Tuesday, 2022-07-26

auniyallogs from nova/db/main/api should come in devstack@n-cond-cell1, super-cond or cpu ?05:31
auniyalit should come under devstack@n-super-cond, but exceptions are not coming05:41
bauzasgood morning Nova07:48
gibio/07:49
bauzasI just sent an email 07:55
bauzastl;dr: HOLD YOUR RECHECKS07:55
gibiI see a nice buzz around the issue in the bug report08:07
bauzaslike I said in -neutron, I'm looking at gerrit now based on mlavalle's comment08:08
gibiI did that already and commented on the bug, but feel free to double check08:09
opendevreviewSylvain Bauza proposed openstack/nova master: DNM: Avoid os-vif 3.0.0  https://review.opendev.org/c/openstack/nova/+/85099808:17
bauzasgibi: ^08:17
gibibauzas: thanks08:18
sean-k-mooney[m]why?08:53
sean-k-mooney[m]what change in os-vif do you think is related08:53
sean-k-mooney[m]the trunk bridge patch?08:55
sean-k-mooney[m]https://github.com/openstack/os-vif/commit/75b290fb2a8f706583e0c12c5c5a4c0fc80e6481 ?08:56
sean-k-mooney[m]if its related to that then the but is in neutron08:58
bauzassean-k-mooney: we're discussing it alot in the -neutron room09:03
bauzasbasically, neutron wanted to defer the trunk creation logic to os-vif but since we have rolling upgrades, looks like we now fail09:04
bauzasit looks to me the upgrade approach was invalid and we should have waited nova to be fully upgraded 09:04
opendevreviewSylvain Bauza proposed openstack/nova master: Block os-vif 3.0.0  https://review.opendev.org/c/openstack/nova/+/85099809:07
sean-k-mooney[m]ill hop over09:07
sean-k-mooney[m]but the nova and os-vif code is right i think09:07
sean-k-mooney[m]but they may have dropped the bridge cleanup code to soon in neutron09:08
bauzassean-k-mooney: it seems so yeah09:08
bauzassome logic was possibly modified in Neutron09:09
bauzasthis would explain why nova-next and grenade were failing09:09
bauzasfor the ovs-hybrid-plug job, I don't k09:09
*** tosky_ is now known as tosky09:10
opendevreviewSylvain Bauza proposed openstack/nova master: DNM for testing drop of os-vif to 2.8.0  https://review.opendev.org/c/openstack/nova/+/85100609:41
opendevreviewTakashi Natsume proposed openstack/nova-specs master: Create specs directory for Antelope  https://review.opendev.org/c/openstack/nova-specs/+/85100709:48
opendevreviewsean mooney proposed openstack/os-vif master: [WIP] make os-vif jobs multinode and enable trunk testing  https://review.opendev.org/c/openstack/os-vif/+/85101110:26
whoami-rajatdansmith, sean-k-mooney[m] hey, just a reminder to request to review the series of rebuilding volume backed instance, would like to get it in Zed release https://review.opendev.org/c/openstack/nova/+/820368/10:30
bauzasso, https://review.opendev.org/c/openstack/nova/+/851006 has a -1 but not because of the os-vif issue12:16
ratailorbauzas, could you please provide your feedback on https://review.opendev.org/c/openstack/nova/+/844418 ?12:17
bauzasratailor: for the moment, I'm working on another CI issue12:18
bauzasbut ok12:18
ratailorbauzas, no problem. whenever you have time. Thanks!12:18
opendevreviewribaudr proposed openstack/nova master: [WIP] Attach Manila shares via virtiofs (objects)  https://review.opendev.org/c/openstack/nova/+/83940113:02
opendevreviewribaudr proposed openstack/nova master: [WIP] Attach Manila shares via virtiofs (manila abstraction)  https://review.opendev.org/c/openstack/nova/+/83119413:02
opendevreviewribaudr proposed openstack/nova master: [WIP] Attach Manila shares via virtiofs (drivers)  https://review.opendev.org/c/openstack/nova/+/83309013:02
opendevreviewribaudr proposed openstack/nova master: [WIP] Attach Manila shares via virtiofs (api)  https://review.opendev.org/c/openstack/nova/+/83683013:02
opendevreviewribaudr proposed openstack/nova master: [WIP] Bump compute version and check shares support  https://review.opendev.org/c/openstack/nova/+/85049913:02
opendevreviewribaudr proposed openstack/nova master: [WIP] Add metadata for shares  https://review.opendev.org/c/openstack/nova/+/85050013:02
opendevreviewribaudr proposed openstack/nova master: [WIP] Add instance.share_attach notification  https://review.opendev.org/c/openstack/nova/+/85050113:02
opendevreviewribaudr proposed openstack/nova master: [WIP] Add instance.share_detach notification  https://review.opendev.org/c/openstack/nova/+/85102813:02
opendevreviewribaudr proposed openstack/nova master: [WIP] Add shares to InstancePayload  https://review.opendev.org/c/openstack/nova/+/85102913:02
*** blarnath is now known as d34dh0r5313:13
bauzasfwiw https://lists.openstack.org/pipermail/openstack-discuss/2022-July/029725.html13:14
*** dasm|off is now known as dasm13:19
opendevreviewKashyap Chamarthy proposed openstack/nova master: Add a workaround to skip hypervisor version check  https://review.opendev.org/c/openstack/nova/+/85103414:07
opendevreviewKashyap Chamarthy proposed openstack/nova master: Add a workaround to skip hypervisor version check on LM  https://review.opendev.org/c/openstack/nova/+/85103414:14
kashyapgibi: sean-k-mooney: --^ When you get a minu14:17
gibikashyap: we need a launchpad bug for tracking and a releasenotes 14:19
gibiotherwise looks good to me14:19
kashyapgibi: Ah, yes; rel-no15:18
kashyapGonna do it15:18
kashyapgibi: I hope you're cool with the what/why format here :) - https://bugs.launchpad.net/nova/+bug/198285315:23
sean-k-mooneywe dont actully enforce a format upstream so sure15:24
gibikashyap: it a bit strange as a bug report as it does not focus on the faulty behavior, but as sean-k-mooney said it is OK15:24
sean-k-mooneyalthough that makes it sound more like a feature then a bug15:24
kashyapsean-k-mooney gibi Yeah, I can describe a buggy behaviour from OSP usage15:25
* kashyap goes to edit15:25
sean-k-mooneygibi: ya i had the same tought, framing it as what/why focuses on solution not the problem15:25
sean-k-mooneywell ter isnt a but in nova really15:26
gibibut I understand if we don't want to embarrass ourselves with the actual fault we made downstream ;)15:26
kashyapThe "what" can also describe a problem space15:26
sean-k-mooneybut there are day 2 issues that people might hit that his will help wiht15:26
kashyapI was just too lazy to add it15:26
sean-k-mooneyso im ok to tack it as a UX bug or operator nice to have15:26
sean-k-mooneygibi: well im expecting we will close our downstream bz as not a bug honestly15:28
kashyapYeah, that's why the "why" describes an example use-case, instead of an elaborate mistake of a specific OpenStack distro :)15:28
sean-k-mooneybut we will se we might have to kep it open for backport reaons15:28
sean-k-mooneyand doc text ectra15:28
gibiack15:28
sean-k-mooneyjust due to the constratints of the tooling so ya15:28
sean-k-mooneyso if there is no objection im going to trage this as a wishlist bug15:29
gibiworks for me15:31
sean-k-mooneyhttps://bugs.launchpad.net/nova/+bug/1982853 cool done15:34
gibisean-k-mooney: thanks15:38
bauzasreminder : nova meeting in 20 mins 15:39
bauzas#startmeeting nova16:00
opendevmeetMeeting started Tue Jul 26 16:00:04 2022 UTC and is due to finish in 60 minutes.  The chair is bauzas. Information about MeetBot at http://wiki.debian.org/MeetBot.16:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.16:00
opendevmeetThe meeting name has been set to 'nova'16:00
bauzashello folks16:00
gibio/16:00
bauzaswho's around ?16:01
bauzas#link https://wiki.openstack.org/wiki/Meetings/Nova#Agenda_for_next_meeting16:01
elodilleso/16:02
bauzasok let's start, people will join16:02
bauzas#topic Bugs (stuck/critical) 16:02
bauzas#info No Critical bug16:02
bauzas#link https://bugs.launchpad.net/nova/+bugs?search=Search&field.status=New 10 new untriaged bugs (+0 since the last meeting)16:02
bauzas#link https://storyboard.openstack.org/#!/project/openstack/placement 27 open stories (+0 since the last meeting) in Storyboard for Placement 16:02
bauzas#info Add yourself in the team bug roster if you want to help https://etherpad.opendev.org/p/nova-bug-triage-roster16:02
Ugglao/16:03
bauzasgiven I'll be on PTO after the next meeting, I can be the next bug baton owner for this week16:03
bauzaselodilles: sorry, but I'll take it for this week :p16:03
elodilles:'(16:03
elodillesno problem of course :D16:03
bauzashah16:04
gibielodilles: you can have mine if you want ;)16:04
bauzas#info Next bug baton is passed to bauzas16:04
bauzasthat's it for bugs16:04
bauzaswe'll discuss the CI outage in the next topic16:04
bauzasany other bug to discuss ?16:04
bauzaslooks not16:05
bauzas#topic Gate status 16:05
bauzas#link https://bugs.launchpad.net/nova/+bugs?field.tag=gate-failure Nova gate bugs 16:05
bauzasso,16:05
bauzasyou maybe have seen my emails16:05
bauzasthe gate is blocked at the moment16:05
bauzas#link https://bugs.launchpad.net/nova/+bug/1940425 gate blocker16:06
bauzasI triaged the bug status to invalid for nova as this is an os-vif/neutron issue16:06
bauzas#link https://lists.openstack.org/pipermail/openstack-discuss/2022-July/029725.html16:07
bauzasthe ML thread ^16:07
bauzasplease don't recheck your changes until we merge an os-vif version blocker16:07
bauzasfor 3.0.016:07
bauzasnow, once the os-vif blocker will be merged, we'll still need to find a way to use os-vif 3.0.x 16:08
bauzasdo people want to know more ?16:09
bauzasor can we continue ?16:10
bauzasmmmm ok16:10
bauzasthen, let's continur16:11
bauzasthe periodic job runs16:11
bauzas#link https://zuul.openstack.org/builds?project=openstack%2Fplacement&pipeline=periodic-weekly Placement periodic job status 16:11
bauzas#link https://zuul.openstack.org/builds?job_name=tempest-integrated-compute-centos-9-stream&project=openstack%2Fnova&pipeline=periodic-weekly&skip=0 Centos 9 Stream periodic job status16:11
bauzas#link https://zuul.opendev.org/t/openstack/builds?job_name=nova-emulation&pipeline=periodic-weekly&skip=0 Emulation periodic job runs16:11
bauzasI don't see any problem with all of them ^16:12
chateaulavalways good to see passing jobs16:12
gibi:) 16:12
bauzaswould love to see nova-next and grenade passing too, but meh :)16:12
gibiwe cannot have it all :)16:13
sean-k-mooneyhttps://review.opendev.org/c/openstack/tempest/+/85024216:13
sean-k-mooneyis stilll pending16:13
bauzasyup, i guessed it16:13
sean-k-mooneyto move centos job to perodic weekly only16:13
sean-k-mooneycan we ping anyone in tempest to expidite that16:13
bauzasas I saw we continue to have check and gate jobs for those16:13
bauzassean-k-mooney: maybe gmann ?16:14
sean-k-mooneyperhaps16:14
dansmithyeah gmann and kopecmartin 16:14
sean-k-mooneyilll add them as a review16:14
bauzascool16:15
bauzas#info Please look at the gate failures and file a bug report with the gate-failure tag.16:15
bauzas#info STOP DOING BLIND RECHECKS aka. 'recheck' https://docs.openstack.org/project-team-guide/testing.html#how-to-handle-test-failures16:15
bauzasnext topic then16:16
bauzas#topic Release Planning 16:16
bauzas#link https://releases.openstack.org/zed/schedule.html16:16
bauzas#info Zed-3 is in 5 weeks16:16
bauzas#topic Review priorities 16:17
bauzas#link https://review.opendev.org/q/status:open+(project:openstack/nova+OR+project:openstack/placement+OR+project:openstack/os-traits+OR+project:openstack/os-resource-classes+OR+project:openstack/os-vif+OR+project:openstack/python-novaclient+OR+project:openstack/osc-placement)+label:Review-Priority%252B116:17
bauzasah shit16:17
bauzasI forgot to change the url16:17
bauzasshould create a bit.ly I guess16:17
bauzas#undo16:18
opendevmeetRemoving item from minutes: #link https://review.opendev.org/q/status:open+(project:openstack/nova+OR+project:openstack/placement+OR+project:openstack/os-traits+OR+project:openstack/os-resource-classes+OR+project:openstack/os-vif+OR+project:openstack/python-novaclient+OR+project:openstack/osc-placement)+label:Review-Priority%252B116:18
bauzas#link https://review.opendev.org/q/status:open+(project:openstack/nova+OR+project:openstack/placement+OR+project:openstack/os-traits+OR+project:openstack/os-resource-classes+OR+project:openstack/os-vif+OR+project:openstack/python-novaclient+OR+project:openstack/osc-placement)+(label:Review-Priority%252B1+OR+label:Review-Priority%252B2)16:18
bauzasas you see, we need to merge the 2.91 patch16:18
bauzasbut given the CI outage, we can't do it yet16:19
sean-k-mooneyi think the requirements patch should merge in then next hour16:19
bauzasany review-prio to discuss ?16:19
sean-k-mooneyso likely wont be much longer16:19
bauzassean-k-mooney: oh haven't seen an update yet16:19
sean-k-mooneyit has +w i think16:19
sean-k-mooneyso it should be in the gate as we speak16:19
bauzasnice16:20
bauzaswe already have the nova change for i16:20
bauzasit16:20
bauzascan we move to the stable branches topic ?16:20
sean-k-mooneyyes16:22
bauzascool16:22
bauzas#topic Stable Branches 16:22
* bauzas gives the mic to elodilles16:22
elodilles:)16:22
elodillesyes, so16:23
elodillesactually i haven't updated the stable section, as the state is the same as last week :/16:23
elodilles#info stable branch status / gate failures tracking etherpad: https://etherpad.opendev.org/p/nova-stable-branch-ci16:23
elodillesand train is blocked :/16:23
elodillesi could spend only a little time to look at the train gate issue16:23
elodilles(devstack-gate is failing to install python3-yaml)16:24
sean-k-mooneyis the shell issue still blocked byu another issue16:24
elodillessean-k-mooney: yepp16:24
elodillesif only nova-grenade would fail i'd suggest to set it non-voting 'temporarily'16:24
elodilles(especially as it is stable/train which is quite old)16:24
sean-k-mooneyya that might be the path forward16:25
elodillesbut unfortunately nova-live-migration also fails with the same issue16:25
elodillesso it would mean two jobs :/16:25
sean-k-mooneyfailing to install python3-yaml16:26
elodillesthough the good part would be to have less intermittent failure to catch :P16:26
sean-k-mooneyis it a dist tools failure16:26
sean-k-mooneyi.e. pip refusing to upgrade it16:26
elodillessean-k-mooney: i don't know as it gets timed out16:26
elodillesand we don't have logs either :/16:26
sean-k-mooneybecasue python3-yaml is one of those packages i alway unitlly form the disto before i run devstack16:26
sean-k-mooneyack16:26
elodillesi mean only the jobs-output.txt16:26
elodillesthe last line we see is the processing triggers from libc-bin16:27
elodillesthen it hangs for 2 hrs16:27
sean-k-mooneyoh hum + /opt/stack/new/devstack-gate/functions.sh:apt_get_install:L71:   sudo DEBIAN_FRONTEND=noninteractive apt-get --assume-yes install python3-yaml16:28
sean-k-mooneyso its explcitly being installed form the distor in that job16:28
elodillesyes, from devstack-gate16:28
sean-k-mooneywhich is the the opicit of what i normally do16:28
elodilleslocally this works fine for me with the same packages (if i checked them correctly)16:29
elodillesso still could not reproduce16:29
elodillesthat's it for what i can tell for now :/16:30
sean-k-mooneyack16:31
sean-k-mooneyit does look like it jut hang proceeign the triggers fo libc-bin16:32
sean-k-mooneyas a workaround we could add a pre playbook to do a full update/upgade of the packages and preinstall that16:32
sean-k-mooneyand see if it helped16:32
chateaulavi wonder if there is some weird package prompt stalling the process16:32
elodillesyes, otherwise we should see some continuation from the devstack-gate script16:32
sean-k-mooneysudo DEBIAN_FRONTEND=noninteractive apt-get --assume-yes install python3-yaml16:33
sean-k-mooneyso its being pulled in form that but16:33
sean-k-mooney--assume-yes and noninteractive16:33
sean-k-mooneyshoudl disable all prompts16:33
sean-k-mooneyanyway we proably can move on16:33
chateaulavill see if I cant dedicate some time this week to take a look. 16:33
elodillesthanks!16:34
bauzascool, moving on then16:34
bauzasand thanks all for discussing it16:34
elodillesfor the workaround suggestion, too, sean-k-mooney 16:34
elodillesbauzas: ++16:35
bauzas#topic Open discussion 16:35
bauzas(bauzas) will be on PTO between Aug 3rd and Aug 29th, who would want to chair the 3 meetings ?16:35
* bauzas needs to visit Corsica16:35
bauzasor do we need to cancel those ?16:35
bauzasI'd prefer the former honestly, given we're on zed-316:36
gibiI think I can take them16:36
bauzasgibi: <316:36
bauzasI'll be back on time before the 3rd milestone16:36
gibi9th, 16th, 23rd16:36
bauzasshould even be around on IRC on Monday but shhhtttt16:36
bauzas#action gibi to chair 3 nova meetings (Aug 9th, 16th and 23rd)16:37
Ugglabauzas, can we define the microversion for virtiofs/manila ?16:38
bauzasI'll chair next week's meeting16:38
bauzasUggla: good question16:38
bauzasI was about to write an email for asking people to look at https://etherpad.opendev.org/p/nova-zed-microversions-plan16:38
bauzasbut given we had the CI outage from last week, I didn't had time to do it16:38
bauzasto be fair, I'll write the email to ask people to propose their microversion changes if they think they're already done16:39
bauzasUggla: for the moment, plan to use 2.9316:40
Ugglaok sounds good to me.16:40
bauzasmelwitt wrote something for last meeting but we hadn't time to discuss about it AFAIR16:41
bauzas (melwitt) I will not be at the meeting today but in case you missed  it, I'm seeking input regarding terminology used in the currently named  "ephemeral encryption" feature spec and patches: https://lists.openstack.org/pipermail/openstack-discuss/2022-July/029546.html.  If you have thoughts on whether the terminology should be changed to  something other than "ephemeral encryption", please comment on the ML  thread or on this pa16:41
bauzastch https://review.opendev.org/c/openstack/nova/+/764486 (thank you gibi for adding a comment 🙂)16:41
bauzassounds a new naming bikeshed \o/16:42
bauzasany opinions about a possible good name instead of "ephemeral encryption" ?16:43
bauzasreminder : I'm terrible at naming things16:43
bauzassean-k-mooney: to be fair, we name "ephemeral disks" disks that are created and aren't volumes16:44
bauzasand are persisted16:44
sean-k-mooneyephmeral disks are only the disk created form flavor.ephmeral16:44
sean-k-mooneyso to me it wrong ot ever use it in any other context16:44
bauzasif you have a flavor with DISK_GB=10, correct me if I'm wrong but you'll get a 10GB disk that will be an "ephemeral disk"16:45
bauzas(not a BFV I mean)16:45
sean-k-mooneyno16:45
sean-k-mooneythat is incorrect terminology16:46
sean-k-mooneyand what im objecting too16:46
gibilocal vs volume, root vs ephermeral, these are the terminologies in the API doc16:46
sean-k-mooneyright16:46
bauzashttps://docs.openstack.org/arch-design/design-storage/design-storage-concepts.html16:46
bauzasEphemeral storage - If you only deploy OpenStack Compute service (nova), by default your users do not have access to any form of persistent storage. The disks associated with VMs are ephemeral, meaning that from the user’s point of view they disappear when a virtual machine is terminated.16:46
bauzaswe name them "ephemeral" because we delete the disk when the instance is removed16:47
gibiand we delete volumes as well if so configured :D16:47
sean-k-mooneywhihc is not correct16:47
sean-k-mooneydeleting somting at the end of its lifetime does not make it epmeral16:47
bauzastechnically, you can have ephemeral disks on shared storage (NFS)16:48
sean-k-mooneyif we delete a volume we also expect the data to be removed16:48
sean-k-mooneyor ceph16:48
sean-k-mooneyif you use the rbd image_backend16:48
bauzassean-k-mooney: I don't disagree with you, I'm just explaining our docs16:48
sean-k-mooneyyep16:49
bauzasnot only our upstream docs btw... https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/16.2/html/creating_and_managing_instances/con_types-of-instance-storage_osp :)16:50
sean-k-mooneyso i would prefer something like local_disk_encryption, instance_disk_encypriton or similar 16:50
sean-k-mooneybauzas: right but there are many other things wrong with our downstream docs16:50
gibi+1 on local disk encryption16:50
bauzasI'd prefer local disk encryption over instance disk encryption16:51
sean-k-mooneyim not going to block this progressing on the name16:51
sean-k-mooneyi just find the usage of ephmeral to rever to storage tied to the vms lifecycle to be mildly insulting16:52
bauzasbecause an instance disk can either be "ephemeral" (from local libvirt disk), or something else16:52
bauzastechnically, this ephemeral storage is actually our virt driver local storage16:53
bauzasdepending on the virt driver you use16:53
sean-k-mooneyits not nessiarly local16:53
bauzastell ironic about it :)16:53
sean-k-mooneywell for libvirt or vmware16:53
bauzasyeah, see, I made that mistake16:53
sean-k-mooneythe storage can be clusterd or remote16:53
bauzasshared storage is ephemeral16:53
bauzasso, not local 16:54
sean-k-mooneythis is why i dont like using that term 16:54
melwitthey o/ thanks for discussing the naming bikeshed :) just wanted to get some confidence that ppl would prefer a name change and lessen the possibility of someone coming to the review later and saying "why did you change this, it should be changed back to ephemeral"16:54
sean-k-mooneyephemeral is ambiguious16:54
bauzasas is "evacuate" :D16:54
bauzasour Gods of Naming didn't help16:54
sean-k-mooneymaybe we need dansmith to write another blog post16:55
bauzasproblem solved.16:55
sean-k-mooneymelwitt: ack so i guess the real question is16:55
bauzas#action dansmith to write a blogpost16:55
sean-k-mooneydo we think channging the name will make thing clearer16:55
dansmithheh16:55
bauzasoh, snap, he saw it16:55
melwitt:)16:55
bauzas#undo16:55
opendevmeetRemoving item from minutes: #action dansmith to write a blogpost16:55
dansmithIn general, I'm not for renaming things like this16:55
bauzasI'm not against renaming ity16:56
bauzasI just wanted to make sure we all agree on what this is16:56
bauzasephemeral is a bad name, but that's a name we already use16:56
dansmithbecause you'll end up with all old docs being inaccurate for new stuff, and people who already understand this will also have to change16:56
melwittthat was one of my concerns16:56
bauzasif we pick something else, this has to be better understandable about what it is16:56
bauzasyeah, if we need to write some doc explaining "ephemeral" == "this new thing" this is bad16:57
bauzashence the challenge16:57
gibiso this is considered a non fixable terminology mistake of the past?16:58
bauzaslike tenant ? :)16:58
gibiwe are fixing tenant16:58
sean-k-mooneywell to me that doc that is erfernce is not a nova doc16:58
bauzasI know, I'm opening a can of worms16:58
sean-k-mooneyso we could jsut fix it16:58
bauzasand pretend it never existed, heh ? :)16:58
bauzaswe're running out of time, but for the sake of the conversation, let's continue16:59
sean-k-mooneywell from my point of view the only thing that nova ever said was ephemreal is the falvor.ephemeral storage disks16:59
bauzasI'll just formally end the meeting at the top of the hour16:59
sean-k-mooneymelwitt: are we encypting the flavor.epmermal disks by the way16:59
sean-k-mooneyor just root and swap17:00
bauzasproblem is17:00
sean-k-mooneyi thikn we will be encypting all 3 types17:00
bauzasroot is also "ephemeral"17:00
bauzas(depending on the conf options)17:00
sean-k-mooneybauzas: it depend on the difinition17:00
sean-k-mooneyform our api its not17:00
melwittsean-k-mooney: what is "flavor.ephemeral"? it is encrypting the root disk and any other attached local disks17:00
sean-k-mooneyin our flavor we have 3 types of storage17:01
bauzascorrect, the point is that *by default, we don't do any difference between root disk and other local (or non-local on shared) disk17:01
sean-k-mooneyroot, swap and ephemeral17:01
sean-k-mooneyhttps://docs.openstack.org/nova/latest/user/flavors.html17:01
melwittok, this is encrypting root and ephemeral, and not swap17:01
bauzasright17:01
bauzasabout the new feature17:02
sean-k-mooneyso we proably should be encyrpting swap too but we can maybe add that next cycle17:02
bauzasswap is out of scope AFAICT17:02
sean-k-mooneyim not sure why it would be17:03
sean-k-mooneywe declared it out of scope for this cycle i guess17:03
sean-k-mooneybut i would hope it woudl get done before we condire this fully complete17:03
bauzasbecause swap isn't using QEMU file-based storage ?17:03
sean-k-mooneyit is17:03
sean-k-mooneydepending on your backend17:04
bauzasf***17:04
sean-k-mooneyit will use a qcow files or a rbd volume17:04
bauzasI'm not expert on swap17:04
bauzasthen, all disks (root, swap and others) go into a same bucket17:04
bauzaswhich is by default the virt driver storage backend17:04
melwittbasically this is encrypting things that are under the 'ephemerals' and 'image' keys in block_device_info: https://review.opendev.org/c/openstack/nova/+/826529/7/nova/virt/driver.py#10717:05
melwitt'swap' has its own key in block_device_info17:05
sean-k-mooneyso ephemerals should be the storage form flavor.ephemeral_gb17:06
bauzasdo we know if https://docs.openstack.org/nova/latest/configuration/config.html?highlight=ephemeral#DEFAULT.default_ephemeral_format is also used for root and swap ?17:06
sean-k-mooneyimage is presumable the storage form flavor.root_gb17:07
melwittI don't know the reason swap is not included and I just checked the specs again and don't find it mentioned why17:07
sean-k-mooneybauzas: no i belive that is for ephemeral_gb only17:07
sean-k-mooneybauzas if you dont specy how you want flavor.ephemeral_gb to be devied up on the server create api request17:08
sean-k-mooneywe use that config to determin the format17:08
sean-k-mooneyand we provide a single ephemeral disk17:08
sean-k-mooneybut you can ask for nova to provide multiple disks as long as the total is equal to or less then flavor.ephemeral_gb17:09
bauzaslooks like I need to end this meeting17:09
bauzasbut let's continue17:10
sean-k-mooneythis gets modeled in the block device mapping info passed in the api request17:10
sean-k-mooneyack17:10
bauzas#endmeeting17:10
opendevmeetMeeting ended Tue Jul 26 17:10:07 2022 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)17:10
opendevmeetMinutes:        https://meetings.opendev.org/meetings/nova/2022/nova.2022-07-26-16.00.html17:10
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/nova/2022/nova.2022-07-26-16.00.txt17:10
opendevmeetLog:            https://meetings.opendev.org/meetings/nova/2022/nova.2022-07-26-16.00.log.html17:10
bauzasI'm trying to see whether we have 'ephemeral' as an API contract besides the ephemeral disks17:10
sean-k-mooneythat depends on the definition of ephemeral17:11
sean-k-mooneyif you use ephemeral or not to imply falult tollerance then no17:11
sean-k-mooneyif you are using it to imply tied to lifecycle fo vm its no differnt then root other thatn its not included in snapshots17:12
bauzasfrom a BDM perspective, nothing changes between a root disk and an ephemeral disk, right? (by default, I mean)17:12
sean-k-mooneynothign major17:12
sean-k-mooneythey are slithgly differnt in that root will have an imave as it source17:13
bauzascorrect17:13
sean-k-mooneyand ephemeral disk are always empty17:13
stephenfintotally unrelated to ephemeral stuff, but I think we're finally reading to switch from mock to unittest.mock17:13
stephenfinThe removal of Python 3.6 support simplifies our job significantly since the unittest.mock lib there was buggy as hell17:13
sean-k-mooneystephenfin: i think i saw you respin that last week17:13
stephenfinDo we want to prioritize reviews of that patch so we can avoid rebase hell? https://review.opendev.org/c/openstack/nova/+/714676/17:13
stephenfinI think melwitt and sean-k-mooney have been interested in that in the past17:14
stephenfininterested/involved17:14
stephenfinsean-k-mooney: Yeah, I meant to do it sooner and, um, forgot17:14
sean-k-mooneyya so we likely need to land the unshleve to host patch first but then i would be open to landing that next17:14
sean-k-mooneybauzas: is the unshelve to host patch ready to emrge once the gate issue is resolved17:15
stephenfinIt's simultaneously low priority (mock works just fine) and high priority (so much chance of merge conflicts)17:15
bauzasstephenfin: that makes me a bit nervous17:15
sean-k-mooneybauzas: https://review.opendev.org/c/openstack/requirements/+/851002/ is merged by the way17:15
stephenfinsean-k-mooney: fine by me. I'm just raising this up now before I forget again 0:)17:15
sean-k-mooneybauzas: why?17:15
bauzasstephenfin: sean-k-mooney: because we're on zed-317:15
sean-k-mooneybauzas: i really dont think it will break anything17:16
bauzasand we'll create a ton of merge conflicts17:16
sean-k-mooneybauzas: right we punted it last time for the same reason17:16
sean-k-mooneybauzas: so i would prefer to either do this after zed-3 but beofre rc117:16
sean-k-mooneyor do it sooner rather then later17:16
bauzasdon't disagree17:16
gibido it sooner17:17
gibilater means we will have the same discussion again in the future :)17:17
bauzasafter zed-3 and before zed-rc1 seems reasonable, as is after we branch zed17:17
sean-k-mooneycan we merge it now17:17
stephenfinthe merge conflicts should only happen in the imports section since I've been able to remove most of the workarounds that were needed to workaround Python 3.6 bugs in unittest.mock 17:17
sean-k-mooneythet gate shoudl not be blocked17:17
sean-k-mooneythen merge the api change patches17:17
bauzasare we done with the ephemeral bikeshed ?17:18
sean-k-mooneybauzas: well we can continue but this was a nice break17:18
stephenfin(to be specific, only nova/tests/functional/regressions/test_bug_1781286.py and nova/tests/fixtures/nova.py needed more that a simple import replacement)17:18
melwittstephenfin: I added a small change in a func test in that patch ages ago to help it pass CI at the time. I wasn't sure if I should be +2ing it bc of that17:19
bauzasfor that unittest usage, well, I'd appreciate if we could target it post-FF17:19
bauzasbut cores are free to vote17:19
sean-k-mooneyif we agree ahead of time that this is oke to merge say on the monday after FF to give time for rechecks17:20
sean-k-mooneythen im ok with that otherwise i would prefer to merge it this week17:20
sean-k-mooneyinstead of defering to next cycle again17:20
gibiat FF we might have potential FF exceptions that and we will run against the clock17:20
stephenfinmelwitt: Yeah, that's the change in nova/tests/functional/regressions/test_bug_1781286.py. I've touched that code also at this point so I wouldn't like us to both abstain17:20
gibiimposing merge conflict that point feels worse than imposing it now17:21
sean-k-mooneyyep17:21
stephenfinbauzas: I'd rather deal with merge conflicts now than when the gate is hammered before FF17:21
stephenfinpersonally17:21
sean-k-mooneysame17:21
bauzasthen, looks like you have your two cores17:21
bauzas:)17:21
stephenfinWow, I proposed this in March 24 2020. 3 days before lockdown #1 (in Ireland, anyway)17:22
stephenfinOh, the innocence of pre-COVID times :)17:22
sean-k-mooneyheh17:22
sean-k-mooneyi set Review-Priority +2 so ill review this today17:23
sean-k-mooneybauzas: melwitt do we want to continue talking about ephemeral17:23
gibiyou can ping me tomorrow to review it17:23
* bauzas sends hugs to the 191 merge conflicts :)17:23
stephenfingibi: will do (y)17:23
stephenfinbauzas: They had it coming17:24
melwittstephenfin: ok, I'll help review if gibi or sean-k-mooney end up not being able to for some reason17:24
* sean-k-mooney wishes i had time to blakify the codebase without chanign linelenght too17:24
* sean-k-mooney whishes i had done the manual fixes in a seperat patch followed by the automated patch17:25
bauzasstephenfin: I don't see a hacking rule preventing us to import the mock lib17:26
stephenfinbauzas: it's here. I just need to rebase it https://review.opendev.org/c/openstack/nova/+/70876817:27
bauzasstephenfin: so I guess you ask reviewers to make sure we don't pull that lib again ?17:27
stephenfinI can do that now17:27
bauzasstephenfin: ok, gtk17:27
sean-k-mooneyack was just going to say seperate patch please17:27
gibiI need to disappeare. see you tomorrow17:29
sean-k-mooneygibi: o/17:29
sean-k-mooneybauzas: back to ephmeral if you look at https://github.com/openstack/python-openstackclient/commit/4da4b96296c6b6d4351ebd47e32d5049a88211f1#diff-6759a29d0fccaa3a8d26137549b909fa3b3925b71d5318f30fe9fe7021f8558eR1227 you will see how osc construts the bdms17:29
sean-k-mooneyfor swap and ephemeral17:30
sean-k-mooneyif you pass --swap or --epmeral to the server create17:30
* bauzas when I see https://review.opendev.org/c/openstack/nova/+/714676/ => https://www.youtube.com/watch?v=JDaD2G9xjMc17:30
bauzasI'm just wondering whether we'll be able to merge things :D17:30
opendevreviewKashyap Chamarthy proposed openstack/nova master: Add a workaround to skip hypervisor version check on LM  https://review.opendev.org/c/openstack/nova/+/85103417:31
sean-k-mooneybauzas: this is perhaps better https://github.com/openstack/python-openstackclient/blob/4da4b96296c6b6d4351ebd47e32d5049a88211f1/openstackclient/tests/unit/compute/v2/test_server.py#L2734-L2785=17:31
bauzassean-k-mooney: about the ephemeral encryption name, I can try to propose something17:31
sean-k-mooneysure17:32
bauzas"local BDM encryption"17:32
sean-k-mooneymaybe17:32
sean-k-mooneythey might not alwasy be type local17:32
dansmiththat's the proposed title of the feature/spec?17:33
bauzasdansmith: I said earlier, I'm terrible at naming17:33
bauzasI'm just quite giving up17:33
dansmithI think the problem is, things were named, and then we implemented lots of features that blurred all the lines we had, which makes not only the existing names less than ideal, but also makes it hard to accurately describe what we're talking about17:33
dansmithwhich pretty much means there's hardly any point in a rename of anything, IMHO17:34
bauzasthis feature is about to encrypt BDMs which are local17:34
bauzashence the countername17:34
dansmith"non-volume data disks" is probably the most accurate, but I mean, good lord :)17:34
sean-k-mooneybauzas: it also works for ceph volumes allocated by nova17:34
sean-k-mooneyso local is not really right17:34
dansmithbauzas: not always local :)17:34
sean-k-mooneyits about encypting non cinder sotrage17:34
melwittswap isn't included so "non-volume data disks" doesn't even work :P17:35
* bauzas has a serious headache by now, after the whole CI outage and now the massive mock change we gonna pull17:35
sean-k-mooneyalthough personally i woudl have perfered if it also work for bfv17:35
sean-k-mooneymelwitt: very true17:35
dansmithmelwitt: but swap isn't encrypted right? that's why I put "data" in there :)17:35
sean-k-mooneydansmith: what would the flavor extra specs be for that17:36
melwittephemeral_and_root_disk_encryption 😂 17:36
dansmithnvdd_encryption=FML17:36
melwittdansmith: oh, I see. yeah swap is not encrypted17:36
melwittI didn't get the "data" emphasis17:36
sean-k-mooneyhonestly i prefer that name to what we have17:36
sean-k-mooneytoo bad we cant just call it  nova_disk_encryption17:37
sean-k-mooneyi.e. encyuption for stuff nova owns17:37
dansmithnon-volume is the key I think, but it's not very nice to say/write17:38
sean-k-mooneyya17:38
sean-k-mooneyflavor_disk_encryption?17:39
sean-k-mooneybut no17:39
sean-k-mooneythat break for bfv17:39
sean-k-mooneywell sort of17:39
melwittand swap17:39
dansmiththe argument here is about the naming of the extra spec?17:39
sean-k-mooneyya swap i would honestly just add17:39
dansmithI mean to be honest, this feels like a *massive* waste of time17:39
sean-k-mooneydansmith: yes extra spec and image property17:40
dansmithhow about nova_disk_encryption <- excludes cinder, and if you're worried about swap, then add swap to it and move on17:40
sean-k-mooneyi would be happy with ^17:40
bauzaswfm17:41
sean-k-mooneyi just was not sure if we are ment to use the project name in things like this17:41
bauzaslet's just write a very good config option doc17:41
dansmithnova_ on this case makes it clear we're talking about the things nova owns17:41
sean-k-mooneysure and by config option doc you mean flavor validator doc17:41
dansmithcompute_ would work too I guess, but I think it's less clear17:41
sean-k-mooneyack ya nova_ i liek more but i would be happy with compute_17:42
sean-k-mooneyso either i think are ok17:42
dansmithnova_ and promise to never ever discuss this again? :)17:42
bauzasor say it 5 times in front of a mirror ?17:43
sean-k-mooney:)17:43
sean-k-mooneymelwitt: are you ok with that?17:43
melwittI guess. I'm focused on making sure everyone's happy with the name. I checked https://docs.openstack.org/glance/latest/admin/useful-image-properties.html and don't find any other project names there so that feels a bit weird, but ¯\_(ツ)_/¯ 17:45
bauzasvirt_disk_encryption ?17:45
dansmithintrinsic_disk_encryption, managed_disk_encryption17:45
bauzasdamn, I stepped into the ...17:46
dansmithvirt_ is not specific enough I think, because cinder disks aren't real disks17:46
dansmithnon_volume_disk_encryption.. uglier, but no project name and more accurate17:46
dansmithOR17:46
bauzasoh, I was poorly refering to the fact we defer the disk creation to the underlying virt driver17:46
dansmithwe explain that "ephemeral disks" are "everything but volume disks" and then we're back to the start!17:46
melwitt😂 ahhhhhh17:47
sean-k-mooneyya i still fine ephmeral kind of insulting to refer to novas storage but thats just me17:47
bauzas(18:56:47) bauzas: if we pick something else, this has to be better understandable about what it is17:47
bauzas(18:57:19) bauzas: yeah, if we need to write some doc explaining "ephemeral" == "this new thing" this is bad17:47
bauzas(18:57:39) bauzas: hence the challenge17:47
dansmithsean-k-mooney: you're fine or you "find" ?17:48
bauzascores, would appreciate a quick +2 on os-vif blocking https://review.opendev.org/c/openstack/nova/+/850998/217:48
sean-k-mooneywe have had customer go to great lents to do terible things because they found the term ephemeral unackceptable17:48
sean-k-mooneydansmith: i find17:48
bauzasor because they considered hostnames be FQDNs ?17:48
bauzassorry, this was easy17:49
sean-k-mooneybauzas: we should not need that anymore17:49
dansmithpersonally I think ephemeral is a pretty accurate name17:49
sean-k-mooneybauzas: its blocked in upperconstraits17:49
dansmithif they're local disks on compute nodes, then ephemeral means "they could go away if a single computer dies"17:49
sean-k-mooneyyep and so could cinder volumes17:49
dansmithand even when they're on ceph, they're treated not nearly as precious as data volumes17:49
sean-k-mooneydansmith: most of them dont provide ha17:49
bauzassean-k-mooney: we're pulling u-c on all jobs ?17:49
dansmithsean-k-mooney: but the intent behind a volume is long-term storage generally17:50
sean-k-mooneybauzas: we shoudl be17:50
bauzasI thought this was only on tox targets17:50
bauzasand not tempest17:50
sean-k-mooneybauzas: it would be a bug if we were not17:50
melwittsean-k-mooney: what's the customer did something bad bc they didn't like the term?17:50
opendevreviewStephen Finucane proposed openstack/nova master: hacking: force explicit import of python's mock  https://review.opendev.org/c/openstack/nova/+/70876817:50
sean-k-mooneywell i wont mention there name but they instead on doing pci passthough fo a raid controler intor ther vm because they did not accept novas ephmeral storage as accpetable17:51
bauzassean-k-mooney: I'll propose a new rev for https://review.opendev.org/c/openstack/nova/+/838976, we'll see17:51
dansmithsean-k-mooney: but that's just because either they didn't read docs, understand, or our docs suck17:52
dansmithand if they're making decisions based on their perception of one word without knowing what it means in context, then they probably have many other problems17:53
sean-k-mooneydansmith: yes they messed up the server toplogy and didnt realise we did not supprot block deivce passethough other then via cinder17:53
dansmithlike thinking cinder volumes are actually burnt ashes of former disks17:53
dansmithand nova computes explode to make new space clouds17:53
melwittlol @ burnt former disks17:55
sean-k-mooneyi think im to the point where i dont care anymore17:56
dansmithwell, jokes aside, I think we probably have to just pick something that won't be perfect and move on17:57
sean-k-mooneyack18:02
sean-k-mooneyill leave that up to melwitt 18:03
sean-k-mooneyif you want to keep it as is fine if we want to make it better for some defintion of better then also fine18:03
stephenfinbauzas: In case you didn't see, refreshed https://review.opendev.org/c/openstack/nova/+/70876818:04
* stephenfin drops to go pick apples in the sunshine 🍎👋18:05
opendevreviewEric Fried proposed openstack/nova master: hacking: force explicit import of python's mock  https://review.opendev.org/c/openstack/nova/+/70876818:42
opendevreviewSylvain Bauza proposed openstack/nova master: Reproducer for bug 1951656  https://review.opendev.org/c/openstack/nova/+/85067318:55
opendevreviewSylvain Bauza proposed openstack/nova master: Handle mdev devices in libvirt 7.7+  https://review.opendev.org/c/openstack/nova/+/83897618:55
sean-k-mooneystephenfin: heh removing the power vm driver is one way to adress that todo19:11
sean-k-mooneystephenfin: did we deprecate it last cycle due to the python2/3 issues19:11
opendevreviewsean mooney proposed openstack/nova master: Remove the PowerVM driver  https://review.opendev.org/c/openstack/nova/+/85034619:21
sean-k-mooneystephenfin: just rebased that on top of the hacking change ^ i think the fucntioal test is unrelated19:22
sean-k-mooneyit was a polison failure acessing os.uname in the libvirt driver19:22
sean-k-mooneyso that is not related to thsi change i think19:22
*** hemna3 is now known as hemna21:14
*** dasm is now known as dasm|off22:17

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!