| opendevreview | melanie witt proposed openstack/cyborg master: Serve agent RPC under the validated resource provider name https://review.opendev.org/c/openstack/cyborg/+/1015965 | 00:50 |
|---|---|---|
| sean-k-mooney | o/ | 14:08 |
| sean-k-mooney | sorry i should have started the irc meeting a few minutes ago | 14:09 |
| sean-k-mooney | ill start it now in one sec | 14:09 |
| jgilaber | np, was about to ping you | 14:09 |
| sean-k-mooney | #startmeeting cyborg | 14:10 |
| opendevmeet | Meeting started Tue Nov 17 14:10:15 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. | 14:10 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 14:10 |
| opendevmeet | The meeting name has been set to 'cyborg' | 14:10 |
| sean-k-mooney | ping: amoralej bogdando jgilaber rlandy chandankumar skovili melwit | 14:10 |
| jgilaber | o/ | 14:10 |
| sean-k-mooney | we will give folks a minute or two to gather then we will start | 14:10 |
| sean-k-mooney | #link https://etherpad.opendev.org/p/openstack-cyborg-irc-meeting#L49 | 14:11 |
| sean-k-mooney | the adgenda is pretty light on topics but we have some revies and bugs to look at | 14:12 |
| sean-k-mooney | if you have anything to add feel free | 14:12 |
| sean-k-mooney | for now lets go directly to reviews and we can loop back ot topics at the end | 14:12 |
| sean-k-mooney | #topic reviews | 14:13 |
| sean-k-mooney | #link https://review.opendev.org/q/topic:%22bug/2157586%22 | 14:13 |
| sean-k-mooney | so these are 2 pathcs form chandan | 14:13 |
| sean-k-mooney | https://review.opendev.org/c/openstack/cyborg/+/997945?usp=search | 14:14 |
| sean-k-mooney | #link https://review.opendev.org/c/openstack/cyborg/+/997945?usp=search | 14:14 |
| sean-k-mooney | #link https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/992210?usp=search | 14:14 |
| sean-k-mooney | both related to https://bugs.launchpad.net/openstack-cyborg/+bug/2157586 | 14:14 |
| sean-k-mooney | Cyborg Device Profile API rejects OWNER_CYBORG trait | 14:15 |
| sean-k-mooney | last week we dicussed the validation and whether or not we shoudl rely on os-triats for that | 14:15 |
| sean-k-mooney | we agreed to remove the validation instead | 14:15 |
| sean-k-mooney | and rely on placement enforcign the tratis requirements | 14:15 |
| sean-k-mooney | that is what the first patch does | 14:16 |
| sean-k-mooney | #link https://review.opendev.org/c/openstack/cyborg/+/997945 | 14:16 |
| sean-k-mooney | the second change | 14:16 |
| sean-k-mooney | #link https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/992210 | 14:16 |
| sean-k-mooney | adds tempets testing of the placement trait reporting | 14:17 |
| sean-k-mooney | i think in general these just need code review | 14:17 |
| sean-k-mooney | but directionally i think they are ok | 14:17 |
| jgilaber | I've reviewed both patches, I think they are ok | 14:17 |
| sean-k-mooney | ack are there any questions or concerns on them? | 14:18 |
| sean-k-mooney | if not i think we can try to land them this week | 14:18 |
| jgilaber | there is an open thread on the tempest one about the use of 'placement_max_microversion=latest' | 14:18 |
| jgilaber | it's not wrong, just unnecessary | 14:18 |
| jgilaber | but yes, I think we can land them soon | 14:18 |
| sean-k-mooney | ya we could use a lower version with the relevent one for the trate info | 14:18 |
| sean-k-mooney | *trait info | 14:19 |
| sean-k-mooney | i have not reviewd that yet so ill respond in gerrit | 14:19 |
| sean-k-mooney | ok if no other comments lets move on to bugs, we have 4 new ones this week | 14:20 |
| sean-k-mooney | #topic bugs | 14:20 |
| sean-k-mooney | #link https://bugs.launchpad.net/openstack-cyborg/+bug/2174353 obj_make_compatible() raises ObjectActionError for objects with nested object fields (ExtARQ, DriverDevice) | 14:21 |
| sean-k-mooney | melwitt: i belvie ^ was one you foudn whne auditing our current ovo usage | 14:21 |
| sean-k-mooney | the tl;dr is that we do not have the code requried to backlevel nested ovos | 14:22 |
| sean-k-mooney | which is not surpsing as cybrog was not addign any code to backlevel at all | 14:23 |
| sean-k-mooney | which leads into the second bug | 14:23 |
| sean-k-mooney | #link https://bugs.launchpad.net/openstack-cyborg/+bug/2174354 | 14:23 |
| sean-k-mooney | ARQ object: project_id and attach_handle_uuid were added without version bumps; obj_make_compatible is missing | 14:23 |
| sean-k-mooney | it also was not alwasy bumping the versions when requried | 14:23 |
| sean-k-mooney | both of these bugs violate the intended usage of ovo for versioning object serisalistion | 14:24 |
| sean-k-mooney | they would break rooling upgrades in any project that support roolling upgrades | 14:25 |
| sean-k-mooney | however cyborg does not | 14:25 |
| jgilaber | both seem valid and important | 14:25 |
| jgilaber | is there any dependence between them? | 14:25 |
| sean-k-mooney | yes i tend to agree, no i do not belive so | 14:25 |
| sean-k-mooney | we can retoactively add the object version for the latter | 14:26 |
| sean-k-mooney | it wont really have an impact | 14:26 |
| sean-k-mooney | as the fiels were added so long ago that that | 14:26 |
| sean-k-mooney | branches that did nto have them are long eol | 14:27 |
| sean-k-mooney | the first bug however will need to be fixed before i can add the object verion tests as aprt of the ovo indirection sepc | 14:27 |
| sean-k-mooney | im inclidne to set https://bugs.launchpad.net/openstack-cyborg/+bug/2174354 as high | 14:28 |
| sean-k-mooney | and https://bugs.launchpad.net/openstack-cyborg/+bug/2174353 as medium | 14:28 |
| sean-k-mooney | sorry | 14:28 |
| sean-k-mooney | other wasy around | 14:28 |
| sean-k-mooney | https://bugs.launchpad.net/openstack-cyborg/+bug/2174353 the issue with nested object fields | 14:28 |
| sean-k-mooney | should be high | 14:28 |
| jgilaber | ok, sounds good | 14:28 |
| sean-k-mooney | https://bugs.launchpad.net/openstack-cyborg/+bug/2174354 the fact that an object version was not done is medium | 14:29 |
| sean-k-mooney | because everyoen who woudl have been broken by that wer boken in 2019 | 14:29 |
| jgilaber | yes, that makes them less critical | 14:31 |
| sean-k-mooney | ok i added a comment ot each set triaged and added tags | 14:32 |
| sean-k-mooney | https://bugs.launchpad.net/openstack-cyborg/+bug/2174355 is more of the same | 14:32 |
| sean-k-mooney | #link https://bugs.launchpad.net/openstack-cyborg/+bug/2174355 | 14:32 |
| sean-k-mooney | ExtARQ obj_make_compatible does not handle version 1.1 changes | 14:32 |
| jgilaber | yeah same, medium seems fine for this | 14:33 |
| sean-k-mooney | so before last cycle when i asked folks to start addign obj_make_compatiable | 14:33 |
| melwitt | o/ | 14:33 |
| sean-k-mooney | cybrog did not have any defintions of that if i recall | 14:33 |
| jgilaber | this does have a dependency on the first one | 14:33 |
| sean-k-mooney | i know some of the change perhaps in chandands serise backfile some of these methods | 14:33 |
| sean-k-mooney | i guess so yes | 14:34 |
| sean-k-mooney | at some point in the next 12-18 months im expecting use to repalce all our ovos with v2 objects | 14:35 |
| sean-k-mooney | mainly because i think we need to concodliate so we nolonge have top levlel and driver copies of the objects | 14:35 |
| sean-k-mooney | thta does not mean we cant add the backlevleing supprot | 14:36 |
| sean-k-mooney | however there are no supproted deployment senarios wehre that code would be requried | 14:36 |
| sean-k-mooney | i.e. all stable release supprot the latest version of the ovos so we should never have to backlevel | 14:37 |
| sean-k-mooney | for the saem reason as the other bug i think this should be traiges as medium | 14:37 |
| sean-k-mooney | what do folks think? | 14:37 |
| jgilaber | +1, medium seems right | 14:38 |
| jgilaber | they are not critical, but would be good to fix them | 14:39 |
| sean-k-mooney | ok left a comment | 14:41 |
| sean-k-mooney | that brings us to the final bug reported by melwitt | 14:41 |
| sean-k-mooney | #link https://bugs.launchpad.net/openstack-cyborg/+bug/2174459 | 14:41 |
| sean-k-mooney | API -> agent RPC (fpga_program) targets the agent by device hostname, which does not match the agent's RPC server name | 14:42 |
| sean-k-mooney | so ^ is a disconnecte between confg.HOST adn hypervior_hostname | 14:42 |
| sean-k-mooney | i think | 14:42 |
| sean-k-mooney | melwitt: do you want to cover this topic | 14:43 |
| melwitt | the fpga programming test is apparently the only one that makes a RPC call to the agent, somehow | 14:43 |
| sean-k-mooney | i have some comemtn on older reviwe wher ei point out that https://github.com/openstack/cyborg/blob/master/cyborg/accelerator/drivers/gpu/nvidia/sysinfo.py#L318 | 14:44 |
| sean-k-mooney | is kind fo problematic | 14:44 |
| melwitt | it's fine if you do :) but yeah it's that the RPC queue name is based off the short name (CONF.host) but then the RPC call is targeting a name based off the FQDN | 14:44 |
| melwitt | I proposed a patch that may or may not be the right fix, but it's just something to look at | 14:45 |
| sean-k-mooney | ack are you happy to assing https://bugs.launchpad.net/openstack-cyborg/+bug/2174459 to yourself | 14:46 |
| sean-k-mooney | i thinkt his is valid and medium | 14:47 |
| sean-k-mooney | ah i was wonderign if it was related to https://github.com/openstack/cyborg/commit/0e9a632f3165af8d108301d7838ba2792de64222 or not | 14:48 |
| melwitt | sure. I forgot the automation no longer does assignment | 14:48 |
| sean-k-mooney | really we need to have host and hostname on the devices at least internally | 14:48 |
| melwitt | yes I think it is related to that, that's what created the deviation | 14:48 |
| sean-k-mooney | the same way an instnace has host and hypervioer_hostname | 14:48 |
| sean-k-mooney | hum i see your fixing this by changing the rpc endpoint | 14:49 |
| sean-k-mooney | im not sure that is really correct | 14:50 |
| sean-k-mooney | the reaosn i did the split is its not genrelaly correct to assume the rpc endpoint name and placment resouce provier for a comute are the same | 14:50 |
| melwitt | yeah, I wasn't sure how to approach this | 14:50 |
| sean-k-mooney | ya its tricky | 14:52 |
| sean-k-mooney | im tryign to think what the upgrade impactg ot both aprohces are | 14:52 |
| sean-k-mooney | if we proceded with yoru change conf.host nolonger has any value | 14:52 |
| sean-k-mooney | it exist to defein the rpc endpoint that identifies the agent | 14:52 |
| sean-k-mooney | and we use the hostname as the identifyer by defualt | 14:53 |
| sean-k-mooney | since it should be unique, it also has to be a value that is agreed between nova and neutron and cybrog | 14:53 |
| sean-k-mooney | seperatly the compute node root resouce provdier name is not CONG.host | 14:54 |
| sean-k-mooney | tis the comptue nodes hypervioer_hostname | 14:54 |
| melwitt | yeah, it does feel weird to make it act different than the usual way presumably all other services work. so the other way would be to make the caller somehow use the right queue name rather than the resource provider name? | 14:54 |
| sean-k-mooney | ya so i think we need ot add a second filed on the device objects | 14:55 |
| sean-k-mooney | to hold a refence to the service (or the queue topic) i.e CONF.host | 14:55 |
| sean-k-mooney | im just not sure how to do that in a backprotable way | 14:55 |
| sean-k-mooney | the orgianl problem i was tryign to solve is nova and cybrog did not agree on the name of the resouce provider | 14:56 |
| sean-k-mooney | cyborg expected the short hostname, libvirt provided the fqdn when there was a relvent domain | 14:56 |
| melwitt | yeah not sure how adding a second field on device could be backportable | 14:58 |
| sean-k-mooney | the only way to do it would be to stashit in an exitng colum | 14:58 |
| melwitt | (I have to join another call in a minute, sorry) | 14:58 |
| sean-k-mooney | no worries we could revert my eiarlry patch adn reintuduce it in master with a new filed | 14:59 |
| sean-k-mooney | im nust concerned that woudl also break peole | 14:59 |
| sean-k-mooney | it was possible to change CONF.hsot to "make it work" before | 14:59 |
| sean-k-mooney | and that also a workaround here | 14:59 |
| sean-k-mooney | lets reviesit this next week | 14:59 |
| sean-k-mooney | or on irc | 14:59 |
| sean-k-mooney | are we agreed its valid and high or medium | 15:00 |
| jgilaber | sounds good, I need to spend some time looking at this | 15:00 |
| sean-k-mooney | ill set it to high | 15:00 |
| sean-k-mooney | that more or less it for this week | 15:00 |
| jgilaber | yes, sounds right | 15:00 |
| sean-k-mooney | anyone open to takign the meting next week? | 15:01 |
| jgilaber | I can do it | 15:01 |
| sean-k-mooney | thank jgilaber | 15:01 |
| sean-k-mooney | in that case thanks for coming and lets call it there | 15:01 |
| sean-k-mooney | #endmeeting | 15:01 |
| opendevmeet | Meeting ended Tue Nov 17 15:01:55 2026 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 15:01 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/cyborg/2026/cyborg.2026-11-17-14.10.html | 15:01 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/cyborg/2026/cyborg.2026-11-17-14.10.txt | 15:01 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/cyborg/2026/cyborg.2026-11-17-14.10.log.html | 15:01 |
| *** user is now known as Guest1182 | 18:00 | |
| opendevreview | melanie witt proposed openstack/cyborg master: Carry the agent RPC identity on reported devices https://review.opendev.org/c/openstack/cyborg/+/1015965 | 18:47 |
| melwitt | sean-k-mooney: changed ^ to add a service_host field to devices. note I will be out next week but we can discuss before/after that | 18:51 |
| sean-k-mooney | melwitt: ack cool | 19:06 |
| sean-k-mooney | i think that is certenly a better long term fix for master | 19:07 |
| sean-k-mooney | the real question is for stable are we ok with jsut documentign that CONF.host shoudl equal CONF.agent.resource_provider_name | 19:08 |
| sean-k-mooney | that is the workaorund for now just make the two the same. today they only differ if socket.getHostname() and socket.getFQDN() retun differnet values | 19:09 |
| melwitt | that's what I'm thinking | 19:12 |
| opendevreview | melanie witt proposed openstack/cyborg master: Carry the agent RPC identity on reported devices https://review.opendev.org/c/openstack/cyborg/+/1015965 | 19:13 |
| sean-k-mooney | melwitt: are you on pto form tomorrow? | 19:21 |
| melwitt | sean-k-mooney: no, not until next week | 19:23 |
| sean-k-mooney | oh ok im going to finish up shortly be we can caht about the the patch more tomroow so | 19:23 |
| melwitt | basically I am out next week only | 19:24 |
| melwitt | not sure how much of a rush it is, just didn't want you to be disappointed next week if I'm not around | 19:25 |
| melwitt | and I figured better to put upload something sooner than later so we don't forget about it | 19:25 |
| sean-k-mooney | i think now that we know its an issue as long as we reolve it this cycle its fine | 19:29 |
| sean-k-mooney | so important but not urgent | 19:30 |
| * melwitt nods | 19:31 | |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!