Shuquan:
It is great to see the progress being made by your team. Gerry Kopec has already provided guidance to you for line 27 (only a subset of the patch is intended
to be upstreamed). Also Dean provided the unsquashed patches for line 37 yesterday.
One suggestion I have for all the nova changes in your list would be for your primes to have a meeting with Gerry since he is familiar with all of the nova high
priority bugs in your list. Gerry can answer questions you may have about any of the patches and possibly provide suggestions on what to put in the launchpads that you create. I’ll leave it to you to work out a day and time that works best for both your
primes and Gerry.
Frank
From: Jones, Bruce E [mailto:bruce.e.jones@intel.com]
Sent: Wednesday, March 13, 2019 6:36 PM
To: starlingx-discuss@lists.starlingx.io
Subject: [Starlingx-discuss] FW: Distro.openstack meeting notes Mar 12th 2019
Here is an update on 99Cloud’s work on OpenStack patch upstreaming. The master spreadsheet has been updated to reflect this update.
They are requesting help on a couple of the line items:
Line 27 primary: server-groups best-effort and group size: Server group soft-affinity is achieved in pike version, why not to use it but to develop the best-effort?
And what is the server group size used for, I can't think about a necessary scene to finish the spec.
Line 37 Configurable auto page refresh: Can someone (@Dean, @Frank) send the un-squashed patches to Shuquan please?
brucej
From: Shuquan Huang [mailto:huang.shuquan@99cloud.net]
Sent: Wednesday, March 13, 2019 1:59 AM
To: Jones, Bruce E <bruce.e.jones@intel.com>
Cc: 王亚 <wang.ya@99cloud.net>;
朱博祥 <zhu.boxiang@99cloud.net>;
张鲲鹏 <zhang.kunpeng@99cloud.net>
Subject: Re: [Starlingx-discuss] Distro.openstack meeting notes Mar 12th 2019
Hi Bruce,
This is the updated from 99Cloud.
6 |
vCPU model |
Ya Wang |
Because the feature is frozen, this spec has to be deferred to the train release. Last week I submit two patches into review system, the main work was refactor part of libvirt driver’s code
and add a public method in scheduler utils to get flavor traits. Now the review is pause because we have to wait for re-proposed spec merged. The spec has one +2. |
20 |
rbd disks: convert from source format to raw Intel to fix upstram |
Boxiang Zhu |
Last week, Matthew suggested to remove the LVM related changes out of the patch. He thought that it might be easier to split it into a separate follow-up patch,
as this is the change that will require a release note. Therefore, I have split it into two patches. And I also added some unit tests and improved the release notes. Just ready for review. I will ping Matthew and some other nova core reviewers to review the
patch. |
23 |
Do not bypass scheduler when using forced hosts |
Boxiang Zhu |
I have pushed a patch to the review system. But it still need more time to discuss with nova community to find whether it is suitable. |
25 |
Live block migration fails following auth token expiration |
Ya Wang |
The bug has been solved and I’ve verified it. On the other hand, I’ve submit a patch to update the nova’s document. The reason for the previous failure is because
the service user's role does not match the config item: service_token_roles which is not clear in the document. |
26 |
Drop support for forced hosts for live migrate |
Boxiang Zhu |
The bug has been solved and I’ve verified it. The nova community has the same bp like 'remove-force-flag-from-live-migrate-and-evacuate'. The bp adds a micro version(named
as 2.68) of api to remove the force flag. I have test the function by myself. And it will not accept the force flag if I specify the os-compute-api-version is 2.68. So I think the bp completely meets our requirements. |
27 |
Unsquash: primary: server-groups best-effort and group size |
Kupeng Zhang |
This patch contains two features, best-effort server group and group size. The best-effort feature is similar to soft-affinity/soft-anti-affinity of server group
in nova community, and the group size is a new feature which need to submit a spec. We are discussing how to work for them. There are two questions to ask. Server group soft-affinity is achieved in pike version, why not to use it but to develop the best-effort?
And what is the server group size used for, I can't think about a necessary scene to finish the spec. |
37 |
Configurable Auto Page Refresh |
Kupeng Zhang |
I can’t get the patch from the link in the excel. It’s inside the intel github. Could you send this to us? |
On Mar 12, 2019, at 10:24 PM, Shuquan Huang <huang.shuquan@99cloud.net> wrote:
Sorry, I didn’t notice it shifts to PDT. I’ll summary the progress from my side and send it out tomorrow.
On Mar 12, 2019, at 10:21 PM, Jones, Bruce E <bruce.e.jones@intel.com> wrote:
It’s over. All meetings shifted one hour earlier due to the US/Canada time change.
From: Shuquan
Huang [mailto:huang.shuquan@99cloud.net]
Sent: Tuesday, March 12, 2019 7:08 AM
To: Jones, Bruce E <bruce.e.jones@intel.com>
Subject: Re: [Starlingx-discuss] Distro.openstack meeting notes Mar 12th 2019
Hi Bruce,
It seems I can’t join the meeting. It shows “The host has another meeting in progress”
Is the meeting rescheduled?
From: <starlingx-discuss-bounces+huang.shuquan=99cloud.net@lists.starlingx.io>
on behalf of "Jones, Bruce E" <bruce.e.jones@intel.com>
Date: Tuesday, March 12, 2019 at 9:55 PM
To: "starlingx-discuss@lists.starlingx.io" <starlingx-discuss@lists.starlingx.io>
Subject: [Starlingx-discuss] Distro.openstack meeting notes Mar 12th 2019
·
·
·
§
§
·
·
·
·
·
_______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss
<image001.png>
<image002.png>
黄舒泉 Shuquan Huang
技术总监 Technical Director
九州云信息科技有限公司 99Cloud Co. Ltd.
邮箱(Email): huang.shuquan@99cloud.net
手机(Mobile): +86-18616211612
电话(Tel): 021-61207665
地址(Addr): 上海市局门路427号1号楼206
Room 206, Bldg 1, No.427 JuMen Road, ShangHai, China
网址(Site): http://www.99cloud.net