Tuesday, 2026-11-03

« Monday, 2026-11-02 Index Wednesday, 2026-11-04 »
opendevreviewchandan kumar proposed openstack/cyborg-specs master: Repropose NVMe secure cleanup spec for 2027.1  https://review.opendev.org/c/openstack/cyborg-specs/+/100571606:34
opendevreviewchandan kumar proposed openstack/cyborg-tempest-plugin master: Add per-driver Cyborg-to-Placement device test  https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/99221009:50
opendevreviewchandan kumar proposed openstack/cyborg master: Validate device profile traits against os-traits library  https://review.opendev.org/c/openstack/cyborg/+/99794510:03
opendevreviewchandan kumar proposed openstack/cyborg-tempest-plugin master: Add per-driver Cyborg-to-Placement device test  https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/99221010:17
jgilaber#starmeeting cyborg14:00
jgilaber#startmeeting cyborg14:01
opendevmeetMeeting started Tue Nov  3 14:01:07 2026 UTC and is due to finish in 60 minutes.  The chair is jgilaber. Information about MeetBot at http://wiki.debian.org/MeetBot.14:01
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:01
opendevmeetThe meeting name has been set to 'cyborg'14:01
jgilaberHi! Who is around today?14:01
jgilabercourtesy ping sean-k-mooney amoralej bogdando rlandy chandankumar skovili melwit14:02
rlandyo/14:02
sean-k-mooneyo/14:02
jgilaberfeel free to add topics to the agenda https://etherpad.opendev.org/p/openstack-cyborg-irc-meeting#L5214:04
jgilaberlet's get started14:04
jgilaber#topic 2025.1 to unmaintained14:04
jgilaberjust a quick announcement14:05
jgilaberthe patch moving the stable/2025.1 branch to unmaintained merged on Friday14:05
jgilaber#link https://review.opendev.org/c/openstack/releases/+/101093114:05
sean-k-mooney+114:07
sean-k-mooneythansk for taking care of that and the other releases14:07
jgilaberno problem14:07
jgilaberI also created a patch to drop the 2025.1 jobs from the tempest plugin14:07
jgilaber#link https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/101301514:07
jgilaberand once that is merged, I think that we're set14:08
sean-k-mooneyoh we can actully do that now cool14:09
jgilabernothing else from side, does anyone have any question/comment?14:09
sean-k-mooney+2 can we push a dnm to the stabel branches to confirm 14:10
sean-k-mooneythat depens on this14:10
sean-k-mooneyactully we shoudl have no refences to the remvoed jobs on stable right14:10
sean-k-mooney2025.1 was the last one to refence it14:10
sean-k-mooneythat is basiclly the last check i want to do before +w14:10
jgilaberchecking now14:12
jgilaberhttps://opendev.org/openstack/cyborg/src/branch/stable/2025.2/.zuul.yaml14:12
jgilaberwe use the parent of the job we're deleting on the stable branches14:12
jgilaberbut I have DNM patches like https://review.opendev.org/c/openstack/cyborg/+/996306 that I can updated to double check14:12
sean-k-mooneyack if you would not mind doing that we can likely merget his later today14:13
sean-k-mooneyi think we are good14:13
jgilabersure14:13
sean-k-mooneybut better safe then sorry14:13
opendevreviewJoan Gilabert proposed openstack/cyborg stable/2025.2: [DNM] Test CI changes  https://review.opendev.org/c/openstack/cyborg/+/99630614:13
jgilaberI'll ping in channel once they run14:14
jgilaberso I think we can move on14:14
opendevreviewJoan Gilabert proposed openstack/cyborg stable/2026.1: [DNM] Test CI changes  https://review.opendev.org/c/openstack/cyborg/+/99630714:14
melwitto/14:14
jgilaber#topic Reviews14:15
jgilaberI have one tempest plugin patch14:15
jgilaber#link https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/100804314:15
jgilaberit adds an api test for https://review.opendev.org/c/openstack/cyborg/+/1008027 which we already merged14:15
jgilaberand backported to all the stable branches14:15
sean-k-mooneyso in general we are not ment to have negitive api tests14:17
sean-k-mooneyso im nod sure this is really required14:17
sean-k-mooneyi gues this si fine14:18
sean-k-mooneybut i would nt expect use to do this every time we fix a validation bug14:18
sean-k-mooneyhttps://docs.openstack.org/tempest/latest/HACKING.html#negative-tests14:19
jgilaberit somewhat overlapping with the unit test, but since we have a file in the api test for negative device profile tests I've added it14:19
sean-k-mooneyand https://docs.openstack.org/tempest/latest/field_guide/api.html14:19
sean-k-mooneyare the relevent secionts14:19
sean-k-mooneythe api guid says They should send not only good data, but bad data at the API and look for error codes.14:19
sean-k-mooneybut the general guide discuages there use14:20
jgilaberyeah, the guide is quite clear, I'm ok dropping the change14:21
sean-k-mooneyya si the idea is that the api sample test that mel added14:21
sean-k-mooneywould be hte way we test this fucntionall14:21
sean-k-mooneybut lets proceed wit this for now since its cheap and written14:21
sean-k-mooneyi just woudl nto default to adding test like this for every bug fix14:21
jgilaber+114:22
jgilabernoted14:22
sean-k-mooneymelwitt: any comment or concern with that? i see chandan is already +114:22
opendevreviewJoan Gilabert proposed openstack/cyborg stable/2026.2: [DNM] Test CI changes  https://review.opendev.org/c/openstack/cyborg/+/101389514:22
sean-k-mooneyill +2 it for now and ill loop back to it in a day or so14:23
melwittsounds ok to me14:23
sean-k-mooneyi think we can move on and comemnt asyc if there is a concern14:23
sean-k-mooneyok well if there is not disagrement i can jsut approve14:23
sean-k-mooneyok ill do that14:24
jgilaberthanks sean-k-mooney 14:24
jgilaberdoes anyone have any patch to highlight for reviews?14:24
jgilaberotherwise we can move to bug triage14:24
sean-k-mooneyam just that 14:25
sean-k-mooneyi did not have time to update the maintainer guide14:25
sean-k-mooneyso i will try and do that before the next meeting14:25
sean-k-mooneyi hope we can come to a concenos on it by then and we can merge and start makeign the required acl changes in gerrit14:26
sean-k-mooneythats all i had for that14:26
jgilaberack, sounds good, thanks14:26
jgilaber#topic Bugs14:27
jgilaberwe have one bug to triage14:27
jgilaber#link https://bugs.launchpad.net/openstack-cyborg/+bug/217244714:27
jgilaberthis came up during a review IIRC14:27
jgilaberit's valid and quite severe to me14:28
jgilaberI haven't checked however if there are more cases like that through the codebases where we're passing around db results14:28
jgilaberwithouth constructing an object from them14:28
sean-k-mooneyoh yes the db row14:29
sean-k-mooneyit is but fixign it will not be straight forward.14:29
sean-k-mooneyit will need an rpc bump14:29
sean-k-mooneywhich means we need the upgrade_levels supprot first14:30
sean-k-mooneywhich means i dont ink we can fix this in a backportable way14:30
sean-k-mooneymelwitt: i dont hink you saw that review but tl;dr is we are shoving a raw dict (a db row) over the rpc bus when programing an fpga14:31
jgilaberwhy do we need an RPC bump? don't the current ovs have the required fields already?14:31
sean-k-mooneyinstead of an ovo14:31
melwittyeah was just reading through the bug report :/ 14:31
sean-k-mooneyjgilaber: because changing the type of the parmater form a dict to an ovo will break any compute-agnet that recives it if they are not upgraded14:31
sean-k-mooneyand no in this case i think we are missign the ovo14:32
sean-k-mooneybut i woudl have to check in more detail. we may be able to use the driver version of the ovo14:32
sean-k-mooneybtu i dont think there is a toplevel ControlpathID ovo14:33
jgilabermmh right, even though it's wrong, we've established the dict in the contract now14:33
sean-k-mooneywe might get lucky14:33
jgilaberhttps://github.com/openstack/cyborg/blob/8f58f18b90de98a047dabc93a7c5f41b0a4ae656/cyborg/objects/control_path.py#L3014:33
sean-k-mooneyi think we are usign the dict comaptiablity mixin for the ovo14:33
sean-k-mooneybut no that woudl not help14:34
sean-k-mooneyin the serialsised form the stucre is diffent14:34
sean-k-mooneyovo add metadata about there type when they are converted to json14:34
jgilaberok, so we set it to triaged/high?14:39
sean-k-mooneyya i think so14:39
sean-k-mooneywe will want to take oru time to think about the best way forward14:39
sean-k-mooneybut its imporant14:39
sean-k-mooneyso high works for me14:40
melwitt+114:40
jgilaberok, I did just ask claude and it claims that this is the only occurrence of this in the codebase14:40
jgilaberwe'll need to double check that, but at least is not a widespread problem14:40
sean-k-mooneywell that at least promising14:40
sean-k-mooneywe may still need to auit the rest but that something14:40
sean-k-mooneyi dont liek this option14:41
sean-k-mooneybut we also dont offically supprot rooling upgrades yet14:41
sean-k-mooneyso we coudl jsut "fix" it a breakign way14:41
sean-k-mooneybut to me that is a last resort14:41
sean-k-mooneyif we have the rpc bump we woudl jsut send the dict as we do today for 1.0 and the object in 1.x14:42
sean-k-mooneythat is the correct way to do it14:42
jgilaberyes, let's try to do it properly14:42
sean-k-mooneyack ok we can move on for now i think14:43
jgilabersure14:43
jgilaberthat was the last thing in the agenda14:43
jgilaberany last minute topic?14:43
jgilaberI'll take that as a no :)14:44
jgilaberlast thing 14:45
jgilaber#topic Volunteers to chair next meeting14:45
jgilaberany volunteer?14:45
jgilaberif not I can chair again next week14:45
jgilaberok, let's end the meeting here, thanks everyone!14:46
jgilaber#endmeeting14:46
opendevmeetMeeting ended Tue Nov  3 14:46:29 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)14:46
opendevmeetMinutes:        https://meetings.opendev.org/meetings/cyborg/2026/cyborg.2026-11-03-14.01.html14:46
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/cyborg/2026/cyborg.2026-11-03-14.01.txt14:46
opendevmeetLog:            https://meetings.opendev.org/meetings/cyborg/2026/cyborg.2026-11-03-14.01.log.html14:46
chandank`sean-k-mooney: jgilaber Hello, I have updated the nvme spec based on ovo indirection spec https://review.opendev.org/c/openstack/cyborg-specs/+/1005716 , Feel free to add it to your review queue, thank you.15:02
sean-k-mooneyack i saw it in my inbox but have not looked yet15:06
chandank`thank you!15:14
jgilabersean-k-mooney, I went through claude findings about the rpc calls and other than the know one, it only flagged https://github.com/openstack/cyborg/blob/8f58f18b90de98a047dabc93a7c5f41b0a4ae656/cyborg/api/controllers/v2/arqs.py#L44015:21
jgilaberit's passing dicts, which we could serialize into strings15:21
sean-k-mooneywe shoudl only pass scalare types or ovo15:22
sean-k-mooneyno collectoins15:22
sean-k-mooney serisaling them as stings does not help15:22
sean-k-mooney so that is another violation btu this tiem in the conductor rpc api rather then in the agent rpc api15:24
sean-k-mooneythis particallar opertion is one the api coudl have done direcly iself15:24
sean-k-mooneyit did not need a rpc call at all15:25
sean-k-mooneyso we rpaobly shoudl jsut revert https://github.com/openstack/cyborg/commit/1a45115e2b3bf2a1a9c6162256329c2eece518ae15:25
sean-k-mooneybut i need to look into this in more detail to verify that15:25
melwittsean-k-mooney: if it doesn't need a RPC call at all, then reverting that could easily fix it and avoid any need for versioning changes? sounds too good to be true but would be great if it somehow is the case16:23
sean-k-mooneywell teh api has direct db access16:29
sean-k-mooneyand the conducot methoed at least at the time the patch was writen16:29
sean-k-mooneyonly called the object memthod to make the changes16:29
sean-k-mooneywhat im unsuer about is has that rpc grown other actions on teh conductor side since this change was made16:30
sean-k-mooneyi have not checked that16:30
melwittack16:50
opendevreviewmelanie witt proposed openstack/cyborg master: Replace python-glanceclient with openstacksdk  https://review.opendev.org/c/openstack/cyborg/+/101337818:12
opendevreviewMerged openstack/cyborg-tempest-plugin master: device_profile: test that group with no resources: key is rejected  https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/100804318:33
« Monday, 2026-11-02 Index Wednesday, 2026-11-04 »

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!