</a>]
<br>
<b>Sent:</b> Thursday, September 13, 2018 11:47 PM<br>
<b>To:</b> OpenStack Development Mailing List (not for usage questions) <<a href="mailto:openstack-dev@lists.openstack.org" target="_blank" rel="noreferrer">openstack-dev@lists.openstack.org</a>><br>
<b>Subject:</b> Re: [openstack-dev] About microversion setting to enable nested resource provider<u></u><u></u></span></p>
<p class="MsoNormal"><u></u> <u></u></p>
<div>
<div>
<div>
<p class="MsoNormal">Hey Naichuan, <u></u><u></u></p>
</div>
<div>
<p class="MsoNormal">FWIW, we discussed on the missing pieces for nested resource providers. See the (currently-in-use) etherpad
<a href="https://etherpad.openstack.org/p/nova-ptg-stein" target="_blank" rel="noreferrer">https://etherpad.openstack.org/p/nova-ptg-stein</a> and lookup for "closing the gap on nested resource providers" (L144 while I speak)<u></u><u></u></p>
</div>
<div>
<p class="MsoNormal"><u></u> <u></u></p>
</div>
<div>
<p class="MsoNormal">The fact that we are not able to schedule yet is a critical piece that we said we're going to work on it as soon as we can.<u></u><u></u></p>
</div>
<div>
<p class="MsoNormal"><u></u> <u></u></p>
</div>
<div>
<p class="MsoNormal">-Sylvain<u></u><u></u></p>
</div>
</div>
</div>
<div>
<p class="MsoNormal"><u></u> <u></u></p>
<div>
<p class="MsoNormal">On Thu, Sep 13, 2018 at 9:14 AM, Eric Fried <<a href="mailto:openstack@fried.cc" target="_blank" rel="noreferrer">openstack@fried.cc</a>> wrote:<u></u><u></u></p>
<blockquote style="border:none;border-left:solid #cccccc 1.0pt;padding:0in 0in 0in 6.0pt;margin-left:4.8pt;margin-right:0in">
<p class="MsoNormal">There's a patch series in progress for this:<br>
<br>
<a href="https://review.openstack.org/#/q/topic:use-nested-allocation-candidates" target="_blank" rel="noreferrer">https://review.openstack.org/#/q/topic:use-nested-allocation-candidates</a><br>
<br>
It needs some TLC. I'm sure gibi and tetsuro would welcome some help...<br>
<br>
efried<u></u><u></u></p>
<div>
<div>
<p class="MsoNormal"><br>
On 09/13/2018 08:31 AM, Naichuan Sun wrote:<br>
> Thank you very much, Jay.<br>
> Is there somewhere I could set microversion(some configure file?), Or just modify the source code to set it?<br>
> <br>
> BR.<br>
> Naichuan Sun<br>
> <br>
> -----Original Message-----<br>
> From: Jay Pipes [mailto:<a href="mailto:jaypipes@gmail.com" target="_blank" rel="noreferrer">jaypipes@gmail.com</a>]
<br>
> Sent: Thursday, September 13, 2018 9:19 PM<br>
> To: Naichuan Sun <<a href="mailto:naichuan.sun@citrix.com" target="_blank" rel="noreferrer">naichuan.sun@citrix.com</a>>; OpenStack Development Mailing List (not for usage questions) <<a href="mailto:openstack-dev@lists.openstack.org" target="_blank" rel="noreferrer">openstack-dev@lists.openstack.org</a>><br>
> Cc: melanie witt <<a href="mailto:melwittt@gmail.com" target="_blank" rel="noreferrer">melwittt@gmail.com</a>>; <a href="mailto:efried@us.ibm.com" target="_blank" rel="noreferrer">
efried@us.ibm.com</a>; Sylvain Bauza <<a href="mailto:sbauza@redhat.com" target="_blank" rel="noreferrer">sbauza@redhat.com</a>><br>
> Subject: Re: About microversion setting to enable nested resource provider<br>
> <br>
> On 09/13/2018 06:39 AM, Naichuan Sun wrote:<br>
>> Hi, guys,<br>
>><br>
>> Looks n-rp is disabled by default because microversion matches 1.29 : <br>
>> <a href="https://github.com/openstack/nova/blob/master/nova/api/openstack/place" target="_blank" rel="noreferrer">
https://github.com/openstack/nova/blob/master/nova/api/openstack/place</a><br>
>> ment/handlers/allocation_candidate.py#L252<br>
>><br>
>> Anyone know how to set the microversion to enable n-rp in placement?<br>
> <br>
> It is the client which must send the 1.29+ placement API microversion header to indicate to the placement API server that the client wants to receive nested provider information in the allocation candidates response.<br>
> <br>
> Currently, nova-scheduler calls the scheduler reportclient's<br>
> get_allocation_candidates() method:<br>
> <br>
> <a href="https://github.com/openstack/nova/blob/0ba34a818414823eda5e693dc2127a534410b5df/nova/scheduler/manager.py#L138" target="_blank" rel="noreferrer">
https://github.com/openstack/nova/blob/0ba34a818414823eda5e693dc2127a534410b5df/nova/scheduler/manager.py#L138</a><br>
> <br>
> The scheduler reportclient's get_allocation_candidates() method currently passes the 1.25 placement API microversion header:<br>
> <br>
> <a href="https://github.com/openstack/nova/blob/0ba34a818414823eda5e693dc2127a534410b5df/nova/scheduler/client/report.py#L353" target="_blank" rel="noreferrer">
https://github.com/openstack/nova/blob/0ba34a818414823eda5e693dc2127a534410b5df/nova/scheduler/client/report.py#L353</a><br>
> <br>
> <a href="https://github.com/openstack/nova/blob/0ba34a818414823eda5e693dc2127a534410b5df/nova/scheduler/client/report.py#L53" target="_blank" rel="noreferrer">
https://github.com/openstack/nova/blob/0ba34a818414823eda5e693dc2127a534410b5df/nova/scheduler/client/report.py#L53</a><br>
> <br>
> In order to get the nested information returned in the allocation candidates response, that would need to be upped to 1.29.<br>
> <br>
> Best,<br>
> -jay<u></u><u></u></p>
</div>
</div>
<p class="MsoNormal">> __________________________________________________________________________<br>
> OpenStack Development Mailing List (not for usage questions)<br>
> Unsubscribe: <a href="http://OpenStack-dev-request@lists.openstack.org?subject:unsubscribe" target="_blank" rel="noreferrer">
OpenStack-dev-request@lists.openstack.org?subject:unsubscribe</a><br>
> <a href="http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev" target="_blank" rel="noreferrer">
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev</a><br>
> <br>
<br>
__________________________________________________________________________<br>
OpenStack Development Mailing List (not for usage questions)<br>
Unsubscribe: <a href="http://OpenStack-dev-request@lists.openstack.org?subject:unsubscribe" target="_blank" rel="noreferrer">
OpenStack-dev-request@lists.openstack.org?subject:unsubscribe</a><br>
<a href="http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev" target="_blank" rel="noreferrer">http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev</a><u></u><u></u></p>
</blockquote>
</div>
<p class="MsoNormal"><u></u> <u></u></p>
</div>
</div>
</div>

__________________________________________________________________________<br>
OpenStack Development Mailing List (not for usage questions)<br>
Unsubscribe: <a href="http://OpenStack-dev-request@lists.openstack.org?subject:unsubscribe" rel="noreferrer noreferrer" target="_blank">OpenStack-dev-request@lists.openstack.org?subject:unsubscribe</a><br>
<a href="http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev" rel="noreferrer noreferrer" target="_blank">http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev</a><br>
</blockquote></div></div></div>