Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2025-02-04
16:40:46 r-taketn ok, i just added topic today's agenda.
16:40:56 r-taketn Request for replies on the following - About RDT/MPAM feature
16:41:04 r-taketn #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/HIKLBMNMTJR6GFPKFKODECMR56SYPT5J/
16:41:07 bauzas oh sorry
16:41:25 bauzas I removed it because I was thinking it was the topic we discussed last week
16:41:29 bauzas but indeed, this is different
16:42:00 r-taketn Yes. I have another topic to mention from last week.
16:42:08 sean-k-mooney r-taketn: the best way to move forward with that is to start a spec and bring it up to dicuss in the ptg
16:42:08 r-taketn I'm considering the feasibility of MPAM/RDT feature within OpenStack to solve noisy neighbor problems on VM clouds.
16:42:50 sean-k-mooney r-taketn: depending on how its mature it may be something that can be supproted
16:42:54 bauzas agreed
16:42:59 sean-k-mooney its a very invasive change
16:43:27 sean-k-mooney which is partly why the previous efforts were abandonded
16:44:18 r-taketn I intended to discuss the feasibility of this before submitting spec because its implementation will not be simple
16:44:18 bauzas r-taketn: the PTG dates are now known
16:44:33 bauzas r-taketn: would you be able to join us remotely ?
16:45:09 r-taketn sure.
16:45:20 bauzas cool thanks
16:45:29 bauzas then let's discuss this at the PTG
16:45:39 bauzas I need to create an etherpad for tracking those topics
16:45:43 sean-k-mooney r-taketn i was off on firday and yesterday so i have not fully read your latest post on this topic
16:45:53 sean-k-mooney ill see if i can do that later today
16:46:12 r-taketn sean-k-mooney: Thanks
16:46:58 r-taketn It's for confirming. Should I submit a spec before the PTG?
16:47:07 bauzas it would be nice
16:47:17 bauzas fwiw, I just created a PTG etherpad
16:47:20 sean-k-mooney it does not hurt to do so, it give use somethign concrete to dicuss
16:47:22 bauzas #link https://etherpad.opendev.org/p/nova-2025.2-ptg
16:47:38 bauzas r-taketn: could you please add your two topics at the bottom of that etherpad ?
16:47:49 r-taketn ok, i'll prepare
16:47:58 r-taketn bauzas: ok i'll add
16:48:00 bauzas the one we discuss last week and the other one from this week
16:48:09 bauzas r-taketn: thanks
16:48:17 bauzas I'll also add a couple of topics too
16:48:27 bauzas and I'll tell about the PTG etherpad next week
16:49:24 bauzas are we done for today?
16:49:27 masahito is it okay to start adding some topics on the etherpad?
16:49:34 MengyangZhang[m] I have a question. I was wondering if it is still possible to request an exception to add this bp https://blueprints.launchpad.net/openstack/?searchtext=add-burst-length-support to Eproxy. The code change is minor and is ready for review https://review.opendev.org/c/openstack/nova/+/939490
16:49:37 masahito it's quick question.
16:49:46 sean-k-mooney r-taketn: https://review.opendev.org/c/openstack/nova-specs/+/662264 was the spec i wrote for RDT 5 years ago just an fyi.
16:50:03 bauzas masahito: yes indeed, you can add topics
16:50:14 masahito thanks.
16:50:14 sean-k-mooney MengyangZhang[m]: its very very late to do that
16:50:34 bauzas MengyangZhang[m]: yeah unfortunately, we're past approval freeze from 3 weeks now
16:50:42 bauzas and feature freeze is in 3 weeks too
16:51:18 bauzas the main reason why we have that approval freeze is for making sure we have enough capacity to review all the already approved ones
16:51:55 MengyangZhang[m] understood sorry i was away for two weeks, so i will need to resubmit it for 2025-02?
16:51:57 bauzas that wouldn't be fair to unprioritize some reviews by looking at some other patch, whatever big it is
16:52:06 r-taketn sean-k-mooney: thanks, and sorry, i have not checked it.
16:52:17 bauzas MengyangZhang[m]: unfortunately yes
16:52:36 MengyangZhang[m] okk will do
16:53:00 bauzas thanks
16:53:04 bauzas I guess we're done then
16:53:06 bauzas thanks all
16:53:09 MengyangZhang[m] do i need to re-attend nova meeting once that's updated?
16:53:09 sean-k-mooney MengyangZhang[m]: the folder is already created and the specs for next cycle are open for review, so this could be done early next cycle
16:53:53 sean-k-mooney not nessisarly we can just review it once its posted
16:54:02 MengyangZhang[m] okk thanks!
16:54:29 bauzas MengyangZhang[m]: you don't need to attend the meeting in person, but please add it in our agenda for asking for a specless blueprint when you want
16:55:03 MengyangZhang[m] ok do you have a link to the agenda?
16:55:26 bauzas https://wiki.openstack.org/wiki/Meetings/Nova#Agenda_for_next_meeting
16:55:39 MengyangZhang[m] thanks!
16:55:42 bauzas just add it in the open discussion and request for a specless blueprint approval
16:56:56 bauzas see the documentation here for trivial features approval https://specs.openstack.org/openstack/nova-specs/readme.html#trivial-specifications
16:57:30 bauzas or ping me whenever you want
16:57:38 bauzas anyway, we're almost at time
16:57:43 MengyangZhang[m] thanks!
16:57:44 bauzas anything anyone ?
16:58:08 bauzas then thanks all
16:58:12 opendevmeet Log: https://meetings.opendev.org/meetings/nova/2025/nova.2025-02-04-16.03.log.html
16:58:12 opendevmeet Minutes (text): https://meetings.opendev.org/meetings/nova/2025/nova.2025-02-04-16.03.txt
16:58:12 opendevmeet Minutes: https://meetings.opendev.org/meetings/nova/2025/nova.2025-02-04-16.03.html
16:58:12 opendevmeet Meeting ended Tue Feb 4 16:58:12 2025 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)
16:58:12 bauzas #endmeeting
16:58:22 r-taketn Thank you for discussing.
16:58:26 elodilles thanks o/
16:58:31 gibi thanks forlks
16:58:36 gibi and folks even :)
16:58:41 fwiesel Thanks
16:58:43 masahito thanks everyone
17:00:03 sean-k-mooney MengyangZhang[m]: i suspect the current impletion is missing some fucntionatliy to ensure the compute node supprot the feature when we have a mix os openstack version on different compute nodes by the way
17:00:16 sean-k-mooney MengyangZhang[m]:the existin spec does not cover that properlys.
17:00:28 sean-k-mooney ill leave a comment to that effect on teh current one.
17:02:07 ratailor sean-k-mooney, gibi I have to leave now, its too late for me. I will update the patch tomorrow and we can discuss then.
17:02:12 masahito gibi sean-k-mooney: if you have time, please review this bug fix again. The new patchset returns null at list API in both no pinned az case and no request spec case. https://review.opendev.org/c/openstack/nova/+/939658
17:02:22 gibi ratailor: sure. I also don't have time right now to load context there
17:02:36 ratailor gibi, sean-k-mooney ack. Thanks!
17:02:50 sean-k-mooney gibi: ratailor i suggested just adding additional_properties=True
17:03:17 gibi sean-k-mooney: I think stephenfin was against that
17:03:42 sean-k-mooney if we cna restrict that only a filed called properties in the image object then that woudl be better
17:03:44 gibi and I also would like to be at least specific on the level of string keys, with length check
17:03:47 gibi and charset check
17:04:13 sean-k-mooney we could do that
17:04:24 sean-k-mooney what i want to avoid is listing the speicic image properties
17:04:56 sean-k-mooney specicicly becasue adding a new image property is not an api change today so we should not encode the set of options in the respocne schema
17:06:03 sean-k-mooney but im ok with encodeign some general restrictions. image properties are not free from so its less problatic then other files but we can still do some level fo validation
17:06:49 gibi yeah, lets try to be strict about having a field "properties" and that it is an object with string key values with some length and charset requirements
17:07:00 gibi I think we have example of such schema pieces already
17:07:16 sean-k-mooney ack probaly server metadata?
17:09:33 gibi yeah that is a good candidate

Earlier   Later