Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2023-11-02
18:15:53 sean-k-mooney bauzas: we could just make max_instances apply to sriov case
18:16:04 sean-k-mooney since the non sriov case just works
18:16:35 bauzas with the non-sriov gpus, you need to specify all the cards
18:16:49 sean-k-mooney in the device list
18:16:52 bauzas giving them a way to say "I don't care"
18:16:58 bauzas seems good to me
18:17:07 sean-k-mooney ya so just add * to device_addreses
18:17:34 sean-k-mooney well im flexible on how we do that
18:18:09 bauzas we need implementation patches and then we could do anything we want
18:18:09 sean-k-mooney im ok to have a away to say "expose all devices of this type" i just woudl like it to be a little more explict then it is today
18:18:28 bauzas btw. I'm on PTO tomorrow
18:18:29 sean-k-mooney right os if we want to defer the details of how we do that for now
18:18:31 sean-k-mooney that works for me
18:18:41 sean-k-mooney ya im on pto tomorrow and monday
18:18:59 bauzas cool
18:19:10 bauzas I'll try if I can draft something before I leave
18:19:37 bauzas now that I have a machine, this is waaaay easier than before
18:19:57 sean-k-mooney yep hardware both sucks and make life simpler
18:36:04 bauzas sean-k-mooney: I made the wrong assumption, for non-sriov, max_instances is indeed meaningless
18:36:26 sean-k-mooney cool
18:38:09 bauzas so yeah we maybe need a wilcard
18:38:34 bauzas that said, we captured the opposite https://etherpad.opendev.org/p/nova-caracal-ptg#L669
18:38:57 bauzas but I recently discovered that every single day is full of adventures and news when it comes to nvidia
19:44:34 opendevreview sean mooney proposed openstack/nova master: add pyproject.toml to support pip 23.1 https://review.opendev.org/c/openstack/nova/+/899753
20:23:21 pas-ha[m] <bauzas> "dansmith: fwiw, there is also..." <- bauzas: I don't think I force anything with this patch, it should be 100% backward compatible - if no explicit devices are defined together with the single enabled mdev type, everything should work as before.
20:23:48 pas-ha[m] at least that was my intention, not sure I actually tested for that :-)
20:24:38 sean-k-mooney with a single mdev type bu tif you had 2 or more then the config section would be requried for each correct
20:26:14 sean-k-mooney hum so https://review.opendev.org/c/openstack/nova/+/899406/1/nova/virt/libvirt/driver.py#7933
20:26:27 sean-k-mooney is just always regestring the config options
20:26:33 sean-k-mooney but not nessiarly enforicing it
20:26:53 sean-k-mooney because you still defalt it here https://review.opendev.org/c/openstack/nova/+/899406/1/nova/virt/libvirt/driver.py#7958
20:27:34 sean-k-mooney pas-ha[m]: so ya i think your version is tecnically still backward compatible
20:28:33 sean-k-mooney pas-ha[m]: the fact your chagne has not unit test chagne means either its backward compatiable or we have no test coverage for this
20:28:59 pas-ha[m] sean-k-mooney: I bet the latter :-D
20:29:12 sean-k-mooney unfortunetly you might be right
#openstack-nova - 2023-11-03
06:45:24 bauzas pas-ha[m] : correct, you're right my bad, we are not changing the behavior for operators not using the groups, we just do this for those who want to use the groups while they only use one type
08:02:30 bauzas folks I'm off today see you on Monday
14:20:17 dansmith melwitt: have you seen vnc console test fails recently?
14:20:32 dansmith melwitt: I've seen this a couple times now: https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_fa0/899375/5/gate/nova-grenade-multinode/fa011a2/testr_results.html
17:49:23 melwitt dansmith: no, that one is new to me
17:56:36 melwitt dansmith: if the timestamps are synced then it looks like the console proxy was stopped and was not started again at the time of the test failure https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_fa0/899375/5/gate/nova-grenade-multinode/fa011a2/controller/logs/screen-n-novnc.txt
17:58:01 melwitt I also noticed there's no screen-n-super-cond.txt, do we run any jobs that don't use a super conductor?
17:58:32 dansmith idk, but if we do, I bet they'd be grenade jobs
#openstack-nova - 2023-11-06
12:11:55 bauzas dansmith: looks like we need to skip test_get_vnc_console from https://review.opendev.org/c/openstack/nova/+/899375/
12:12:19 bauzas we just start to test it by my change
12:42:07 gibi bauzas: or you need [compute-feature-enabled]vnc_console=true in the tempest config to enable the vnc console
12:57:00 bauzas gibi: I haven't checked the tempest config for this job, I can take a look
13:52:37 bauzas humpf, actually call me stupid, the job worked fine with the vnc test in a previous run
14:09:18 dvo-plv bauzas, Hello, Maybe you will have a time to review https://review.opendev.org/c/openstack/nova-specs/+/895924 ?
14:15:51 bauzas dvo-plv: sorry I'm currently doing a lot of things, but we plan to do a spec review day on tomorrow
14:16:45 dvo-plv Sure, thanks
14:33:13 bauzas dansmith: I know sean-k-mooney is on PTO, but the more I think, the less I like to create a specific config option for SR-IOV GPUs
14:34:31 dansmith bauzas: I didn't follow the rest of your conversation after our gmeet
14:35:11 dansmith are you talking about the max_devices one or what?
14:35:14 bauzas so I think I'll just create a new change for telling that we would support a new wilcard "*" for device_addresses which would be the default
14:35:18 bauzas dansmith: yeah
14:35:44 dansmith okay I'm not sure how that solves the prioblem
14:35:48 bauzas dansmith: I tried to see how we could use max_instances (or max_devices) for both SR-IOV and non-SRIOV GPUs, but it would be dififcult
14:36:08 bauzas so if max_instances would be only for SRIOV GPUs, it would be sadf
14:36:22 dansmith I thought the non-SRIOV ones didn't need it, and that's why the default would be None or -1 to not apply that to those
14:36:27 dansmith why?
14:36:37 dansmith it's very specifically only needed for those, IIRC
14:38:11 bauzas we have two different behaviours
14:38:25 bauzas either we ask for SRIOV operators to provide all the VFs they want,
14:38:46 bauzas or we try them to provide a specific option
14:41:39 bauzas anyway, let me try to provide a specific change and we'll see
14:42:20 dansmith mkay
15:10:36 opendevreview Balazs Gibizer proposed openstack/nova master: [libvirt]Add migration_inbound_addr https://review.opendev.org/c/openstack/nova/+/900203
15:11:28 gibi bauzas, sean-k-mooney: ^^ it belongs to https://blueprints.launchpad.net/nova/+spec/libvirt-migrate-with-hostname-instead-of-ip which I will bring up for approval on the tomorrows meeting
15:13:19 bauzas okay, but too many things I work on
15:13:31 bauzas f nivia
15:17:43 gibi this is just a fyi, we will discuss it further tomorrow on the meeting
16:10:21 opendevreview Merged openstack/nova master: Adding server actions tests to grenade-multinode https://review.opendev.org/c/openstack/nova/+/899375
16:13:31 bauzas \o/
17:32:50 auniyal6 bauzas can this be merged https://review.opendev.org/c/openstack/nova/+/865336, it was created by gibi then sean backported it till wallaby
17:33:28 auniyal6 this whole topic only wallaby this pending https://review.opendev.org/q/topic:bug%252F1970467
17:33:53 auniyal6 this will be helpful for other backports
17:36:30 bauzas sorry, can't do it, other priorities for the moment due to f... nvidia issues
17:37:08 auniyal6 ack
17:39:51 dansmith that seems like a pretty impactful backport to me, just skimming
17:40:09 dansmith I'd like to chat with sean about it tomorrow when he's back before I review/approve
17:40:23 auniyal6 thanks dansmith
17:40:50 bauzas my head is just about to explode
17:41:09 bauzas too many things on the pile, that's it
18:41:06 opendevreview Amit Uniyal proposed openstack/nova stable/yoga: Fix the PCI device capability dict creation https://review.opendev.org/c/openstack/nova/+/899252
18:41:07 opendevreview Amit Uniyal proposed openstack/nova stable/yoga: Store pf_mac_address and vf_num in extra_info https://review.opendev.org/c/openstack/nova/+/899253
18:41:08 opendevreview Amit Uniyal proposed openstack/nova stable/yoga: Translate VF network capabilities to port binding https://review.opendev.org/c/openstack/nova/+/899254
20:15:39 opendevreview Pierre Riteau proposed openstack/nova master: Remove outdated comment about allocation ratios https://review.opendev.org/c/openstack/nova/+/900242
22:16:51 opendevreview Sylvain Bauza proposed openstack/nova master: WIP: add max_instances and a wildcard https://review.opendev.org/c/openstack/nova/+/900250
22:16:52 opendevreview Sylvain Bauza proposed openstack/nova master: WIP: Support to pass PFs in device_addresses https://review.opendev.org/c/openstack/nova/+/900251
22:35:26 opendevreview Steven Relf proposed openstack/nova master: Adding basic auth to dynamic vendordata api calls https://review.opendev.org/c/openstack/nova/+/900252
#openstack-nova - 2023-11-07
07:50:16 bauzas happy spec review day, everyone :)
09:21:56 gibi I did my first pass on the spec. I skipped re-proposes as those probably doesn't need much discussion anyhow
09:22:06 gibi s/spec/specs/
09:28:35 bauzas gibi: I still need to do my duty, still working on f... vgpus
09:35:51 opendevreview Balazs Gibizer proposed openstack/nova master: [libvirt]Add migration_inbound_addr https://review.opendev.org/c/openstack/nova/+/900203
09:38:07 opendevreview Balazs Gibizer proposed openstack/nova master: DNM: test hostname based migration URL https://review.opendev.org/c/openstack/nova/+/900275
09:42:37 opendevreview Sylvain Bauza proposed openstack/nova master: WIP: Support to pass PFs in device_addresses https://review.opendev.org/c/openstack/nova/+/900251
09:42:43 bauzas there we go, I'm done
09:43:02 bauzas I'll just await sean-k-mooney and dansmith for discussing about that

Earlier   Later