Open Stack

>> For me from the sharing disk provider feature perspective the placement
>> aggregate that is needed for the sharing to work, and any kind of nova
>> host aggregate (either synced to placement or not) is independent. The
>> placement aggregate is a must for the feature. On top of that if the
>> operator wants to create a nova host aggregate as well and sync it to
>> placement then at the end there will be two, independent placement
>> aggregates. One to express the sharing relationship and one to express
>> a host aggregate from nova. These two aggregate will not be the same as
>> the first one will have the sharing provider in it while the second one
>> doesn't.

> I tend to agree with the simplicity of this as well.

I have updated the specs as per the agreements at Shanghai PTG.
I still do see some thorny issues especially the way disk_gb information will be returned in the new micro-version os-hypervisor API. Please check "other end user impact" and "Response of os-hypervisors statistics" sections from the specs.

Please review the specs [1] and give your suggestions/feedback.

[1] : https://review.opendev.org/#/c/650188/8

Thanks,
tpatil

________________________________________
From: Matt Riedemann <mriedemos at gmail.com>
Sent: Thursday, November 14, 2019 11:02 PM
To: openstack-discuss at lists.openstack.org
Subject: Re: [nova][ptg] Allow compute nodes to use DISK_GB from shared storage RP by using aggregate relationship

On 11/14/2019 1:45 AM, Balázs Gibizer wrote:
> For me from the sharing disk provider feature perspective the placement
> aggregate that is needed for the sharing to work, and any kind of nova
> host aggregate (either synced to placement or not) is independent. The
> placement aggregate is a must for the feature. On top of that if the
> operator wants to create a nova host aggregate as well and sync it to
> placement then at the end there will be two, independent placement
> aggregates. One to express the sharing relationship and one to express
> a host aggregate from nova. These two aggregate will not be the same as
> the first one will have the sharing provider in it while the second one
> doesn't.

I tend to agree with the simplicity of this as well.

--

Thanks,

Matt

Disclaimer: This email and any attachments are sent in strictest confidence for the sole use of the addressee and may contain legally privileged, confidential, and proprietary data. If you are not the intended recipient, please advise the sender by replying promptly to this email and then delete and destroy this email and any attachments without any further use, copying or forwarding.