Tuesday, 2026-08-25

opendevreviewchandan kumar proposed openstack/cyborg master: Add device_state bind guard  https://review.opendev.org/c/openstack/cyborg/+/99994103:24
opendevreviewchandan kumar proposed openstack/cyborg master: Expose device_state in the device API and add POST /clean endpoint  https://review.opendev.org/c/openstack/cyborg/+/99994203:24
opendevreviewchandan kumar proposed openstack/cyborg master: Deprecate SSD drivers in favour of NVMe driver  https://review.opendev.org/c/openstack/cyborg/+/99994303:24
opendevreviewchandan kumar proposed openstack/cyborg master: Add NVMe driver admin guide and API reference for microversion 2.4  https://review.opendev.org/c/openstack/cyborg/+/99994403:24
opendevreviewchandan kumar proposed openstack/cyborg master: Add per-PF vf_personality sysfs interface  https://review.opendev.org/c/openstack/cyborg/+/99860303:24
opendevreviewchandan kumar proposed openstack/cyborg master: pci-sim: introduce personality ops-table dispatch for VF device types  https://review.opendev.org/c/openstack/cyborg/+/99912503:24
opendevreviewchandan kumar proposed openstack/cyborg master: pci-sim: add NVMe host kernel module  https://review.opendev.org/c/openstack/cyborg/+/99992803:25
opendevreviewchandan kumar proposed openstack/cyborg master: pci-sim: add NVMe VFIO guest probe and host handoff  https://review.opendev.org/c/openstack/cyborg/+/99992903:25
opendevreviewchandan kumar proposed openstack/cyborg master: Add NVMe driver devstack plugin support  https://review.opendev.org/c/openstack/cyborg/+/99994503:25
opendevreviewchandan kumar proposed openstack/cyborg master: Route ARQ unbind through conductor-to-agent cleanup RPCs  https://review.opendev.org/c/openstack/cyborg/+/99993905:21
opendevreviewchandan kumar proposed openstack/cyborg master: nvme: implement NVMeDriver.cleanup()  https://review.opendev.org/c/openstack/cyborg/+/99994005:21
opendevreviewchandan kumar proposed openstack/cyborg master: Add device_state bind guard  https://review.opendev.org/c/openstack/cyborg/+/99994105:21
opendevreviewchandan kumar proposed openstack/cyborg master: Add device_state bind guard and fix fake driver topology  https://review.opendev.org/c/openstack/cyborg/+/99994106:42
opendevreviewchandan kumar proposed openstack/cyborg master: Expose device_state in the device API and add POST /clean endpoint  https://review.opendev.org/c/openstack/cyborg/+/99994206:42
opendevreviewchandan kumar proposed openstack/cyborg master: Deprecate SSD drivers in favour of NVMe driver  https://review.opendev.org/c/openstack/cyborg/+/99994306:42
opendevreviewchandan kumar proposed openstack/cyborg master: Add NVMe driver admin guide and API reference for microversion 2.4  https://review.opendev.org/c/openstack/cyborg/+/99994406:43
opendevreviewchandan kumar proposed openstack/cyborg master: Add per-PF vf_personality sysfs interface  https://review.opendev.org/c/openstack/cyborg/+/99860306:43
opendevreviewchandan kumar proposed openstack/cyborg master: pci-sim: introduce personality ops-table dispatch for VF device types  https://review.opendev.org/c/openstack/cyborg/+/99912506:43
opendevreviewchandan kumar proposed openstack/cyborg master: pci-sim: add NVMe host kernel module  https://review.opendev.org/c/openstack/cyborg/+/99992806:44
opendevreviewchandan kumar proposed openstack/cyborg master: pci-sim: add NVMe VFIO guest probe and host handoff  https://review.opendev.org/c/openstack/cyborg/+/99992906:44
opendevreviewchandan kumar proposed openstack/cyborg master: Add NVMe driver devstack plugin support  https://review.opendev.org/c/openstack/cyborg/+/99994506:44
opendevreviewchandan kumar proposed openstack/cyborg master: Add device_state bind guard and fix fake driver topology  https://review.opendev.org/c/openstack/cyborg/+/99994107:55
opendevreviewchandan kumar proposed openstack/cyborg master: Expose device_state in the device API and add POST /clean endpoint  https://review.opendev.org/c/openstack/cyborg/+/99994208:56
opendevreviewchandan kumar proposed openstack/cyborg master: Deprecate SSD drivers in favour of NVMe driver  https://review.opendev.org/c/openstack/cyborg/+/99994308:56
opendevreviewchandan kumar proposed openstack/cyborg master: Add NVMe driver admin guide and API reference for microversion 2.4  https://review.opendev.org/c/openstack/cyborg/+/99994408:57
opendevreviewchandan kumar proposed openstack/cyborg master: Add per-PF vf_personality sysfs interface  https://review.opendev.org/c/openstack/cyborg/+/99860308:57
opendevreviewchandan kumar proposed openstack/cyborg master: pci-sim: introduce personality ops-table dispatch for VF device types  https://review.opendev.org/c/openstack/cyborg/+/99912508:57
opendevreviewchandan kumar proposed openstack/cyborg master: pci-sim: add NVMe host kernel module  https://review.opendev.org/c/openstack/cyborg/+/99992808:57
opendevreviewchandan kumar proposed openstack/cyborg master: pci-sim: add NVMe VFIO guest probe and host handoff  https://review.opendev.org/c/openstack/cyborg/+/99992908:58
opendevreviewchandan kumar proposed openstack/cyborg master: Add NVMe driver devstack plugin support  https://review.opendev.org/c/openstack/cyborg/+/99994508:58
opendevreviewsean mooney proposed openstack/cyborg master: api: Expose PCI managed mode in ARQ 2.4  https://review.opendev.org/c/openstack/cyborg/+/99457510:49
opendevreviewsean mooney proposed openstack/cyborg master: doc: Add generic PCI driver guide  https://review.opendev.org/c/openstack/cyborg/+/99457610:49
opendevreviewsean mooney proposed openstack/cyborg master: devstack: Configure pci-sim Cyborg PCI testing  https://review.opendev.org/c/openstack/cyborg/+/99457710:49
opendevreviewsean mooney proposed openstack/cyborg master: devstack: Add NVIDIA vGPU variant-driver setup helpers  https://review.opendev.org/c/openstack/cyborg/+/99457810:49
opendevreviewsean mooney proposed openstack/cyborg master: conductor: Retain missing allocated devices  https://review.opendev.org/c/openstack/cyborg/+/100225310:49
opendevreviewsean mooney proposed openstack/cyborg master: objects: Add conductor indirection for OVOs  https://review.opendev.org/c/openstack/cyborg/+/100225410:49
opendevreviewsean mooney proposed openstack/cyborg master: objects: Add remote accelerator inventory access  https://review.opendev.org/c/openstack/cyborg/+/100225510:49
opendevreviewsean mooney proposed openstack/cyborg master: agent: Reconcile PCI attach handle metadata  https://review.opendev.org/c/openstack/cyborg/+/100225610:49
opendevreviewsean mooney proposed openstack/cyborg master: pci: Report managed attach metadata  https://review.opendev.org/c/openstack/cyborg/+/100225710:49
chandankumarHello, Cyborg team meeting is in another 58 mins, Add your topics here: https://etherpad.opendev.org/p/openstack-cyborg-irc-meeting#L4513:01
chandankumarHello, time to for team meeting14:01
chandankumar#startmeeting cyborg14:01
opendevmeetMeeting started Tue Aug 25 14:01:34 2026 UTC and is due to finish in 60 minutes.  The chair is chandankumar. 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
chandankumarwho all are around here?14:01
chandankumaro/14:01
gamioo/14:01
melwitto/14:02
chandankumar#courtesy ping rlandy sean-k-mooney bogdando[m] 14:02
sean-k-mooneyo/14:02
rlandyo/14:03
chandankumarWe are just getting started, Feel free to add your topics here https://etherpad.opendev.org/p/openstack-cyborg-irc-meeting#L4514:03
chandankumar#link https://etherpad.opendev.org/p/openstack-cyborg-irc-meeting#L4514:03
chandankumarLet me get started with the first topic14:03
chandankumar#topic Feature Freeze week14:04
chandankumarI have added this topic14:04
chandankumarAug 24- 28 is milestone 3 and feature freeze is happening14:04
chandankumarrelease team proposed to do final release of client lib14:04
chandankumar1002056: Release final python-cyborgclient for 2026.2 Hibiscus | https://review.opendev.org/c/openstack/releases/+/100205614:04
chandankumar#link 1002056: Release final python-cyborgclient for 2026.2 Hibiscus | https://review.opendev.org/c/openstack/releases/+/100205614:05
chandankumarI checked the open patches for cyborgclient https://review.opendev.org/q/project:openstack/python-cyborgclient+status:open14:05
chandankumarnothing seems to be pending to be merged there.14:05
chandankumarFeel free to take a look at the release patch and add your votes there14:06
sean-k-mooneyya the client has been pretty quite14:06
sean-k-mooneywe obvioly cleaned it up at the start of the cycle14:07
sean-k-mooneybut i think your corect that we dont expect to merge much else this cycle14:07
melwittgo us14:07
sean-k-mooneythanks for reviwing the release patch14:08
chandankumarYes, there is one patch from melwitt https://review.opendev.org/c/openstack/python-cyborgclient/+/995877 which I think will go in openstackclient14:08
chandankumarbut we can talk about that later on14:08
chandankumarMoving ahead14:09
chandankumarSince this week soft string freeze is also going on14:09
chandankumarI wanted to highlight the status of 2026.2 approved spec implementations14:10
chandankumarFor mdev, cyborg-vgpu-support on nova side https://review.opendev.org/q/topic:%22cyborg-vgpu-support%22 is merged now14:10
chandankumarRest of the mdev generic driver is waiting for refresh14:10
chandankumarfor vfio-variant drivers supprot for pci driver, sean-k-mooney is already working on refreshing the patches14:11
chandankumarhttps://review.opendev.org/q/topic:%22bp/support-vfio-variant-driver-managed-mode%2214:12
sean-k-mooneyyes i did a refersh yesterday but i was not happy with on aspect14:12
sean-k-mooneyspecificly how the condutor was responibel for validating if we were tryign to change the managed mode on an allcoated driver14:12
sean-k-mooneyso i pivoted the approch last night14:12
sean-k-mooneythe new approch start by fixing https://bugs.launchpad.net/openstack-cyborg/+bug/201751314:13
sean-k-mooneyand then build out the initall supprot for ovo remotable methods14:13
sean-k-mooneyand then move the check to the cybrog-agent where it shoudl alwasy have been14:13
sean-k-mooneyhowever that is a non trivial change in scope14:13
sean-k-mooneysoe im not sure if we will still complet that this cycle14:14
sean-k-mooneyi pushed the inital version fo that this somrining14:14
sean-k-mooneybut i need ot reviwe it in more detial14:14
sean-k-mooneythere was too much technical debt in the perior approch for me to be comfortabel merging it hence the pivot14:14
sean-k-mooneyi have also broken the patch up a bit as part of this to make it a bit easier to review but i think im going to focus on advancing the nvme work14:15
melwitt++14:16
chandankumarok, thank you for sharing the background and pairing up the tech debt also14:17
chandankumarI will go over them14:17
chandankumarComing to nvme14:17
chandankumarthank you sean-k-mooney for reviewing nvme series, there is a lot of code movement in between different patches,14:17
chandankumarNow the CI is passing 14:17
chandankumarBut I need to resolve the comments, I will do that tomorrow14:18
sean-k-mooneyack im currenly in the process of preparing a devstack to test it with the pci sim end to end as well14:18
sean-k-mooneydid you have a chance to test it with a real nvme device?14:18
chandankumarI also need to validate again with libvirt nvme setup14:18
chandankumarI was developing it with pci-sim14:19
chandankumarsean-k-mooney: Not yet, I will try to get one machine next week and validate there14:19
sean-k-mooneyso i think this will slip into next cycle 14:20
chandankumarOne more thing, I am not sure we are going to get this merged this week14:20
sean-k-mooneyi dont want to merge it without the real hardware testing14:20
sean-k-mooneythere are part we can but not the full feature14:20
chandankumarHow do we ask for feature freeze exception based https://docs.openstack.org/cyborg/latest/contributor/release-guide.html#week-following-milestone-3 or move it to next release14:20
chandankumaryes totally agree, we need a proper validation on real hardware14:21
sean-k-mooneyasyc -> mailing list post, or syncounly in the meeting after ff14:21
sean-k-mooneyso next wekk if you have a host to test it on we coudl condier grant a FFE till the end ot the week14:21
sean-k-mooneyotherwise we will jsut need to resubmit the spec and wew can continue workign on it after RC1 is cut14:22
sean-k-mooneyits effectvly only a 2 week delay if we are still happy with the direction14:22
chandankumargreat, I will update the status based on real hardware availability.14:22
chandankumarin next meeting14:22
chandankumarthank you, that's it on nvme14:23
chandankumarone last thing on milestone 314:23
chandankumarMarking the max microversion for the release14:24
sean-k-mooneyunless the manage mode feature lands it will be unchanged at 2.314:24
sean-k-mooneyso there isnt really much to do14:24
chandankumarok then we are good here14:24
sean-k-mooneyhttps://docs.openstack.org/cyborg/latest/contributor/rest_api_version_history.html14:24
sean-k-mooneywe only need to mark it really if we add on in a cycle14:25
chandankumarok14:25
sean-k-mooneyalthough we could backfile the markers for 2.2 and 2.314:25
sean-k-mooneyi suspect that was missed but that is just a docs change14:26
sean-k-mooneyso it can merge up until rc114:26
chandankumarI will take a look how to add the marker for 2.2 and 2.314:26
chandankumarok14:26
sean-k-mooneynova tend to update the marker even if there istne an api change in a relase14:26
sean-k-mooneywe can do that but im not sure if its needed14:27
chandankumarok14:27
sean-k-mooneyat least for a few years nova tended ot have 1 or more api versoin a release14:28
chandankumarsean-k-mooney: can you share the patch from nova which does that after the meeting.14:29
sean-k-mooneysure14:29
chandankumarany more question on milestone 3?14:29
chandankumarif not then moving to reviews section14:30
chandankumar#topic Reviews14:30
chandankumarHere is the link to ongoing reviews14:30
chandankumar#link https://etherpad.opendev.org/p/cyborg-2026.2-status14:30
gamiohttps://review.opendev.org/c/openstack/cyborg/+/1000240 device_addresses allow-list for the intel nic driver.14:30
chandankumarwe have already covered about 2026.2 spec implementations14:30
chandankumargamio: go ahead14:30
gamiomelwitt already gave +1.14:30
gamiocase-matching is a separate follow-up, bug 2163799. this patch stays as is.14:31
gamiosean-k-mooney: no rush this week. if you get a chance before FF, that'd be great.14:31
sean-k-mooneychandankumar: its just https://github.com/openstack/nova/commit/73724fef9a66c4df3d018e7368067f883b1ed9e2 but if you have more then 1 or two release where it does nto change it gets too long IMO14:31
chandankumarI will take a look this one14:31
chandankumarthank you sean-k-mooney 14:32
sean-k-mooneycase-matching you mean case insentivity right14:32
gamiosean-k-mooney: yes, case insensitivity. right now the compare is exact-case14:33
sean-k-mooneythe pci adress is actully an int value as are the vendor id and product id adn what we normally see is jut a sting represeation14:33
gamioso a wrong-case or unmatched address is silently skipped instead of warned.14:33
sean-k-mooneyya so eventually we shoudl just normlaise the vlues 14:34
sean-k-mooneyby lowercasing them before comparing14:34
sean-k-mooneythere is no info encoded in teh casing14:34
gamiosean-k-mooney: agreed, that's exactly what i had in mind for the follow-up.14:34
sean-k-mooneyack14:34
chandankumarI also need to check case matching in nvme config14:35
chandankumarthank gamio!14:36
chandankumarThere are few open reviews from rlandy around docs14:36
chandankumarlet me paste it here14:37
chandankumar#link https://review.opendev.org/c/openstack/cyborg/+/1001774 Improve installation guide index14:37
chandankumar#link https://review.opendev.org/c/openstack/cyborg/+/1001779 Reorganize documentation landing page14:37
rlandythanks chandankumar 14:37
chandankumar#link https://review.opendev.org/c/openstack/cyborg/+/1001755 Move support matrix to admin docs14:37
rlandythese can wait until after FF if needed14:37
chandankumar#link https://review.opendev.org/c/openstack/cyborg/+/1001767 Add agentic scaffolding and contributor docs14:37
rlandyjust adding them here for reference14:37
chandankumarok, I have added them to cyborg-2026.2 status page14:38
chandankumarThank you rlandy for working on improving the docs.14:38
chandankumarwe will go over them.14:39
chandankumarAnything you want to highlight on these patches14:39
sean-k-mooneyrlandy: https://review.opendev.org/c/openstack/cyborg/+/1001767 need a rebase14:39
rlandyok - will take care of that14:39
sean-k-mooneyeitehr on top of the latest of the parent or direclty onto master14:39
rlandyack14:40
rlandyI assume the agentic one needs more work - I will bring that one back after FF for discussion14:40
sean-k-mooneyi have not looked at it yet14:41
sean-k-mooneybut ill let you know later today14:41
sean-k-mooneyskimming it its mostly the right shape14:41
sean-k-mooneyso it may be ok but we will see14:41
sean-k-mooneythanks for pickign that up14:42
chandankumarthat's it from review, moving to next section14:43
chandankumar#topic Bugs14:43
chandankumarwe have one bug from openstack-cyborg14:43
chandankumar#link https://bugs.launchpad.net/openstack-cyborg/+bug/2163799 device_addresses: case-sensitive match silently mishandles wrong-case or absent entries (nic, gpu)14:43
chandankumarI think this bug is related to the follow up improvement coming after https://review.opendev.org/c/openstack/cyborg/+/100024014:45
sean-k-mooneyyes14:47
gamioplan is a small shared helper in accelerator/common/utils.py (same module as parse_address), used by nic and gpu for now. stacked on top of 1000240.14:47
gamionot touching nvme yet, just keeping the helper generic enough for later.14:47
sean-k-mooneywe don tneed it for nvme14:48
sean-k-mooneyso this would only be a temproy helper i think14:48
sean-k-mooneybecuase the pci bus manager woudl elminat it14:48
chandankumarthe bug seems to be valid14:49
gamiosean-k-mooney: right, makes sense given the direction toward device_spec 14:49
sean-k-mooneythe util fucntion is fine to adress this in the short term14:49
sean-k-mooneyand for stable branches14:49
sean-k-mooneybut its a mini feature so i woudl triage this as wishlist14:50
chandankumarstatus?14:50
sean-k-mooneyit impvoes the ux but the fact its case sensitive todya is not by itself a bug14:50
sean-k-mooneybut this is a ux improvment14:50
sean-k-mooneyand its backward compatbale so i think we wuld backprot this14:50
sean-k-mooneychandankumar: triaged and wishlist woudl be my assesment14:51
chandankumarDone, thank you!14:51
sean-k-mooneygamio: i dont think we need to log a warning14:51
sean-k-mooneyeither we supprot case insentitive or we dont14:51
sean-k-mooneyif we supprot it then we shoudl not warn if we dont then the silent ignore is not really a bug14:52
sean-k-mooneywe can chat about that on the review14:52
chandankumarok14:52
gamiosean-k-mooney: got it, no warning either way then. just the case-insensitive compare. that simplifies the follow-up nicely.14:52
sean-k-mooneycase inssestive compare and we proably want to normalise the value in the db to lowercase14:53
sean-k-mooneybut we need to consider the impact of that14:53
sean-k-mooneyso i would focus on the compare for now14:54
sean-k-mooneyeventually we want a pci device form any driver to look the same as every other driver14:54
gamiosean-k-mooney: sounds good, compare-only for now. keeps the follow-up small and low risk.14:55
sean-k-mooneyat least when it comes to things like the adress14:55
sean-k-mooneycool i assigend https://bugs.launchpad.net/openstack-cyborg/+bug/2017513 to myelf by hte way14:56
sean-k-mooneysince i started working on that bug14:56
chandankumarok, thank you!14:56
chandankumarMoving to next topic14:56
chandankumar#topic open discussion14:56
chandankumarAnything you want to highlight or want to discuss now is the time14:57
sean-k-mooneyjust an fyi14:57
sean-k-mooneyredhat has a company day firday/sunday14:57
sean-k-mooneyso i and other will not be around im alos taking monday off14:57
chandankumarenjoy the time off!14:58
chandankumarthank you sean-k-mooney for mentioning that14:58
chandankumarMoving to last one14:58
chandankumar#topic Volunteers to chair the next meeting14:58
rlandyI can take it ... 14:59
chandankumarAnyone up to lead next irc meeting?14:59
chandankumarthank you rlandy!14:59
rlandywill hand to others after FF14:59
chandankumarwe are right on time14:59
chandankumarthank you all for attending14:59
chandankumar#endmeeting14:59
opendevmeetMeeting ended Tue Aug 25 14:59:37 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)14:59
opendevmeetMinutes:        https://meetings.opendev.org/meetings/cyborg/2026/cyborg.2026-08-25-14.01.html14:59
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/cyborg/2026/cyborg.2026-08-25-14.01.txt14:59
opendevmeetLog:            https://meetings.opendev.org/meetings/cyborg/2026/cyborg.2026-08-25-14.01.log.html14:59
rlandythanks chandankumar for running this week's call15:00

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