Tuesday, 2026-03-10

« Friday, 2026-03-06 Index Wednesday, 2026-03-11 »
opendevreviewchandan kumar proposed openstack/cyborg master: Add NVMeVirt emulated NVMe device support for devstack  https://review.opendev.org/c/openstack/cyborg/+/97708305:54
opendevreviewchandan kumar proposed openstack/cyborg master: Add NVMeVirt emulated NVMe device support for devstack  https://review.opendev.org/c/openstack/cyborg/+/97708306:03
opendevreviewchandan kumar proposed openstack/cyborg master: Add NVMeVirt emulated NVMe device support for devstack  https://review.opendev.org/c/openstack/cyborg/+/97708306:39
opendevreviewchandan kumar proposed openstack/cyborg master: Added ansible-lint and bashate pre-commit hook  https://review.opendev.org/c/openstack/cyborg/+/97978807:12
opendevreviewchandan kumar proposed openstack/cyborg master: [WIP] Add cyborg greande job  https://review.opendev.org/c/openstack/cyborg/+/97980210:12
opendevreviewchandan kumar proposed openstack/cyborg master: [WIP] Add cyborg greande job  https://review.opendev.org/c/openstack/cyborg/+/97980210:52
opendevreviewchandan kumar proposed openstack/cyborg master: Rename devstack CYBORG_CONF_FILE to CYBORG_CONF for consistency  https://review.opendev.org/c/openstack/cyborg/+/97980511:30
opendevreviewchandan kumar proposed openstack/cyborg master: Add NVMeVirt emulated NVMe device support for devstack  https://review.opendev.org/c/openstack/cyborg/+/97708311:32
opendevreviewchandan kumar proposed openstack/cyborg master: Added ansible-lint and bashate pre-commit hook  https://review.opendev.org/c/openstack/cyborg/+/97978811:32
opendevreviewchandan kumar proposed openstack/cyborg master: [WIP] Add cyborg greande job  https://review.opendev.org/c/openstack/cyborg/+/97980211:51
sean-k-mooneyo/13:04
sean-k-mooneyactully the irc call is not for an hour13:05
sean-k-mooneywe moved our internal meeeting an hour early but the upstream irc call is at its normal time13:05
sean-k-mooneyso we normally dont have a break in between but we do today13:05
sean-k-mooneyrlandy: ^13:05
rlandyyes - correct13:07
skoviliack13:09
chandankumarCyborg meeting starts in 27 mins, Please add your agenda here: https://etherpad.opendev.org/p/openstack-cyborg-irc-meeting#L4313:33
opendevreviewchandan kumar proposed openstack/cyborg-tempest-plugin master: [WIP] Add cyborg pci nvme test  https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/97984713:59
chandankumarHello, who all are around for cyborg meeting?14:01
rlandyo/14:02
jgilabero/14:02
chandankumar#startmeeting cyborg14:02
opendevmeetMeeting started Tue Mar 10 14:02:23 2026 UTC and is due to finish in 60 minutes.  The chair is chandankumar. Information about MeetBot at http://wiki.debian.org/MeetBot.14:02
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:02
opendevmeetThe meeting name has been set to 'cyborg'14:02
bogdando[m]o/14:02
chandankumar#courtesy ping sean-k-mooney amoralej jgilaber 14:03
chandankumarfeel free to add topics to the agenda14:03
amoralejo/14:03
chandankumar#link https://etherpad.opendev.org/p/openstack-cyborg-irc-meeting#L4314:04
sean-k-mooneyhi o/14:05
chandankumarstarting with the first topic14:05
chandankumar#topic testing nvme pci device tempest tests in Upstream CI14:05
chandankumarI was working on adding nvmevirt devstack plugin for cyborg to test nvme pci device in CI https://review.opendev.org/c/openstack/cyborg/+/97708314:06
chandankumarI was working on adding the tempest scenario test https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/979847 where we can expose the nvme pci device to the vm14:07
chandankumarBut while creating the vm, On nove compute side, I am hitting unsupported configuration: host doesn't support passthrough of host PCI devices14:07
sean-k-mooneyya so nova can virutalize a iommu but you need to enable it in the image uploaded to the provider14:08
chandankumarLater on I realized that I need to enable IOMMU on the host to make sure pci passthroug works14:08
sean-k-mooneyalthernitivly i think there are 2 other options14:08
sean-k-mooney1 enable the softiommu on the kernle command line14:08
sean-k-mooney2 set teh mdoule option for vfio-pci to allow "unsafemode"14:08
sean-k-mooneymodprobe vfio enable_unsafe_noiommu_mode=1. 14:09
chandankumarthanks sean-k-mooney Do we have any job upstream using any of these options?14:09
sean-k-mooneyno14:09
chandankumarDo we need to change the nodeset label also? or current one will work fine14:09
sean-k-mooneywithout those we could still do some limited testing14:09
sean-k-mooneyjust not passhtough to a vm14:09
sean-k-mooneyi..e we could bind an arq and unbindi it formn tempet14:10
sean-k-mooneyin a whitebox style plugin we coudl aslo ssh in write some data to the device and hten asert that its cleaned when unbound14:10
sean-k-mooneyit would be preferable to test it via nova howver14:10
sean-k-mooneywith standard tempest14:10
chandankumarbind and unbind arg sounds good14:12
chandankumarI will break the existing test into multiple tests to have bind/ubind as well as passthrough to vim via standard tempest tests14:13
chandankumarThank you! I will update it in next meeting how it goes.14:13
bogdando[m]a naive question, can't we reboot zuul node in CI? that breaks devstack, right?14:14
sean-k-mooneyyes and no14:14
sean-k-mooneyreboots dont break devstack14:14
sean-k-mooneyand we can remboot the vms 14:14
sean-k-mooneywe woudl do that before installing devstack however14:14
sean-k-mooneyif the kernel is compiled with supprot for it14:16
sean-k-mooneyecho 1 | sudo tee /sys/module/vfio/parameters/enable_unsafe_noiommu_mode14:16
bogdando[m]so worth exploring?14:16
sean-k-mooneyis the simples option in the exsitng pre playbook14:16
sean-k-mooneywithout that we can still do some testign via whtiebox we just can boot a vm. lets see if ubuntu 24.04 or debian 13 ship with that kernel module partmer14:17
sean-k-mooneyif not we can revisit how we approch this14:17
sean-k-mooneyliek addign a new image to the ci and nodepool lable with the viommu enabled14:18
sean-k-mooneywe dont wnat to do that by defualt as there is a non trivial performace hti to that14:18
chandankumarI will check with my existing nvmevirt preplaybook14:18
chandankumaryes, we can revisit this topic in next meeting based on testing how it goes.14:19
chandankumarAny more questions on this topic?14:20
sean-k-mooneyubuntu@openclaw:~$ grep CONFIG_VFIO_NOIOMMU /boot/config-$(uname -r)14:20
sean-k-mooneyCONFIG_VFIO_NOIOMMU=y14:20
sean-k-mooneyso ubuntu 24.04 appears to have it enabeld at least if we install the mainline kernel14:20
bogdando[m]\o/14:20
sean-k-mooneybut now lets see what your ci run shows14:20
sean-k-mooneyif we need to install the mainline (upstream kernel) we can do that but we need to verify it in a vanilla 24.0414:21
chandankumaryup, I will verify it locally with mainline kernel14:22
chandankumargood point!14:22
chandankumarmoving to next topic14:23
chandankumar#topic Reviews14:23
chandankumarthere is one review14:23
chandankumar#link 979805: Rename devstack CYBORG_CONF_FILE to CYBORG_CONF for consistency | https://review.opendev.org/c/openstack/cyborg/+/97980514:23
chandankumarWhile I was working on nvmevirt devstack patch, I found that In cyborg we use CYBORG_CONF_FILE var for cyborg conf in devstack14:24
sean-k-mooneyhum i think its currently CYBORG_CONF_FILE to disambiguate with CYBORG_CONF_DIR14:24
chandankumarbut in Nova or other services we use <SERVICE>_CONF, so I thought about changing it to nova format14:24
sean-k-mooneywell nova has more then one conf it uses with mutlipe variables14:25
sean-k-mooneyi.e. NOVA_CPU_CONF14:25
sean-k-mooneyi think14:25
sean-k-mooneywe could make this change im just not sure it by us much14:26
sean-k-mooneyif we do this i woudl make it fallback to the old name14:27
sean-k-mooneyso CYBORG_CONF=${CYBORG_CONF:-${CYBORG_CONF_FILE}}14:28
chandankumarok make sense, It will not break existing user14:28
chandankumarI will update the review after this call. 14:28
sean-k-mooneyack14:28
chandankumarsince there is no more reviews moving to next topic14:29
chandankumar#topic Bugs14:29
chandankumar#link https://bugs.launchpad.net/openstack-cyborg/+bug/1954888 Docs: state how to use bound accelerators14:29
chandankumarwe discussed this bug in the last meeting14:29
sean-k-mooneyya so im incliend to ether use this to track descirbing how to create the nova flavor and device profile14:30
bogdando[m]looks like it should stay deferred unless standalone cyborg gets on the radar14:30
sean-k-mooneyand that workflow 14:30
sean-k-mooneyor using it to track a contibutor doc or closing as wont fix14:31
sean-k-mooneyright standalon cyborg is not really a thing today so we shoudl nto have end user docs for that14:31
sean-k-mooneyso either we have a conceptal do for devleoper/contibutors14:31
sean-k-mooneyor we docuemnt the nova flow for end admins ro end users14:32
sean-k-mooneybut i dont think we shoudl spend time on stand alone14:32
sean-k-mooneywe could close this and create 2 seperate bugs for those other two docs gaps14:32
jgilaber+1 I think there is value in documenting the standard workflow using nova, but I would not go further than that for now14:32
sean-k-mooneyi.e. 1 for admin doc for configuing and testign the standard wrokflow and 2 a contibutor doc to serve as a technial deep dive/refence of howt he services interact?14:33
chandankumarDo we have a reference for nova flow doc link which we can add as a reference in the bug?14:34
sean-k-mooneyi dont think this really exist out side fo the spec today14:35
sean-k-mooneyi dont recall htere being a nova doc for ti but there may be14:35
chandankumarok14:36
sean-k-mooneyi will file 2 seperate bugs and close this when done14:36
chandankumarthanks sean-k-mooney!14:37
sean-k-mooneyi can likely create a draft for the docs in the next few weeks if no one else does14:37
chandankumarok14:37
chandankumarany more question on this bug.14:38
chandankumarmoving to next one14:38
chandankumar#link https://bugs.launchpad.net/openstack-cyborg/+bug/1955285 arq delete no valid14:38
sean-k-mooneythat is just invliad14:39
sean-k-mooneyopenstack accelerator arq delete c48f5ac0-0bff-44c0-9106-1850b355c5ca woudl be deliting arq c48f5ac0-0bff-44c0-9106-1850b355c5ca14:39
sean-k-mooneynot the arq for instance c48f5ac0-0bff-44c0-9106-1850b355c5ca14:39
sean-k-mooneywe might have a docs bug or a help text bug if it imples its the vm uuid14:40
bogdando[m]how reasonable is having this in CLI assuming only service user should be allowed to do that low level things for/via nova?14:41
sean-k-mooneybut that is not what "openstack accelerator arq delete" means per the openstack client naming conventions 14:41
sean-k-mooneybogdando[m]: it some thing we will need to enfoce going forard14:41
jgilaberthere is an api call to delete arqs using the instance uuid Ie1894311b0e384ab52b1b3dfe0eb50618eef6c9f14:41
jgilabersorry pasted wrong14:41
jgilaberhttps://docs.openstack.org/api-ref/accelerator/#delete-accelerator-requests-by-instance-uuid14:42
sean-k-mooneyits only vaild for a user to delete an arq if its not asssocated with an instance14:42
sean-k-mooneyack let me review that seperatly14:42
sean-k-mooneythis clie shoudl be callign the other one14:42
sean-k-mooney/v2/accelerator_requests?arqs={accelerator_request_uuid} not /v2/accelerator_requests?instance={instance_uuid}14:43
sean-k-mooneyboth exist and the nameing convetion of osc woudl requrie we call the first path14:43
sean-k-mooneyto call the second we woudl need --instance or similar14:43
sean-k-mooneybut the secodn is invlid from my persective14:43
sean-k-mooneyim saying that as a nova core not a cybrog one14:44
bogdando[m]the original use case seemed to be remove all arqs for a running instance14:44
sean-k-mooneyyep that not supproted by nova14:45
jgilaberfor what is worth the usage test of openstack accelerator arq delete is clear enough 14:45
jgilaberusage: openstack accelerator arq delete [-h] <uuid> [<uuid> ...14:45
jgilaber  <uuid>  UUID(s) of the accelerator_request(s) to delete.14:45
sean-k-mooney+114:45
jgilaberno mentions of instance uuids14:45
sean-k-mooneythat confims the expecation in the bug reprot is incorect14:45
jgilaberyep, we can close it14:46
bogdando[m]at least for now14:46
sean-k-mooneyi willl likely need to adress /v2/accelerator_requests?instance={instance_uuid} in my srback spec14:46
sean-k-mooneybut i suspect wwe will need to remove it in a future microversion14:46
sean-k-mooneyill  dig into the history of it14:47
bogdando[m]may be it makes sense in future, as housekeeping things, like defragment vram to allow repartitioning14:47
sean-k-mooneylets not specualte on it now14:47
bogdando[m]ack14:48
sean-k-mooneyjgilaber: do you want to leave a comment on the bug or will i14:48
sean-k-mooneyi woudl liek to link to the doc sting when closing it14:48
chandankumarhttps://github.com/openstack/cyborg/commit/1f9a808f1d965063a2744c6ea6ddeb76a5d57061 adds some does around second option14:48
jgilaberI can do it14:48
sean-k-mooneyjgilaber: ++14:48
sean-k-mooney chandankumar  nova only supprots 1 ARQ today14:49
sean-k-mooneyi think this was added in preperation for cybrog device via neuton ports14:50
sean-k-mooneysince that fits the time frame but that never happend in nova or neutron14:50
chandankumarok14:51
sean-k-mooneyshall we move on to the next bug or finish early?14:51
chandankumaryes14:51
chandankumar#link https://bugs.launchpad.net/openstack-cyborg/+bug/1955286 cyborg cmd exec meeting keystone authentication fail14:51
sean-k-mooneythat looks like a deployment isseu14:53
sean-k-mooneythey are gettign a 401 form the micoversion endpoint14:53
sean-k-mooneywhich shoudl not requrie auth14:53
sean-k-mooneyso without more info mation i think we close this as incompl.ete14:54
sean-k-mooneybases on the path this looks liek a devstack deployment not production14:54
sean-k-mooney/opt/stack/data/CA/int-ca/ca-chain.pem14:54
sean-k-mooney/opt/stack/data is where devstack store service state14:54
chandankumarI will take care of this one14:55
sean-k-mooneyack14:55
chandankumarthere is one more bug, but I will move it to next meeting14:55
sean-k-mooneywe could see if we have the same bevhior in our local devstack but i dont think this si a real issue14:56
sean-k-mooneyack14:56
chandankumarmoving to last topic14:56
chandankumar#topic volunteer for next meeting14:56
rlandyI'll do it since I skipped my last round14:56
chandankumarthanks rlandy !14:56
chandankumar#topic open discussion14:56
chandankumarAnyone wants to bring any topic before closing the meeting.14:57
chandankumarif not, then let me close the meeting14:58
chandankumarthank you all for joining the meeting!14:58
chandankumar#endmeeting14:59
opendevmeetMeeting ended Tue Mar 10 14:59:02 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.2026-03-10-14.02.html14:59
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/cyborg/2026/cyborg.2026-03-10-14.02.txt14:59
opendevmeetLog:            https://meetings.opendev.org/meetings/cyborg/2026/cyborg.2026-03-10-14.02.log.html14:59
« Friday, 2026-03-06 Index Wednesday, 2026-03-11 »

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