Tuesday, 2026-06-23

opendevreviewMerged openstack/nova master: Fix cover env HTML testr results missing unit tests  https://review.opendev.org/c/openstack/nova/+/99315100:50
opendevreviewSeyeong Kim proposed openstack/nova master: Avoid serializing orphaned instance in log  https://review.opendev.org/c/openstack/nova/+/99257400:59
opendevreviewBrian Haley proposed openstack/nova master: Change neutron API calls to use project_id  https://review.opendev.org/c/openstack/nova/+/99432101:44
opendevreviewGhanshyam Maan proposed openstack/nova master: DNM: Test removal of enforce_scope  https://review.opendev.org/c/openstack/nova/+/98647901:45
gmaannova cores, these are the changes for the nova to merge before oslo.policy enforce_scope flag is removed, please check https://review.opendev.org/q/hashtag:%22remove-enforce-scope-flag%22+AND+hashtag:%22nova%22+status:open01:50
*** bauzas7 is now known as bauzas01:54
opendevreviewJoan Gilabert proposed openstack/nova master: move compile earlier  https://review.opendev.org/c/openstack/nova/+/95051609:17
opendevreviewJoan Gilabert proposed openstack/nova master: Add mtty support for testing fake mdevs  https://review.opendev.org/c/openstack/nova/+/89810009:17
opendevreviewJoan Gilabert proposed openstack/nova master: Rename vtpm job and add mtty support for vgpu test  https://review.opendev.org/c/openstack/nova/+/92214009:17
gibisean-k-mooney: could we land https://review.opendev.org/c/openstack/nova/+/936093 ? There was a bit of back an forth on IRC and in the review and then we hit some deadline and then we forgot about the patch. 10:13
gibiwe got yet another downstream report around interface MTUs and live migration10:13
gibiI know this is not the full fix but it is a start10:14
sean-k-mooneygibi: o/10:20
sean-k-mooneyamy proiably i need to load context10:20
sean-k-mooneygibi: so the issue that is not adressed is you cant change the mtu on live migration and we need to handel that10:22
sean-k-mooneybut ya we can likely proceed10:22
sean-k-mooneyim trying to see if i had a specific concern in the past beyond that i.e. somethign that woudl be hard to fix later beyond it being a partial fix10:23
sean-k-mooneygibi: so looking at my old comment the main ask i had was to add a functional regression test10:25
sean-k-mooneygibi: so that would still be nice to have but i dont have an objection to the approch to fixign it10:26
gibiOK. Lets see If I get some time to add a functional reproducer. If so I will ping you that patch 10:28
gibithanks for confirming that the direction is OK10:28
sean-k-mooneyif you want i can just have ai create one now and i can adress my own request10:28
gibiif you have the capacit I'm OK with that10:30
opendevreviewJoan Gilabert proposed openstack/nova master: Report OWNER_NOVA trait on resource providers  https://review.opendev.org/c/openstack/nova/+/99429910:37
opendevreviewAnton Iacobaeus proposed openstack/nova master: libvirt: Add Intel TDX host capability detection  https://review.opendev.org/c/openstack/nova/+/99442110:55
sean-k-mooneybauzas: while i think of it im now +2 on https://review.opendev.org/c/openstack/nova/+/988777 if you want to readd yours11:14
opendevreviewJoan Gilabert proposed openstack/nova master: move compile earlier  https://review.opendev.org/c/openstack/nova/+/95051611:18
opendevreviewJoan Gilabert proposed openstack/nova master: Add mtty support for testing fake mdevs  https://review.opendev.org/c/openstack/nova/+/89810011:18
opendevreviewJoan Gilabert proposed openstack/nova master: Rename vtpm job and add mtty support for vgpu test  https://review.opendev.org/c/openstack/nova/+/92214011:18
opendevreviewAnton Iacobaeus proposed openstack/nova master: libvirt: Add Intel TDX host capability detection  https://review.opendev.org/c/openstack/nova/+/99442111:29
opendevreviewMasanori Ueno proposed openstack/nova master: Add regression test for bug #2153425  https://review.opendev.org/c/openstack/nova/+/98877713:13
opendevreviewMasanori Ueno proposed openstack/nova master: live-migration: Ensure limits are set before _call_livem_checks_on_host  https://review.opendev.org/c/openstack/nova/+/99021213:16
*** ykarel_ is now known as ykarel13:26
opendevreviewsean mooney proposed openstack/nova master: Fix network mtu metadata disappears after attaching  https://review.opendev.org/c/openstack/nova/+/93609313:32
opendevreviewsean mooney proposed openstack/nova master: Add regression tests for interface attach MTU bug  https://review.opendev.org/c/openstack/nova/+/99443713:32
opendevreviewsean mooney proposed openstack/nova master: Fix network mtu metadata disappears after attaching  https://review.opendev.org/c/openstack/nova/+/93609313:33
sean-k-mooneygibi: that also needed a rebase but ^ should cover all my nits13:33
gibisean-k-mooney: thanks 13:35
sean-k-mooneyi didnt quite have tim eto push that before my meeting as i was waiting for tox runs13:35
opendevreviewErlon R. Cruz proposed openstack/nova master: Adding regression test to reproduce bug 1741364  https://review.opendev.org/c/openstack/nova/+/99359713:51
opendevreviewErlon R. Cruz proposed openstack/nova master: Add RBD XML update functionality for migration process  https://review.opendev.org/c/openstack/nova/+/97403213:51
*** raukadah is now known as chandankumar14:02
opendevreviewJoan Gilabert proposed openstack/nova master: Report OWNER_NOVA trait on resource providers  https://review.opendev.org/c/openstack/nova/+/99429914:32
fricklernova-ceph-multistore is failing on unmaintained/2024.1, I don't really want to bug people about it much, but maybe if someone knows an easy fix we can apply that instead of just dropping the job https://zuul.opendev.org/t/openstack/build/570a4539d0f24452ab6c5f9d75e3450a16:09
melwitthm, "oslo_config.cfg.NoSuchOptError: no such option replication in group [volume-feature-enabled]" must be something like, cinder-tempest-plugin is not branched but maybe tempest on unmaintained/2024.1 is pinned cc gmaan 16:25
gmaanyes, i think it is pinned16:28
opendevreviewTakashi Kajinami proposed openstack/nova master: compute: Validate SEV/SEV-ES support at start up  https://review.opendev.org/c/openstack/nova/+/99434216:28
gmaanyes tempest 45.0.0 is used and it does not have 'replication ' config option https://github.com/openstack/tempest/blob/45.0.0/tempest/config.py16:29
gmaanso test should be skipped in that case 16:29
gmaanwe need to pin cinder-tempest-plugin also so that it will not pick the latest tests from it16:31
melwittmakes sense16:37
gmaanwill propose fix after meeting16:39
gmaanmelwitt: sean-k-mooney you might be interested in these. nova side work needed for oslo.policy enforce_scope flag removal https://review.opendev.org/q/hashtag:%22remove-enforce-scope-flag%22+AND+hashtag:%22nova%22+status:open16:39
melwittgmaan: ack will look16:41
gmaanthanks16:41
opendevreviewTakashi Kajinami proposed openstack/nova master: libvirt: Ignore SEV-ES when SEV-SNP is detected  https://review.opendev.org/c/openstack/nova/+/99434316:55
sean-k-mooneygmaan: oh cool17:01
sean-k-mooneylooks pretty mechnaical17:02
sean-k-mooneywhile while somewhat tedious to review is at elast not overly compex17:02
opendevreviewLajos Katona proposed openstack/nova master: Add regression test to repoduce bug 1854844  https://review.opendev.org/c/openstack/nova/+/99177217:03
sean-k-mooneyim slightly suprised the nova change depens on the placment one but i guess that is down to the fixture?17:04
gmaansean-k-mooney: I think yes, somewhere it was failing and I need to do placement one first17:08
sean-k-mooneyi am sumwhat suprised that we need to update teh gabit tests17:14
sean-k-mooneyspecific for system scoped rp creation17:15
sean-k-mooneygiven i was not expecting system scoped roels to really be valid based on https://github.com/openstack/placement/blob/master/placement/policies/resource_provider.py17:15
sean-k-mooneygiven admin or service is "role:admin or role:service"17:16
sean-k-mooneybut i guess when scope enforcment is disabeld we are merging system and doamin roels into the set of token roles?17:17
sean-k-mooneygmaan: basiclly im asking did we change the behvior or was gabbi just not emululatign this properly17:17
sean-k-mooneyfor this change https://review.opendev.org/c/openstack/placement/+/987568/4/placement/tests/functional/gabbits/aggregate-legacy-rbac.yaml#5517:18
gmaansean-k-mooney: yes, it is same as nova test change. the update is to the tests which were testing scope disable and system_admin can do the things17:19
gmaansean-k-mooney: those are needed to create the resource for other tests like update test, for example $ENVIRON[[RP_UUID]17:20
gmaanpreviously system_admin was allowed in those tests and it created the resouce17:20
gmaani thought of remove the legacy tests but till we have the legacy policy, we should keep them tested17:20
gmaanBasically, it remove the scope disable tests and update a few that were dependent on the test resource creation by scope disable tests17:23
opendevreviewTakashi Kajinami proposed openstack/nova master: compute: Validate SEV/SEV-ES support at start up  https://review.opendev.org/c/openstack/nova/+/99434217:36
opendevreviewTakashi Kajinami proposed openstack/nova master: libvirt: Ignore SEV-ES when SEV-SNP is detected  https://review.opendev.org/c/openstack/nova/+/99434317:36
opendevreviewTakashi Kajinami proposed openstack/nova master: Fix wrong mock for properties  https://review.opendev.org/c/openstack/nova/+/99450517:36
opendevreviewTakashi Kajinami proposed openstack/nova master: Drop dependency on [libvirt] num_memory_encrypted_guests  https://review.opendev.org/c/openstack/nova/+/99450617:36
opendevreviewSahana proposed openstack/nova master: compute: remove redundant exception args from LOG.exception calls  https://review.opendev.org/c/openstack/nova/+/99452218:37
opendevreviewmelanie witt proposed openstack/nova master: Add non-voting IPv6 ceph live migration job  https://review.opendev.org/c/openstack/nova/+/98462318:57
opendevreviewmelanie witt proposed openstack/nova master: Add non-voting IPv6 ceph live migration job  https://review.opendev.org/c/openstack/nova/+/98462319:12
gmaanfrickler: melwitt this will fix the nova-ceph-multistore job on 2024.1, we cna see devstack-plugin-ceph-tempest-py3  passing there https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/99450319:31
gmaanbut it seems gate is broken for doc job there so I am leaving this until someonw fix the gate. cannot spend more time on those. That is one of the reason i try not to touch unmaintained branches as they need lot of gate fixes before one merge :)19:32
gmaanCC: elodilles just to make you aware about ceph job fix19:32
fricklergmaan: thanks, I can try to look at the docs issue. and yes I can understand your reluctance, thx for checking this anyhow19:49
melwittgmaan: ah makes sense19:57
opendevreviewMerged openstack/nova master: move compile earlier  https://review.opendev.org/c/openstack/nova/+/95051620:01
opendevreviewmelanie witt proposed openstack/nova master: Add non-voting IPv6 ceph live migration job  https://review.opendev.org/c/openstack/nova/+/98462320:08
erlonsean-k-mooney: hey Sean,  can you take a look in the RBD fix we discuss the other day? See if thats how you wanted the regression tests, please let me know if I'm missing anything: https://review.opendev.org/c/openstack/nova/+/97403220:12
sean-k-mooneyerlon: yes i started lookign at that elaire today20:14
sean-k-mooneyi have the regression test open i na tab somewhere20:14
sean-k-mooneyerlon: lookign at ti brifly yes it looks more or less correct ill try and review it proeprly but probaly tomorrow20:16
erlonnice, thanks20:16
sean-k-mooneyi see melwitt updated https://review.opendev.org/c/openstack/nova/+/984623/21 as well so we shoudl get som new reuslt of the ipv6 test in a few hours20:17
melwittyeah fingers crossed it doesn't run into another issue. this is the first time I try it with depends-on patches proposed to the proper repos. last fail was for database host using ipv4 address. I'll keep iterating on it20:19
sean-k-mooneyis there a reason you made the ipv6 connectivity opt in instead of opt out?20:25
sean-k-mooneywe often have ipv6 in the jobs20:25
sean-k-mooneyeven if its not the ipv6 job20:25
sean-k-mooneywe just dont use those adress for the services by default20:25
melwittno reason other than to keep existing behavior the same20:25
sean-k-mooneyack20:26
sean-k-mooneyim not sure which is the better default i just said i would ask20:26
sean-k-mooneyincasue it was because you hit a bug or something i was not aware of20:26
melwittoh yeah. not yet anyway haha20:26
sean-k-mooneyso if you respin you might want to mention ipv6 in the readme for that roles https://opendev.org/zuul/zuul-jobs/src/commit/6c0c643b8cffa1f61cf7af34f29d8a7e4d1bb114/roles/multi-node-bridge/README.rst20:27
sean-k-mooneymelwitt: they do not use molecule in that repo20:29
sean-k-mooneybut you may also want to exted https://opendev.org/zuul/zuul-jobs/src/commit/6c0c643b8cffa1f61cf7af34f29d8a7e4d1bb114/test-playbooks/multinode/multi-node-bridge.yaml20:29
sean-k-mooneyto test the ipv6 functionlatiy20:29
sean-k-mooneybut i dont really review code in that repo oftent so im not sure what the normal expecation are20:30
melwittgood point, I didn't know about those20:30
sean-k-mooneyfungi and clarkb are inovled so i knwo there woudl be tests but i am not very familar with how they are setup but i guess this is just usign zuul to run the test playbooks directly20:31
sean-k-mooneylooks like tey are using tox and unittest to run them at a quick glance20:31
sean-k-mooneyalthough that may just be for the inline moduels20:32
fungiyeah, clarkb can probably explain it better since he wrote a lot of it, but the idea initially was to have an abstract foundation for multi-node devstack jobs to build on20:32
sean-k-mooneyfungi: for the most part it works quite well. fungi context is melaine is adding supprot for multi node ipv620:34
fungiin particular, avoiding dependence on the network topology available to the test nodes themselves and setting up an encapsulated layer 2 network between them that would provide consistency20:34
sean-k-mooneyso we can test live mgiration with ipv6 ceph20:34
fungijust to be clear, you want to tunnel v6 over the bridge (that shouldn't be a problem but maybe we baked in assumptions)20:35
sean-k-mooneyfungi: yep that the idea. melwitt has some devstack josb mostly workign with pre-playbooks in nova and is now tryign to do it properly in zuul-jobs and devstack20:36
melwittwith barely any knowledge 😩 it just failed again. obviously I must have lost something while splitting out from the pre-playbooks into zuul-jobs. will have to figure this out20:38
fungiright, afaik we didn't want that role to restrict what kind of traffic you can send over it, modulo obvious mtu adjustments20:39
sean-k-mooneyyou may need to set  bridge_configure_ipv6_address: true in the subnodes as well20:39
sean-k-mooneyhttps://review.opendev.org/c/openstack/nova/+/984623/21/.zuul.yaml#41620:39
fungialso i think there may be some limitations or at least limitations carrying broadcast and multicast traffic on it20:39
sean-k-mooneyfungi: its using vxlan20:40
fungier, or at least inefficiencies20:40
sean-k-mooneyso it shoud be fine20:40
sean-k-mooneyya its a partial mesh using ovs20:40
sean-k-mooneyso all multicast becoms bradcast which si copied intrenlly t o all spokes in the mesh20:40
fungii think vxlan relies on multicast groups underneath to carry broadcasts though? and our cloud providers don't let the multicast traffic through20:40
sean-k-mooneyso its a littele inefficnet but it shoudl work20:40
fungiah okay20:41
sean-k-mooneyfungi: not with ovs20:41
sean-k-mooneybut with linux bridge yes20:41
sean-k-mooneyovs only supprot unicast tunnels20:41
fungiyou are far more familiar with all of this than i am20:41
sean-k-mooneybut that actully makes it more portable since you do not rely on the cloud provdier supproting multicast grouping at the fabic level20:41
fungiall i can really say is afaik we didn't intentionally omit support for sending ipv6 packets over the bridge, so if it's not working i'd consider that a bug worth fixing20:42
sean-k-mooneyit comes form addign ovs-dpdk suppreot neutron/nova and maintianing it for years:)20:42
melwittsean-k-mooney: oh yeah, that makes sense. in pre, I was doing it for hosts: all but didn't put under both20:42
fungi`hosts: all` should run the same task on every test node (but not the zuul executor, which needs `hosts: localhost` instead, not really relevant to this playbook anyway)20:43
sean-k-mooneyfungi: i have also impletend my own version fo the role 2 before although the most recent version i got ai to do :)20:45
sean-k-mooneyfungi: https://github.com/SeanMooney/ard/tree/master/ansible/roles/ard_multinode_bridge does basically the same thign but it uses linux bridge and gre tunnels to avoid installign ovs20:46
sean-k-mooneyi used to just use the zuul jobs one20:46
sean-k-mooneyim not entrily sure im happy with how the new version works but i was tryign to make it compatiabe with our operator based isntaller and i dint want to fight with its installation and management of ovs20:48
sean-k-mooneythat fact that ovs survies reboot makes the ovs based approch we use upstream less complex in many ways20:49
sean-k-mooneymy the way the mtu caluation does have to be adjuscted for ipv6 but the role aslready subtratc 70 bytes which si enough to have ipv6 geneve trafic go over the bridge20:56
sean-k-mooneymelwitt: https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/multi-node-bridge/tasks/common.yaml#L103-L13920:56
melwitthm ok20:58
sean-k-mooneytechinaly the 70 bytes of overad is the bare minium for geneve to work over the bridge20:59
sean-k-mooneyand it shoudl work for neutron in general20:59
clarkbits not geneve its vxlan iirc20:59
clarkband that is why there isn't ipv621:00
sean-k-mooneyi use a 100 byte safty buffer in my verions21:00
fungisean-k-mooney: yeah, we originally used gre i believe, but it isn't allowed in all out providers21:00
clarkbthere is/was work to convert it to geneve and drop the use of the ansible bridge module21:00
sean-k-mooneyclarkb: yes the overly is21:00
fungiat least in some openstack clouds, the kernel conntrack plugin for gre isn't running21:00
sean-k-mooneyclarkb: actully that not the part the matters21:00
sean-k-mooneywe have other code in devstack playbooks to calulate teh neutron mtu config21:00
sean-k-mooneythat the one that need to accomidate geneve whwen ovn is used21:01
clarkbhttps://review.opendev.org/c/zuul/zuul-jobs/+/95939321:01
sean-k-mooneyfungi: ya gre is not an ip protocol so its often blockded21:01
clarkbyes the original implementation used gre and linux bridges21:01
sean-k-mooneyi depebated about gre or vxlan and for 8 bytpes of mtu vxlan is definly more portabl since its just udp form the network layer21:02
clarkbthat didn't work because we couldn't get gre to pass even with wide open security groups so we switched to vxlan as it is udp21:02
sean-k-mooneyclarkb: :) yep i remember21:02
clarkband now ansible has dropped supprot for much of the machinery that we relied on to setup vxlan and ovs so we're switching back to linux bridge and taking the opportunity to use geneve in that change21:02
clarkbthat work has clearly stalled out and probably should be picked up again. I think it is very close if not ready21:03
clarkband the reason we didn't support ipv6 is taht netiher vxlan nor gre could support it iirc21:04
clarkbor maybe it was just vxlan that couldn't21:04
sean-k-mooneyboth can21:04
clarkbI think vxlan can now21:04
sean-k-mooneyat both the underlay and over the mesh21:05
clarkbbut it was a v4 only protocol21:05
sean-k-mooneyit dependied on the kernel version21:05
sean-k-mooneyand ovs version21:05
clarkbin any case if you want to make changes like this getting 959393 in first is probably a prereq21:05
sean-k-mooneyvxlan suppoted it but we could supprot it on all distos we supproted if i recll21:05
sean-k-mooneywell it does not stricly ned the refacto to linux bridge21:05
sean-k-mooneybut that cerntly viable21:06
clarkbit does not but I will nto review a chnge that creatse a merge conflict21:06
clarkbmaybe other people are willing to review it and land it21:06
opendevreviewmelanie witt proposed openstack/nova master: Add non-voting IPv6 ceph live migration job  https://review.opendev.org/c/openstack/nova/+/98462321:06
sean-k-mooneythat is one of my version of that https://github.com/SeanMooney/ard/tree/master/ansible/roles/ard_multinode_bridge21:06
sean-k-mooneyi have a much older version too somehwere21:06
clarkbthe current code only works because I vendored the entirety of the ansible module that is no longer supported into that role21:06
sean-k-mooneyya i replaced all of tha twith a systemd unit file that calls a bashcript21:07
clarkbwe really need to stop using that stuff and leaving it as is is fine. But if you want to make changes then we should do the right thing to fix the underlying ansible stuff21:07
sean-k-mooneyhttps://github.com/SeanMooney/ard/blob/master/ansible/roles/ard_multinode_bridge/templates/ard-bridge-setup.sh.j221:07
clarkband that work is basically done so there isn't any reason not to21:07
sean-k-mooneyclarkb: ya https://review.opendev.org/c/zuul/zuul-jobs/+/959393 cerntely cleans up a lot21:08
clarkbrereading things it stalled beacuse people using ovs as a hard dependency weren't properly installing/settting up that dep and relying on that role instead21:08
clarkbI think we can basically say sorry you need to fix that and move forward21:09
clarkbwhich is also what I basically said back when I was trying to move this forward21:09
sean-k-mooneyya. is ther a reason it swapt to geneve?21:09
sean-k-mooneyi know it says "This is a more modern and flexible approach than the21:09
clarkbI think because geneve is the future and more flexible21:09
sean-k-mooneyprevious vxlan implementation."21:09
clarkbright vxlan had a bunch of gotchas like the ipv6 support or lack of it21:10
clarkbgeneve exists to solve all those problems with a new protocol21:10
clarkbso we should use it since those problems hit us previously21:10
sean-k-mooneythe main one that geneve solve is actully fragmentation21:10
sean-k-mooneyvxlan explity does not allow fragmentation21:10
sean-k-mooneyof the vxlan packet21:10
sean-k-mooneygeneve i belvie does21:10
sean-k-mooneywhich makes it more flexible acorss l3 subnets21:11
sean-k-mooneyclarkb: but i guess i agree building on that patch woudl be good for the reason above21:12
sean-k-mooneyanyway dinner is ready o/21:12
clarkbya I think the next step is to followup with ykarel to do what we can to properly capture ovs dependencies then proceed with the switch in zuul-jobs then add ipv6 support21:13
clarkbprobably don't even need to make ipv6 support optional just add it21:13
clarkbykarel: can you followup on https://review.opendev.org/c/zuul/zuul-jobs/+/959393 and indicate if properly managing jobs' dependency on ovs si a possibility? That really shouldn't be something we rely on a side effect from a base job role to handle21:14
opendevreviewmelanie witt proposed openstack/nova master: Add non-voting IPv6 ceph live migration job  https://review.opendev.org/c/openstack/nova/+/98462322:00
opendevreviewmelanie witt proposed openstack/nova master: Add non-voting IPv6 ceph live migration job  https://review.opendev.org/c/openstack/nova/+/98462322:51
opendevreviewmelanie witt proposed openstack/nova master: Add non-voting IPv6 ceph live migration job  https://review.opendev.org/c/openstack/nova/+/98462323:52

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