Earlier  
Posted Nick Remark
#openstack-cyborg - 2026-02-24
14:31:56 sean-k-mooney but we are not in a super rush here
14:32:11 jgilaber I'll review it after the meeting
14:32:36 sean-k-mooney ack ill check back then thrusday or friday adn respond to any feedback if there is any
14:33:21 jgilaber thanks, I think we can move to the next one
14:33:23 jgilaber #link https://review.opendev.org/c/openstack/cyborg/+/977020/3
14:33:46 sean-k-mooney so the next review is the release note prelude
14:33:50 sean-k-mooney #link https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_dfc/openstack/dfcb37a19d884363afaee29efe3c877d/docs/unreleased.html
14:34:03 sean-k-mooney this will become the 2026.1 prelude when we cut the release
14:34:34 sean-k-mooney rlandy: suggested i add a refence to docuemantion so i added it in the last revision
14:35:03 sean-k-mooney im happy to refine it more if there is addtional feedback but i would like to merge this by the end of the week if no outstanding issues are raised
14:35:15 rlandy I'll review again
14:35:17 sean-k-mooney the hard deadlien for merging this is before rc1
14:35:26 sean-k-mooney so we have 3 weeks
14:35:36 sean-k-mooney rlandy: ack thanks
14:37:00 sean-k-mooney the only ohter comment i have on that is we have the opertunity to add a marketing highlights as part of the release
14:37:27 sean-k-mooney i do not know if there is anythign we woudl ike to do in that regard but now is a good time to start thinking about that
14:38:08 sean-k-mooney i was thinkign fo reuseing the prelude to some degree adn put more of a focus into that for next cycle
14:38:37 rlandy +1
14:39:56 sean-k-mooney lets keep moving ot get though the adgenda unless there is anything else on this topic. ^ is technially the last change we need for the release
14:40:53 jgilaber ack thanks sean-k-mooney
14:40:58 jgilaber next review
14:41:00 jgilaber #link https://review.opendev.org/c/openstack/cyborg/+/977083
14:41:22 sean-k-mooney so ^ is a work in progress change that chandan has been working on
14:41:47 sean-k-mooney it is using a project https://github.com/snu-csl/nvmevirt
14:41:59 sean-k-mooney to provide a kernel module to emulate nvme devices
14:42:23 sean-k-mooney this we hope to use to test the pci driver and a future nvme/ssd driver
14:42:38 sean-k-mooney this is making good progress but need more work before it is ready
14:42:58 sean-k-mooney the next review in the list was https://review.opendev.org/c/openstack/project-config/+/977194
14:43:09 sean-k-mooney which add the project to our zuul tenant
14:43:16 sean-k-mooney so that we cna have zuul prepare the git repo for us
14:43:21 sean-k-mooney that is now merged
14:43:30 sean-k-mooney #link https://review.opendev.org/c/openstack/project-config/+/977194
14:44:00 sean-k-mooney if folks have tiem to review over the next week or two we will likely hold htat until after RC1 is cut but its good to review
14:45:29 sean-k-mooney my proposal is to review the branch status and review if we are ready to crate the rc1 tag during next weeks meeting
14:46:26 sean-k-mooney any comment or questions?
14:47:06 sean-k-mooney if not the last 2 reviews also go togehter
14:47:34 jgilaber next one is
14:47:36 jgilaber #link https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/976215
14:48:01 sean-k-mooney so this unfortually depend on https://review.opendev.org/c/openstack/cyborg/+/976224
14:48:23 sean-k-mooney which is currently blocked by a bug in the cherry-pick check which is fixed by https://review.opendev.org/c/openstack/cyborg/+/977793
14:48:42 sean-k-mooney so we will have to merge ^ on master and cherry-pick it back
14:49:02 sean-k-mooney before we can fix the package name and then update the jobs in the tempest plugin
14:49:19 sean-k-mooney none of that is hard but we just need to do it in the corerct order
14:49:43 sean-k-mooney i hope we can resolve that and complete updating the master branch of the tempest plugin in the next week or so
14:50:09 sean-k-mooney that effeitlvy it for reviews form me for now
14:50:39 jgilaber sounds reasonable, any other comments on the reviews? If not we can do some bug triaging
14:51:07 jgilaber #topic Bugs
14:51:23 jgilaber first one
14:51:25 jgilaber #link https://bugs.launchpad.net/openstack-cyborg/+bug/1987380
14:51:47 sean-k-mooney so this is a valid bug IMO, i marked it as medium for now
14:52:06 sean-k-mooney the current nvidia driver only supprot full gpu passthough or mdevs
14:52:27 sean-k-mooney unfortuletly the way it enmerate devices also include the VFs that mdevs are creted form for sriov capable gpus
14:52:46 sean-k-mooney that a bug and the correct fix for now is to filter out any nvida pci device that is a VF
14:53:13 sean-k-mooney long term we wil want to supprot mig and extendign this driver is one option but right now you cannot use this driver with mig VFs
14:53:27 sean-k-mooney it could crash the host kernel on older verions fo rhel for example
14:54:00 sean-k-mooney so this is a relitlivy easy bug for someone to fix but we do not need to fix it imiedealy
14:54:25 sean-k-mooney any comments/concerns with trating this as medium
14:55:06 jgilaber I'm not too familiar with the nvidia driver, but based on the context mediums seems ok
14:56:13 jgilaber any other comment? if not we can take another bug
14:56:24 jgilaber #link https://bugs.launchpad.net/openstack-cyborg/+bug/1954886
14:57:09 sean-k-mooney this i have triaged as high. the tl;dr is we have a issue with teh current api policy that allows you do detach a arch form a nova vm via the cyborg api
14:57:29 sean-k-mooney that is a mild security issue but since you can only break your own vms less so
14:57:56 sean-k-mooney the fix for this is we need to implement SRBAC next cycle and i plan to spen the next week or so investigatin that and preparing a spec
14:58:21 sean-k-mooney i have set this as high to reflect that but im unsure fi there is a backportable solution to this
14:58:25 jgilaber looking at the last comment, do you think we'll we need to anything else on top of the SRBAC implementation for this bug?
14:58:26 sean-k-mooney so this may be a master only fix
14:58:56 sean-k-mooney i dont think we do but we may make this a ptg topic to dicussi t futuer
14:59:19 jgilaber ack, so high seems appropiate
14:59:19 sean-k-mooney for now we are aware of i ti and we can revist. do you want to do the last 2 quickly?
14:59:42 jgilaber let's try
14:59:45 jgilaber #link https://bugs.launchpad.net/openstack-cyborg/+bug/2020178
15:00:39 jgilaber as noted, this one seems to be an exact duplicated of an already closed bug, I think it's safe to close
15:00:50 sean-k-mooney i think so too
15:01:01 sean-k-mooney and there is a comment to that effect on the bug
15:01:09 jgilaber any objection to closing it?
15:01:20 sean-k-mooney nope +1
15:01:42 jgilaber done, last one
15:01:50 jgilaber #link https://bugs.launchpad.net/openstack-cyborg/+bug/1933399
15:03:02 sean-k-mooney ah i see
15:03:06 sean-k-mooney took a second
15:03:23 sean-k-mooney http://cyborg.openstack.svc.region-stackdev.myinspurcloud.com/v2 vs http://cyborg.openstack.svc.region-stackdev.myinspurcloud.com/accelerator/v2
15:03:36 sean-k-mooney it shoudl be http://cyborg.openstack.svc.region-stackdev.myinspurcloud.com/v2/accelerator
15:03:39 sean-k-mooney i think
15:04:03 sean-k-mooney we likely need to verify if this stilll hapens and then restore the orgian proposed fix
15:04:45 jgilaber sounds right, move it back to triaged and medium?
15:04:48 sean-k-mooney im not sure the propsoed fix is correct
15:05:00 sean-k-mooney ya that sounds correct to me
15:05:24 jgilaber ack, so that's all the bugs we had for today
15:05:28 jgilaber final topic
15:05:38 jgilaber #topic Volunteers to chair next meeting
15:05:41 jgilaber any volunteers?
15:05:48 rlandy I'll do it
15:05:53 jgilaber thanks rlandy
15:05:54 sean-k-mooney :)
15:06:04 jgilaber and thanks everyone for participating, see you next week!
15:06:09 sean-k-mooney i guess we should ask if there are any last miniut topics
15:06:10 rlandy oops actually I am out
15:06:31 rlandy I will find someone else
15:06:31 sean-k-mooney i can do it of otehrs are not free
15:06:40 sean-k-mooney ack we can figure it out offline

Earlier   Later