| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-cyborg - 2026-09-08 | |||
| 12:49:31 | opendevreview | chandan kumar proposed openstack/cyborg master: Add NVMe driver devstack plugin support https://review.opendev.org/c/openstack/cyborg/+/999945 | |
| 12:49:31 | opendevreview | chandan kumar proposed openstack/cyborg master: pci-sim: add software MSI-X domain to fake host bridge https://review.opendev.org/c/openstack/cyborg/+/1003255 | |
| 13:42:32 | opendevreview | Merged openstack/cyborg master: Add 2026.2 release prelude https://review.opendev.org/c/openstack/cyborg/+/1004468 | |
| 14:01:55 | opendevmeet | The meeting name has been set to 'cybrog' | |
| 14:01:55 | opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | |
| 14:01:55 | opendevmeet | Meeting started Tue Sep 8 14:01:54 2026 UTC and is due to finish in 60 minutes. The chair is sean-k-mooney. Information about MeetBot at http://wiki.debian.org/MeetBot. | |
| 14:01:55 | sean-k-mooney | #startmeeting cybrog | |
| 14:02:14 | chandankumar | o/ | |
| 14:02:19 | morenod | o/ | |
| 14:02:22 | sean-k-mooney | ping list amoralej bogdando jgilaber rlandy chandankumar skovili melwit | |
| 14:02:25 | jgilaber | o/ | |
| 14:02:32 | rlandy | o/ | |
| 14:02:53 | sean-k-mooney | #link https://etherpad.opendev.org/p/openstack-cyborg-irc-meeting | |
| 14:03:02 | sean-k-mooney | we will start shortly | |
| 14:03:35 | sean-k-mooney | if there are any topics feel free to add them to the adjeda above | |
| 14:03:49 | melwitt | o/ | |
| 14:04:42 | sean-k-mooney | ok i think that is most of the folks that are around today | |
| 14:04:49 | sean-k-mooney | so lets start | |
| 14:05:00 | sean-k-mooney | #topic Support for multiple resource classes per deployable | |
| 14:05:26 | sean-k-mooney | jgilaber: this one is yours | |
| 14:05:45 | jgilaber | yes, I wanted to ask about how to handle this | |
| 14:05:55 | sean-k-mooney | so contex for every one | |
| 14:06:06 | jgilaber | for context, cyborg only supports one resource class per deployable | |
| 14:06:09 | sean-k-mooney | today cyborg modesl resouce provider in placmeent as deployables | |
| 14:06:56 | sean-k-mooney | yep which for the generic mdev drvier or pci driver means that if we want to group mdevs or pci vf by the parent pf | |
| 14:07:14 | sean-k-mooney | they either must all have the same resouce class | |
| 14:07:27 | sean-k-mooney | or we have to model it diffently with nested resouces | |
| 14:07:41 | sean-k-mooney | nested provdiers | |
| 14:08:01 | sean-k-mooney | so i guess your question is shoudl this be its own feature/spec | |
| 14:08:20 | sean-k-mooney | the ablity to have more then one resocue class per deploryable/resource provider | |
| 14:08:23 | jgilaber | exactly | |
| 14:08:29 | sean-k-mooney | im inlcined to say yes | |
| 14:08:37 | jgilaber | in the approved spec we agreed to have one deployable per mdev type and revisit in the future | |
| 14:08:46 | sean-k-mooney | repopose the mdev spec as is with only 1 type | |
| 14:09:01 | jgilaber | but since I have to repropose anyway I figured I'd look into it | |
| 14:09:05 | sean-k-mooney | and seperately propsole a seperate feature to supprot multiple rescoues classes per deployable | |
| 14:09:40 | sean-k-mooney | jgilaber: so without going into the details too much | |
| 14:09:51 | jgilaber | wanted to get people thoughts but that was my idea since this would affect all drivers | |
| 14:09:55 | sean-k-mooney | i think we need to move to modelign the resouce class on teh device or attachment handel | |
| 14:10:31 | sean-k-mooney | currently i belive we use the atibutes tabel to store it correct? | |
| 14:11:12 | jgilaber | yes, I think that's right | |
| 14:11:27 | jgilaber | e.g from the pci driver https://github.com/openstack/cyborg/blob/master/cyborg/accelerator/drivers/pci/pci/sysinfo.py#L71-L74 | |
| 14:11:36 | sean-k-mooney | we maynot need ot modify the device or attach hanel actully it just depend what atributes are assocated with | |
| 14:11:54 | sean-k-mooney | ah so ya that is the problem | |
| 14:11:56 | sean-k-mooney | https://github.com/openstack/cyborg/blob/master/cyborg/db/sqlalchemy/models.py#L144-L156 | |
| 14:12:19 | sean-k-mooney | the Attribute table references the deployable not the device | |
| 14:13:05 | sean-k-mooney | i think this is a generic capablity that all drviers shoudl be able to leverage when needed | |
| 14:13:15 | sean-k-mooney | so to me that makes it its own spec | |
| 14:13:24 | sean-k-mooney | anyone else have an opionon on that? | |
| 14:13:47 | melwitt | I think that sounds a good idea | |
| 14:14:05 | melwitt | keep each piece smaller | |
| 14:14:21 | sean-k-mooney | there is a larger related topic | |
| 14:14:27 | jgilaber | ack, I was leaning towards that so +1 from me | |
| 14:14:45 | sean-k-mooney | i think we need to look at and impvoe the modeling of device, atibutes, deployabels and attachment handels in general | |
| 14:15:07 | sean-k-mooney | even if tha tis just docs to descibe there role | |
| 14:16:07 | sean-k-mooney | right now the exact meaning/uage of all of the above feel a bit fuzzy and it woudl be nice to impvoe that going forward | |
| 14:16:31 | sean-k-mooney | i think this is just one fo the pain point in that model but perhasp its a good one to adress first | |
| 14:17:16 | jgilaber | that's a good point, we can build docs as part of this new spec | |
| 14:17:48 | sean-k-mooney | part of the problem is the atribtues api | |
| 14:17:52 | sean-k-mooney | that was merged | |
| 14:17:58 | sean-k-mooney | did not follow the approves sepc | |
| 14:18:19 | sean-k-mooney | so its an areay of technial debt that we need to adress at some point | |
| 14:19:29 | sean-k-mooney | so we may want to adress that as part of this work and use this as an opterutnity to docuemtn the intended behvior | |
| 14:19:56 | sean-k-mooney | anyway we can dicuss that in a spec or ptg topic unless there is anyting else you want to raise today? | |
| 14:20:18 | jgilaber | no, that is ok for now, we can go into more detail at ptg I think | |
| 14:20:33 | jgilaber | I'll also will have spent some time looking at the details by then | |
| 14:21:07 | sean-k-mooney | ack sound good, we just need to make sure we capture all the contenxt in the evneutal spec | |
| 14:21:47 | sean-k-mooney | ok ill move on to the next topci breifly | |
| 14:22:03 | sean-k-mooney | #topic ptg planning | |
| 14:22:24 | sean-k-mooney | #link https://etherpad.opendev.org/p/cyborg-2027.1-ptg | |
| 14:22:42 | sean-k-mooney | so that is currently empty but we need to start planing the ptg | |
| 14:22:56 | sean-k-mooney | ill add a basic safold shortly | |
| 14:23:25 | sean-k-mooney | but we need to make a request to have a seession | |
| 14:23:33 | sean-k-mooney | adn plan our slots | |
| 14:23:50 | sean-k-mooney | the dates are October 12-16, 2026 | |
| 14:25:02 | sean-k-mooney | do folks have a prefernce for how many days or time slots? | |
| 14:26:03 | sean-k-mooney | the current slots are form last ptg so we could just use the same but lets loop back to this next week and folks can think about it | |
| 14:26:12 | jgilaber | personally, I won't be available on October 12th, so I would prefer to avoid scheduling on that day | |
| 14:26:38 | sean-k-mooney | ok lets use https://etherpad.opendev.org/p/cyborg-2027.1-ptg#L12 | |
| 14:26:59 | sean-k-mooney | and if folks can add there aviablity we can see when will work for folks next week | |
| 14:27:25 | sean-k-mooney | #action review ptg slots in next meeting | |
| 14:28:00 | sean-k-mooney | anything else on this topic? | |
| 14:29:00 | sean-k-mooney | #topic reviews | |
| 14:29:28 | sean-k-mooney | first one 1004438: Move 2026.2 RBAC spec to implemented | |
| 14:29:40 | sean-k-mooney | #link https://review.opendev.org/c/openstack/cyborg-specs/+/1004438 | |
| 14:30:18 | sean-k-mooney | i think this is pretty straigt forward and correct | |
| 14:30:30 | sean-k-mooney | thank jgilaber for proposing that | |
| 14:30:39 | jgilaber | I pushed this yesterday, not urgent, just had not done it yet | |
| 14:30:54 | jgilaber | I also looked at launchpad, I think we're all set there? | |
| 14:31:07 | sean-k-mooney | so that actully a good point | |
| 14:31:11 | sean-k-mooney | for that spec yes | |
| 14:31:38 | sean-k-mooney | but we need to move the other blueprint to 2027.1 | |
| 14:31:40 | sean-k-mooney | https://blueprints.launchpad.net/openstack-cyborg | |
| 14:32:09 | sean-k-mooney | i will move https://blueprints.launchpad.net/openstack-cyborg/+spec/generic-mdev-driver and https://blueprints.launchpad.net/openstack-cyborg/+spec/generic-nvme-driver-with-secure-cleanup now | |
| 14:32:29 | sean-k-mooney | and reset definti/appoved | |
| 14:32:41 | jgilaber | ack, thanks | |
| 14:32:46 | sean-k-mooney | once those are repoposed we can appove them in lauchpad again | |
| 14:33:25 | sean-k-mooney | anything else on this? | |
| 14:33:56 | jgilaber | not from me | |
| 14:34:32 | sean-k-mooney | ok next one is "rc1 and marketing highlights" | |